¹ D. Fryklund, Les changements de signification des expressions de ‘droite’ et de ‘gauche’ dans les langues romanes et spécialement en français 42 (Diss. Uppsala, 1907), and his critic A. Zauner, LGRPh. 30.278 (1909), are agreed that the rise of directus ‘*right’ (rather than 'straight') is traceable to Latin. The identification of 'straight' and ‘right’ cuts across the borders between major language families: Romance, Germanic, and Slavic. Note that replacement of words for ‘left’ and ‘right’ at frequent intervals is peculiar to several languages; MHG zese yielded to G recht, MHG winster was overlaid by G link; OE swidre has given way to right, OE winstre to left.

In Old Spanish, derecho was occasionally used for 'right (side)': Apolonio 159c (add 106b, 264a 'favorable', of winds, < ‘right’, in the language of augurs); Poema de Alfonso Onceno 1313d; Carlos Maynes (ed. Bonilla), fol. 145ro; Rimado de palacio, E, 1518b; Confisión del amante, fol. 128ro. Otherwise derecho, as an adjective, signified ‘upright, straight up, just, fair, legitimate’, cf. derechero ‘fair’ (paralleling justiciero), derechamientre contrasted with infintosamientre (Barlán e Josaphá [ed. Moldenhauer], fol. 139ro). Throughout the Middle Ages the standard word for ‘right’ was diestro: Cantar de Mio Cid, lines 216, 750, 753, 810, 1548, 2010, 2573; Fuero de Avilés §87; Fuero de Guadalajara, see V. R. B. Oelschläger; Fuero de Teruel (ed. Gorosch) §90, §136, §335, §336, §579, §599, §679; Milagros 483b; Santa Oria 58c; Sacrificio de la misa 51a, 219d, 221ad; Apolonio 260a; Santa María Epigciaca (ed. Knust), fols. 9ro, 13ro; Estoria del Rrey Guillelme, fol. 34ro; Cuento del Emperador Ottas, chs. 12, 21 (α diestro et a siniestro); Libro de Josep ab Arimatía, fols. 272vo, 275vo; Barlán e Josaphá, fols. 116ro, 122vo, 168ro; Confisión del amante, fol. 209vo (a diestro e a siniestro). Diestro ‘adroit, skillful’ was infrequent as yet (Confisión del amante, fol. 372vo). OSp. a diestro e a siniestro exactly parallels OFr. destre et senestre, especially in descriptions of blows dealt out by valiant fighters surrounded by superior forces (cf. Godefroy's quotation from the Chevalerie Vivien, MS Boulogne-sur-mer 192, fol. 85c).

As derecho gradually encroached upon the territory of diestro, the latter, withdrawing from its central position ('right side'), held its own in a marginal area of meaning ('adroit') which in time developed into a new center. Thus the main process—substitution of derecho for OSp. diestro 'right' —on close inspection dissolves into a multitude of minute shifts and advances, with many words originally restricted to the position of mere satellites, auxiliaries, and occasional variants (OSp. derecho ‘right’ beside more frequent diestro; OSp. diestro ‘skilled’ beside mañoso and other dominant expressions) taking the place of long prevalent words. Cf. W. von Wartburg, Betrachtungen über die Gliederung des Wortschatzes und die Gestaltung des Wörterbuchs, ZRPh. 57.296–312 (1937), and the pertinent chapter in the same writer's Einführung in die Problematik und Methodik der Sprachwissenschaft (Halle, 1943). On the relations between dexter and dīrectus, see also FEW 3.62–3,90 (1928).

^1a The changes produced recall those caused by the mutual attraction of near-synonyms or by the integration of individual forms (or small groups of words) into closely coherent major series, such as numerals, the names of the weekdays (Sp. lunes, miércoles after martes, jueves, viernes), and the months (Ptg. Outubro, Novembro after Janeiro, Fevereiro ..., which characteristically initiate the series). See C. Merlo's thesis (1904) and its review by J. Leite de Vasconcelos, RL 8.226–8; the statement on serial replacement, of OFr. ester, seoir, getir and of Lat. hiems and aestās by A. Carnoy, RR 8.172 (1917); the latest studies on the names of the days of the week (H. P. Bruppacher's dissertation and its reviews; articles by W. von Wartburg and A. Henry); the controversy between D. Internoscia, R. A. Hall Jr., and L. Spitzer on the It. poserai series, It. 18.51–4 (1941), 20.198–200 (1943), and 21.154–69 (1944); and Spitzer's analysis (surely not definitive) of Sp. fulano, mengano, zutano, Ptg. sicrano, AILC 1.30-S (1941–2).

² M. Bréal, Essai de sémantique 76 (Paris, 1899); A. Waag, Bedeutungsentwicklung unseres Wortschatzes [based on H. Paul] 48–50 (Lahr i. B.); K. Xyrop, Grammaire historique de la langue française² 1.144 (Copenhagen, 1904), discussing septentriōnālis ˜ merīdiōnālis (instead of *merīdiālis, cf. merīdiānus); Fr. mérovingien ˜ carlovingien (for carolingien); La Suisse allemande ˜ La Suisse romande (for romane); évanouir ˜ épanouir (< OFr. espanir). In his book on French semasiology, Nyrop studies words (substantives, verbs, prepositions) cumulating opposite meanings: their number has sharply decreased in recent centuries, in part no doubt through the interference of grammarians; see Grammaire historique 4.43–9 (Copenhagen, 1913).

The tendency to bring out sharply the association by contrast ‘right’ ˜ ‘left’ by endowing the two words with similar endings is limited, within Indo-European, to Latin and Greek ( but stands apart).

^2a At least two more chapters could have been added to this article: one on the effect of polarization on Romance recomposition (G. Serra, Noterelle etimologiche sarde, ZRPh. 66.361 [1950], stresses that ignōτus, rather than *ignǏtus [cf. agnǏtus, cognǏtus], is duo to the pressure of nōtus), and one on the analogical termination of words (in Old Portuguese, the Gallicism sages ‘wise’ is patterned on simprez ‘naive’; both occur side by side in Vida e feitos de Julio Cesar, fol. 15vo. see BF 3.63. and cf. English male ˜ female).

³ On the intermittent occurrence of sinexter in Late Latin manuscripts, see H. Schuchardt, Der Vokalismus des Vulgärlateins 1.38, 3.12 (Leipzig, 1866–9), and W. Meyer-Lübke, Einführung in das Studium der romanischen Sprachwissenschaft³ 181 (Heidelberg, 1920). Ernout-Meillet, Dictionnaire étymologique de la langue latine² s.v., opportunely recall the reverse spelling sinixtra in Isidor's Origines 11.1.68. A significant abortive attempt to equate the endings of ‘right’ and ‘left’ mentioned by Festus (104.12) is seen in LAETRUM, laetrō(r)sum, after dextrum, dextrōrsum (unless one prefers to assume synonymic attraction by sinistrum, sinistrōrsum); see Ernout-Meillet (who speak of influence), and J. B. Hofmann in his revision of A. Walde's Lateinisches etymologisches Wörterbuch (who speaks of a blend).

⁴ For details, see the excellent entries in Ernout-Meillet; the bisemanticism of laeuus, scaeuus, and sinister (‘auspicious’ ˜ ‘ominous‘) is traced to conflicting rites in the patterns of cultures converging upon archaic Rome (Etruscan vs. Greek).

⁵ On the Indo-European level there was already a greater number of expressions for ‘left’ than for ‘right’, each of them, of necessity, rather thinly represented in the individual languages; see Ernout-Meillet's Dictionnaire. An analogous situation emerged (independently?) in Romance, so that the pattern of distribution of lexical units appears invariable amid the rapid wear and tear and constant renewal of the resources. The names of the male domestic animals, especially those used for breeding, show a similar clash between diversity of forms and immutability of the pattern throughout the Indo-European and the Romance stages.

Basque ezker, Prov. esquer(re), Sp. izquierdo, Ptg. esquerdo were cited by H. Schuchardt, Zum Iberischen, Romano-Baskischen, Ibero-Romanischen, ZRPh. 23.200 (1899), in his polemics against J. Vinson as an argument in favor of the assumed ancient ties between Iberian and Basque; Ptg. esquerdo, Schuchardt surmised, precluded the possibility of late diffusion of the word from the Basque Provinces. Actually Sp. izquierdo was preceded by isquierdo (Escorial Glossary No. 344) which, in turn, goes back to esquierdo; this form, in spreading westward, may have been transformed into esquerdo. Examples of OSp. esquierdo: Libro de Josep ab Arimatía, fol. 275vo; Poema de Alfonso Onceno 1319c, 1334c; Toledo Glossary (ed. Castro) No. 2252. V. R. B. Oelschläger, A Medieval Spanish Word-List 85b (Madison, [1940]), documents the early use of the proper name Esquierdo (Sahagún, a.d. 1171; Mayordomo, a.d. 1203; Burgos, a.d. 1209); cf. OJud.-Sp. izquierdero ‘one tied by his right hand’ (M. Gaspar Remiro, BRAE 5.642 [1918]). The standard word for ‘left’ was anciently seniestro (especially in western texts: Carlos Maynes, fol. 132vo; Barlán e Josaphá, fol. 116ro) and, with metaphony, siniestro (Cantar de Mio Cid, lines 12, 2691, 2694; Burgos, a.d. 1213, see Menéndez Pidal, Documentos lingüísticos: Reino de Castilla, No. 163; Sacrificio de la misa 48c: anda diestro siniestro, 50a, 51b, 200d, 219c; San Millán 37c; Santa Oria 58b; Fuero de Teruel §247, §335, §336; Santa María Egipciaca, fol. 13ro; Cuento del Emperador Ottas, ch. 21; Barlán e Josaphá, fol. 168ro; Rimado de palacio, N 1166c; Confisión del amante, fols. 128ro, 209vo: a diestro e a siniestro; El Corbacho, fol. 10vo: por syniestros senderos, cf. fol. 11ro: por sendas non conoscidas). Moreover, there existed the rare type sinistro: Signos 25b; Barlán e Josaphá, fol. 122vo; Libro de Josep ab Arimatía, fol. 273ro. K. Pietsch, Spanish Grail Fragments 2.119 (Chicago, 1924–5), was undecided whether a Latinism or simplification of the diphthong -ie- before the group -str- (cf. > OSp. aviespa > Sp. avispa ‘wasp‘) was involved. Comparison with It. sinistro strengthens the first alternative, for which the figurative use of sinister, in particular by ecclesiastic writers (anent the segregation of the righteous from the wicked) provides the cultural justification.

At least two Old French designations of ‘left’ are classifiable as Germanic: eslenc, -quɛ goes back to a Frankish stem influenced by clīnāre ‘to bend’, to which Sp. enclenque ‘feeble sickly’ indubitably belongs (to be added to REW³ No. 8029: slink); for older bibliography see Fryklund, op.cit. 36. Fr. gauche (attested toward the close of the Middle Ages) which succeeded in evicting senestre after 1500, is a back-formation from gauchir ‘faire des détours’, as was already known to A. Darmesteter, La vie des mots étudiée dans leurs significations³ 167 (Paris, 1889), who labeled the substitutes droit and gauche as bizarre, obscure, and awkward. Gauchir (documented since the late 14th century) is traced by E. Gamillscheg (1928) and O. Bloch (1932) alike to a cross of OFr. ganchir, guenchir ‘to turn, to twist’ < Frank. *wankjan ‘to make tremble’ (cf. G. wanken) and OFr. gauchier ‘to full, *to walk in an odd way’ < Frank. *walkan. Divergent opinions of Ménage, Littré, Mistral, Diez, Mackel, Jud, Sainéan, Schuchardt, and Thomas are discussed by Fryklund (64–9); Gamillscheg, Etymologisches Wörterbuch der französischen Sprache s.v. (Heidelberg, 1928); and Meyer-Lübke, REW³ No. 9492 (Meyer-Lübke sides with those scholars who eliminate *wankjan as one of the sources of gauchir). The association of ‘left’ with ‘weak’ is self-explanatory; the image of ‘curved’ underlying some expressions for ‘left’ goes back (not necessarily in a direct line) to remote antiquity; see G. S. Lane, Etymological miscellany, Lg. 11.195 (1935), and Ernout-Meillet, Dictionnaire s.v. laeuus, with a reference to Servius' commentary to Georgica 3.55.

Replacements drawn from Latin stock include It. manco, mancino, lit. ‘mutilated’ (REW³ No. 5285); Prov. senec, lit. ‘old, senile’; Sp. redruña, derived from retrō ‘backward’ (REW³ No. 7269); and possibly Ptg. canho, canhoto (REW³ No. 1595a: *canius). It. stanco, Raet.-Rom. čank, Rum. stâng, once traced to stāgnāre, are now assembled under unidentified *stancus (REW³ No. 8225). Prov. bigarro, Sp. zurdo (REW³ No. 50 dismisses Diez's, Baist's, and Spitzer's conjectures) and zoco (< soccu ‘wooden shoe’, with numerous fanciful extensions of meaning?) are of dubious ancestry.

⁶ Sinister < *s°nistros was a dual comparative, with the secondary suffix -tro- superimposed on primary -yes-/-is- as in Magister; the reverse sequence is visible in an-ter-ior; see, in addition to Ernout-Meillet, É. Benveniste, Noms d'agent et noms d'action en indo-européen 118 (Paris, 1948), who significantly regards dexter ˜ laeuus as the original, dexter ˜ sinister as the reshaped pair. This follows presumably from the existence of cognates of laeuus in Greek and Slavic, while sinister is isolated; from anthroponymic data (no group of names comparable to Laenius, Laeninus, Laeca, Laelius clusters around sinister); and from the early spread of the initial consonant of laeuus to laeuir (lēuir) ‘husband's brother’, cf. Skt. devár-, OCS . Scaeuus also has a counterpart in Greek σκαιός, and the diminutive scaeuola developed into a proper name. The findings of Indo-Europeanists and Latinists thus show that the change of sinistru to sinestru, [senεstru] is but the last phase of a much more protracted and intricate process of lexical polarization. Scaeuus and σκαιός, incidentally, were associated by the ancients; in the environment of Greco-Latin bilingualism, scaeuus is believed by some to have evolved the meaning ‘clumsy’ under the pressure of σκαιός, while scaeuitās (Aulus Gellius, Apuleius, Ammianus Marcellinus) was coined in imitation of σκαιότης.

⁷ R. Menéndez Pidal, Cantar de Mio Cid: Texto, gramática y vocabulario 628 (Madrid, 1908–11), traces diestro to dexteru; dextru is possibly the more accurate base. Only the forms without -e- have counterparts in Oscan and Umbrian and were dominant in Imperial Rome; yet the comparative, since Varro's time, was invariably dexterior.

⁸ In Modern Spanish diestro means ‘able, skillful, handy, sly’, and is closely connected with destreza ‘skill, nimbleness’ and ad(i)estrar ‘to guide, to teach, to train’ (attested since the Cantar de Mio Cid); the formulaic sequence a diestro y (a) siniestro ‘right and left, recklessly’ is a mere literary reminiscence. Siniestro, in addition to ‘unlucky, inauspicious’, has, as a substantitive, developed the new special meaning ‘shipwreck, disaster; damage, loss at sea’; the earliest nautical use of siniestro was in reference to adverse winds. Thus, semantically, diestro and siniestro have been drifting apart ever since losing their primacy to derecho and esquierdo (izquierdo) and retreating to small, asymmetric segments of their original areas of meaning. The process took over a millennium: occasionally, dexter is on record in Imperial Latin as signifying ‘skillful’, cf. the correlated adverb dext(e)rē and the abstract dexteritās (after δεξιότης?). On the rivalry between destre and droit, senestre and esclenc, later gauche, in French, see Fryklund, op.cit. 28; O. Bloch, Dictionnaire étymologique 1.233a.

⁹ The parallelism pervades the structure of the polarized word families: cf. dext(e)rā and sinistrā, on the pattern of extrā, used as adverbs, exceptionally also as prepositions; the archaic superlative dextimus (until Sallustius, later dexterrimus in Palladius) is echoed by sinistimus, in speaking of auspices; Pliny the Younger's sinisteritās is modeled on dexteritās; sinistrōrsum (-us) beside laeuōrsum ‘to the left’ calls to mind dextrōrsum ‘to the right.‘ Cf. OFr. senestral ˜ destral, senestror ˜ destror, senestrier ˜ destrier.

¹⁰ On Cat. ensenestrar, ensinistrar, see L. Spitzer, HR 12.177 (1944); on Jud.-Sp. sinistrar ‘to go to the left,‘ Ptg. esquerdear, Sp. izquierdear (yzquierdar: Ferrara Bible, Gen. 13.9), OFr. senestrer (Tristan of Thomas; later Jean Molinet's chronicles), Jud.-Fr. assenétrer, see Fryklund, op. cit. 39; R. Levy, Recherches lexicographiques sur d'anciens textes français d'origine juive (Baltimore, 1932), §116, and The vocabulary of the Escorial Manuscript I. j. 4, HR 11.62 (1943).

¹¹ See the excellent presentation of this state of affairs by Ernout–Meillet, Dictionnaire S.V. SOCER, SOCRUS (948–9).

¹² Characteristically, the prototype of socrus in the parent language referred exclusively to the husband's mother in her relations to the bride and daughter-in-law; it was secondarily extended to include the wife's mother in her relations to the son-in-law. Moreover, in several Indo-European languages, including the Slavic group, the word for ‘father-in-law’ is believed to have been patterned on the word for ‘mother-in-law’, a most unusual sequence of events casting light on the status of the mother-in-law in the ancient family, since normally in such leveling processes the feminine is assimilated to the masculine. Thus, in the Dominican Republic ‘daughter-in-law’ is occasionally expressed by yerna instead of by nuera, to conform to yerno ‘son-in-law’ < generu, on the analogy of hermana ˜ hermano, prima ˜ primo, hija ˜ hijo, sobrina ˜ sobrino (only the oppositions padre: madre and marido: mujer are unshakable), see R. E. Jiménez, Del lenguaje dominicano 107 (Ciudad Trujillo, 1941); the coinage of *nuero ‘son-in-law’ would have had a strongly derogatory effect.

¹³ Ernout-Meillet, Dictionnaire 689, mention much older attempts to replace the -o- of the IE base *snusó- by -ā-, inferred from Skr. snǔṣā, OHG snur, Alb. snusε. The Vulgar Latin process, pieced together with the aid of Romance forms, is thus integrated in a vastly broader development.

¹⁴ The tendency to polarize terms of kinship according to sex rather than according to generation or other categories has lately been very powerful in the Romance languages. The types cognātus ‘*brother-in-law’ and cognāta ‘*sister-in-law’ (the semantic specialization dates from the 4th century), eminently successful in Romance, represent this trend as clearly as does, in a negative way, the rapid obsolescence, within the confines of Latin, of Glōs, glōris ‘husband's sister’ (since Plautus; see P. Festus 87.16), whose Russian cognate (zolóvka) is still in use; of laeuir, lēuir ‘husband's brother’ (Dig. 38.10.4), excluded from literary usage; of ianitrīx < *ianiter ‘wife of the husband's brother’ (CGL 2.446.58: 'σύνvυμφος'; pl. ianitrīcēs in Digesta and, near the threshold of the Middle Ages, in Isidore), patterned on genetrīx, again with Slavic congeners, now extinct; see Ernout-Meillet, Diet. 466, 469, 541, 948–9, who characterize glōs, laeuir, ianitrīx as so many labels for parts of the newly-wed woman's environment, from which the mother-in-law (socrus > VLat. socra) emerges as the undisputed leader. Add to the score of these lexical losses frātria (Festus), frātrissa (Isidore) 'brother's wife', unrepresented in Romance, whereas sorōrius ‘sister's husband’ left traces in Northern France. The innovation *genera (which need not have existed in Latin) is represented in Raeto-Romance (Eng. schendem, Bergell gendra), Catalan (gendra), French Switzerland (Fribourg dzindra), and south-central France (Tarn-et-Garonne, Aveyron dǧendro), aside from Central America (Dom. yerna, see n. 12).

¹⁵ Scholars of the older generations were not aware of the complexity of this stratification; thus, G. Gröber, Vulgärlateinische Substrate romanischer Wörter, ALLG 4.134–5 (1887), placed on a single plane Sp. nuera, Ptg. Cat. Prov. Sic. nora, Rum. noră, noru, It. nuora, Mil. næura, to which he inadvertently added OFr. nore without the requisite qualification. In analyzing the words for the ‘parents-in-law’, Gröber (ALLG 5.472) correspondingly failed to segregate It. suocero, -a from the rest of the reflexes of VLat. socru, socra.

The destinies of VLat. socru, socra were largely shared by cōnsocer (REW³ No. 2166): It. consuocero, Eng. konsör, Sp. consuegro, Ptg. consogro, Lig. sközer ('father-in-law'), with corresponding feminines; note Daim. consegro. But Rum. cuscru presupposes the erratic stress-pattern , which implies dissociation from socru, even as Fr. cousin < cō(n)sobrīnu, with its syncope and abnormal simplification of secondary -sr- to -s- (REW³ No. 2165), drifted away from sobrīnus 'nephew' (REW³ No. 8050) as preserved in Ibero-Romance and the western branch of Raeto-Romance. The type progener ‘granddaughter's husband’ (P. Festus 257.2) is unknown to Romance.

¹⁶ Nura and socra are both documented as current vulgarisms in the anonymous Appendix Probi, tentatively traced to the 3d cent. a.d. (Nos. 169, 170). The core of the territory in which the variants soceru and socera have prevailed is Central (and, initially, Northern) Italy, cf. It. suocero, -a, whence it spread to the south of the Apennine Peninsula, to largely re-populated Sicily and to Northern Sardinia (sozaru), exposed in the late Middle Ages to Genoese and Pisan influence. All over Northern Italy this primary type was subsequently discarded in favor of messer and madonna (originally formulas of respectful address), which led to a sharper cleavage between ‘mother-in-law’ and ‘daughter-in-law’ than elsewhere in the Romania, the polarization in Northern France being re-established by the introduction of belle-mère ˜ belle-fille.

Generu (from gignō, -ere?) has fared worse than socru; notice its replacement by fīliastru, lit. ‘stepson’, in Basse Auvergne, Landes, Puy de Dôme, Picardy, Wallonie (with traces in Old French), and by fīliātus, lit. ‘the adopted one’ (cf. Sp. ahijado) in the Rhône basin; cf. E. Tappolet, Die romanischen Verwandtsch aftsnamen mit besonderer Berücksichtigung der französischen und italienischen Mundarten 129 (Strasbourg, 1895). Socru and generu, the names for the male partners of the marriage contract, have influenced each other as strongly as socra and nura, with the characteristic difference that the younger partner exercized the greater power of attraction: the -er of socer is probably, the -e- of socerī undoubtedly, due to the example of gener, -ī. The relative insignificance of the father-in-law in the older family, overshadowed by the mother-in-law and the son-in-law, is indicated by Arm. skesrayr, lit. ‘husband of the mother-in-law.‘

¹⁷ Phonological conditions may have accelerated the shift, yet Ernout-Meillet's statement (689): '... nurus > nora (où u devant z non suivi de u a passé à o)' leaves out of the reckoning the stratification of Romance forms (which Meyer-Lübke's dictionary puts into bold relief), and seems to project into an implicitly early period a process relatively late, to judge by the aloofness of marginal Romance dialects. On the other hand, Meyer-Lübke, with a Romanicist's competence, quite properly rejects the Proto-Lat. *norus reconstructed by F. Sommer, Lateinischer Vokalumlaut in haupttonigen Silben, IF 11.326 (1900); B. Bianchi, Storia dell' i mediano, AGI 13.198 (1892–4); O. Densusianu, Histoire de la langue roumaine 1.78–9 (Paris, 1901); and A. Ernout, Les éléments dialectaux du vocabulaire latin 204 (Paris, 1909).

¹⁸ Socerus is attested in Plautus, socrus ‘father-in-law’ occurs in glosses, as do socra (CGL 3.181.55: ), socera, and socrua. Late soceriō, -ōnis, an occasional expression for 'father-in-law', and sucerio—which in a single inscription, difficult of interpretation and conceivably involving Germanic interferences (CIL 3.11977, found in Straubing), may signify ‘brother-in-law’ in preference to moribund lēuir (see W. A. Baehrens, Sprachlicher Kommentar zur vulgärlateinischen Appendix Probi 107 [Halle a/S, 1922])—testify to the vitality of the radical socer-, as do present-day Italian dialects. Sidonius Apollinaris' socruālis, -e is distinctly artificial and discrepant with the trend of the spoken language.

¹⁹ See H. Tiktin, Rumänisches Elementarbuch 79 (Heidelberg, 1905); K. Tagliavini, Rumänische Konversationsgrammatik 184–5 (Heidelberg, 1938). In Rumanian, the late passage of nor(u) to noră precluded the diphthongization of the stressed vowel and thus separated noră from soacră ‘mother-in-law’ as regards the stem vowel, leaving it in the company of sor(u), soră ‘sister’ < soror, cf. the plural nurori patterned on surori < sorōorēs. In Italian, suora and nuora likewise form a pair apart, as was observed by Gaston Paris, Phonétique française: o fermé, Rom. 10.49 (1881), cf. his Mélanges linguistiques 248 (Paris, 1906–9); an extension of this similarity to Western Romance, which alone would have warranted Paris' attempt to attribute the reshaping of nurus to the agency of soror, is made impossible by the partial prevalence of germāna over soror, except in the ecclesiastical sphere, under conditions specified in P. Aebischer's meticulous study, ZRPh. 57.211–39 (1937). H. Schuchardt, in 1888, had casually thought of the influence of nouia ‘bride’ on nurus, yet the situation in Ibero-Romance (Sp. novia ˜ nuera, Ptg. noiva ˜ nora) fails to strengthen this hypothesis, which Meyer-Lübke mentioned in his Italienische Grammatik (41), while consistently preferring his own confrontation of nurus, nura with socrus, socra, as first stated in ZRPh. 8.205 (1884). Note that, while it is plausible to retrace Ptg. nora [] to a base with [], a skeptic's possible objection that an original closed o may secondarily have been opened through metaphony as in formoso ˜ formǫsa (cf. the reverse shift in nǫvo ˜ nǫva) cannot easily be countered.

²⁰ Some sources speak indiscriminately of OFr. nore, OProv. nora, Sp. nuera, etc., without specifying that the French form is solely found in the southwest of the territory of the langue d'oïl (Poitou, Aunis), which once belonged to the domain of the langue d'oc: the resulting compromise form nore, at a late date, was temporarily tolerated, rather than absorbed, by Standard French, while speakers were undecided whether to use bru or belle-file. (Godefroy's Dictionnaire quotes nore from Brantôme's Dames galantes, the Chronique bordeloise, and the Dictionnaire des trois langues of 1617.)

VLat. soCRU and socra yielded a diversity of forms in medieval texts: socre (Glossaire latin-français, MS Montpellier H 110; Bible, MS Richelieu 1); soegre (Benoit's Ducs de Normandie; Bible, MS Richelieu 1); soigre (Digestes, MS Montpellier H 4); sougre beside seugre (Conquête de la Morée); suegre beside sogre (Chastoiement d'un père); seucre (Ch. de Louis VIII, a.d. 1211); sevre (var. soivre) and suivre (var. seure) in Dolopathos; suire (in Wace's Brut, beside sogres; in Greban's Mistere de la Passion; in Digestes, MS Montpellier H 47); suere (Rôl. de bans de tréfonds); sire (Letter of Philip V, Arch. munic. de Rouen); soir (Chronique de J. de Stavelot). These forms, used promiscuously for ‘father-in-law’ and ‘mother-in-law’, are separated in Godefroy's documentation (7.590) according to sex, while they are here listed in the order of increasing distance from their Latin prototypes. Socre is visibly a Latinism; sogre, suegre, soegre, soigre, seugre are marked by -g- < -c- as retarded forms, reminiscent of aigre ‘sour’ < ācre, while the diphthongs ue, oe, oi, eu show successive stages in the development of Lat. []; seucre is seugre, partly disguised as a Latinism, by way of concession to socre; sevré, soivre, suivre faintly recall the development rogāre > OFr. rover; suire, sire, seure, soir represent the most advanced phase of the evolution, with total abandonment of the Latin velar.

Latinized bruta, of Germanic parentage (cf. Proto-Goth. *brūdis, Goth, brūþs), was first found in inscriptions at the fringes of the Roman Empire (Friuli, Dalmatia, Moselle area). Apropos of its encroachment upon Romance territory, Meyer-Lübke, in reviewing Tappolet's thesis (LGRPh. 17.131 [18961), properly mentions Rum. nevastă ‘wife’ < Slav. ‘bride’, whose borrowing emphasizes the foreign background and hence the doubly insecure status of the daughter-in-law, while the semantic shift shows the Daco-Romans as adopting the mother-in-law's point of view in classifj'ing the newlyweds. Bru was superimposed on the Gallo-Romans by the Franks pouring over Gaul from the northeast (bru, rare in the literary idiom at present, still persists with great force in eastern patois, inhospitable to bellefille) and its Raeto-Romance cognate is of Alemannic provenience; but bruman ‘groom’, brought to France by the Normans, spread from the north and northwest; see E. Walberg, Sur un mot français d'origine nordique, StX 16.39–49 (1943–4). Tappolet illustrates the shift ‘bride’ > ‘daughter-in-law’ with Western Lombardic (spasa < spōnsa: Bergamo, Palazzuolo, Brescia, Solicino) and Modern Greek (νύμφη).

²¹ Except that the starting point of the paraphrase was beau-frère (since 1386), belle-sœur (since 1463), with belle-mère (since 1454) and beau-père (since 1488), beau-fils and belle-fille (both since 15th cent.) following suit. The first word of ancient stock that succumbed was serorge < sorōriu ‘brother-in-law’, lit. ‘sister's husband’ (opposed to frātria ‘brother's wife‘), occasionally employed down to the 15th century; its status was affected by the collapse of the Old French declension, entailing the loss of seror < sorōre as distinct from suer < soror; add the instability of -rye, -nge in late Old French, recently made clear by N. Dupire. Saire and its variants, consistently ambiguous, fell easy prey to the spread of prefixal beau- (bel-), belle; gendre and bru (brus, bruz, brul) formed the rearguard of the withdrawing host of designations of kin acquired through marriage.

²² See Ernout-Meillet, Dictionnaire² 434.

²³ Grave has, at all times, been used in Spanish with the meanings ‘serious, difficult, hard, bitter’ (Santo Domingo 107d; San Millán 325c; Libro de buen amor 513a, 587b, 1133a, 1301a; Santa Catalina, fol. 17vo; Libro de Josep ab Arimatía, fol. 271vo; Barlán e Josaphá, fol. 120ro); for the rarer connotation ‘cruel, angry’, see Santa Catalina, fol. 20vo; Barlán e Josaphá, fol. 104ro. Note the frequency of the adverb: grabemientre (Glosas Silenses 100); gravemientre (Milagros 667b; also in the mutilated line 9 of Elena y María?); gravemente (Libro de buen amor 887a); gravement (Martirio de San Lorenzo 63b). For ‘heavy’, in the physical sense, Spanish uses pesado, from pē[n]sārk, tenderi; 'to weigh(t)', cf. It. pesante and OFr. pesant (Chanson de Roland), with the derivatives pesanteur and appesantir (16th cent.); pesant was later in part contaminated with besant < Byzantiu", name of a coin. OFr. lourd 'dizzy, giddy' (applying, even in modern Berrichon and some Occitanian dialects, to the staggers of the sheep) is ordinarily traced to *lǔrdu < lùridu ‘pallid, yellowish’, which in Prov. lort, It. lordo ‘dirty’ developed a different facet of its initially complex meaning. Only in modern times has lourd begun to crowd out pesant as the standard term for ‘heavy’. Concomitantly the word-family was reorganized, with alourder (eslourdir) and (eslourdissement) giving way to alourdir and alourdissement and new ties being established with the Ilalianism beslourd (?), balorde, balourd (e); see Bloch and von Wartburg s.v.

The other reason for the preservation of graue in the Iberian Peninsula was the early decline there of leue, replaced by liviano (Santo Domingo 621c, contrasted with pesado; Milagros 67b, as a qualifier of seso ‘brain, mind’, hence the figurative meaning ‘fickle, frivolous’ as early as Libro de buen amor 67c, 1347b; Barlán e Josaphá, fol. 111ro). Note rianda (Libro de buen amor 85a), los livianos (Duelo 49c) ‘lungs’: here again, the derivative is encroaching on the territory of leue; cf. Ptg. leves, Regg. Mod. Mant. slef, Sors. lef, Prov. leu, Cat. lleu. Somewhat different is the development of March. lebbe (REW³ No. .5006). With liviano go OSp. liviandat (Milagros 499c, Duelo 142d, Libro de buen amor 105d), livianeza (Confisión del amante, fol. 81ro; in contrast with graveza, ibid., fol. 258vo?), livianez (ibid., fol. 106ro). Another replacement of leue joined forces with liviano, exerting pressure from without, then deeply penetrating into the Hispanic lexicon: *leuiārius (RKW³ No. 5003) radiated from Northern Gaul to Italy (leggiero), Engadin (liǧer), Friuli (lidzer), Spain (ligero), Portugal (ligeiro), whereas the Catalan-Provençal congener leugier, lleuger indirectly gave rise to It. leggiadro; see F. d'Ovidio, Di alcune infiltrazioni d'italiano settentrionale nell'italiano letterario, Rom. 25.305–9 (1896); REW³ No. 5003, where leggiadria is selected as the starting point (cf. my own recent explanation of Sp. rebelde as patterned on rebeldía). Imported Sp. ligero, in contrast to Fr. léger, adopted the special shade ‘light (of movement), fleet, fast’, leaving the area of meaning ‘light (of weight)‘ to indigenous liviano.

Crushed between these two forces, leue survived vestigially in the vernacular stratum of Portuguese (leves 'lungs') and in the ancient Navarro-Riojan-Aragonese zone, with its back against the solid block of Gallo-Romance (Santa Oria S8d; Santo Domingo 406d); otherwise, it. appears in Old Spanish literature late, intermittently, and in strictly figurative use. at a time when Provençal-Catalan cultural influence was strongly felt (Cancionero de Baena, fol. 143vo, No. 416, anon.: Amigo señor, yo non visto sayo / de gloria tan vana e lieves pesares). Sp. leve and adjectival Ptg. leve are distinctly late Latinisms, restricted to exquisite style. Ptg. leo 'occasion', ao leo ‘at random, at one's discretion’ are due to the diffusion of Prov. leu (cf. beleu ‘soon’, Rouss. belleu 'perhaps'), cognate to Cat. lleu, OFr. lief, Eng. leif, It. lieve, see REW³ No. 5004. OCat. leu thus cumulated the meanings 'easy, easily, soon, fast', and ‘lung’; see Aguiló i Fuster, Diccionari 4.307. Notice also in the Libro de Alexandre, of disputed dialectal background, P 536b: de lieu (misinterpreted in O), P 870a: de leve (O 846a: de lieve) ‘easily’, recognized as imported from Provençal-Catalan by A. Morel-Fatio, Introduction to the Libro de Alixandre xxvii (Dresden, 1906), and by J. Keller, Contribución al vocabulario 116 (Madrid, 1932).

²⁴ Ernout-Meillet, Dictionnaire² 434, 541–2, quote the ‘couple antithétique’ grauis ˜ leuis from Plautus, Trinummus 684; Lucretius 2.225, 5.474; Cicero, Pro Roscio comoedo 2.6; idem, De lege agraria 2.17.45; idem, Pro Deiotaro rege 2.5.

²⁵ It. gravare, Prov. Cat. Sp. Ptg. gravar (Sp. grabar 'to engrave'), Vionn. se gravá 'to repent' might well have been separately listed by Meyer-Lübke s.v. grauō, -āre, and not umped together with the descendants and new derivatives of grauis (REW³ No. 3855). Cf. the Italian, French, Provençal, Catalan, Spanish, and Portuguese reflexes of aggrauāre (REW³ No. 279).

²⁶ Fr. grief has tended to become a substantive, cf. E. grief ‘sorrow’ and Lorr. grié ‘nostalgia’; Meyer-Lübke defines Grand' Comb. εt gri (Franche Comté) by ‘leid tun’. Cf. the adverbial relic grièvement (blessé) < griefment. On Cat. greu ‘grave; feixuc, pesant, molest’, greujar, greuje ‘grave, agravio’, greujos ‘quejoso’, see M. Aguiló i Fuster, Diccionari 4.185–6. The form grieve which Meyer-Lübke labels as Old Spanish is more properly Old Aragonese and represents the farthest advance into the Peninsula of the innovation centered in France; cf. grieu in Fueros de Aragón, B. Nac. MS 458 (ed. Tilander), §190.3 (but grant in the corresponding passage of MS Zaragoza 154); a grief (var. grieu) in the Fuero de Teruel §722.8; J. Yanguas y Miranda, Diccionario de palabras anticuadas que contienen los documentos existentes en los archivos de Navarra (Pamplona, 1854) s.v. algún: greu ‘agravio’, and Fuero de Navarra 3.20.3: greument enfermo (I owe the last two references to M. Gorosch's glossary); cf. further V. García de Diego, Manual de dialectología española 250 (Madrid, 1946)—a passage that requires further sifting; J. Borao, Diccionario de voces aragoneses² 242 (Zaragoza, 1908) s.v. greuge, with an unspecified reference to ancient texts (greuge is, of course, Catalan rather than Aragonese). The Rumanian dialect forms include Mac.-Rum. greu, greao, Megl. greu, grεo, Istr.-Rum. grew, grε, according to S. Puşcariu's etymological dictionary.

²⁷ The inflectional pattern of OFr. grever ‘to grieve, to damage’ (later 'to tax') indicates its descent from *greuāre rather than from grauāre, cf. OProv. grevar. Fr. aggraver was preceded by agrever, cf. OProv. agravar, counterbalanced by agrevar; Swiss agreva 'attirer par feinte, amadouer, faire la cour'. FEW 1.53 (1922): ‘Das Wort hat sich in der älteren Sprache meist, aber nicht durchgehend an greuis angelehnt. Im Französischen kommt seit dem 14. Jh. die nach dem Lateinischen retouchierte Form auf.‘

²⁸ From the reflexes of *leuiāre ‘to alleviate, to ease’, assembled by REW³ No. 5002 (OPis. lebbiare, Lomb. libá, Ven. libar, Sors, levǵar, OFr. legier), one should subtract Sp. lijar ‘to lame, to hurt’, which Cejador's vocabulary traces to P. López de Ayala, Aves de caça, and to D. Sánchez de Badajoz, Recopilación en metro; cf. REW³ No. 4842 *laesiāre, from laedō, -ere; No. 4843 laesiō, -ōnis. The homophone lijar ‘to sandpaper’ is related to lija ‘shark, dogfish, or seal skin’, cf. REW³ No. 5081 *līsius (Germanic?). From *leuiāre the verbal adjective *leuius was extracted in Italo-French territory, including Sardinia (REW³ No. 5006); it did not reach Ibero-Romania in sufficient force to survive.

From adleuiāre (in glosses), there branched off formations in Tuscan, Logudorese, Engadinian, French, Occitanian, Catalan, Spanish, and Portuguese: it is a base peculiar to the entire west of the Empire; see REW³ No. 361. *Subleuiāre ‘to ease’ (REW³ No. 8374) has survived in a more narrowly circumscribed zone: witness Sp. soliviar ‘to lift up, to prop up’ and, with a slight adjustment of the prefix, OFr. sozlegier, which, through association with fashionable solaz ‘solace’, became soulager ‘to console’.

The nuclear area of *leuiāre and variants can thus be squarely identified with Gaul, from which powerful prongs extended alternately to Catalonia and Castile, Raetia, and Northern (including North-central) Italy. Small wonder, then, that *greuiāre, supposed to echo *leuiāre, should have originated in Gallo-Romance: cf. OFr. gregier, OProv. greujar. Prov.-Cat. greuje is a postverbal noun.

²⁹ Ernout-Meillet's Dictionnaire s.v. leuis succinctly describes the semantic departure in Imperial Latin of leuō, -āre from its original course. Note that in Ibero-Romance, where leuis suffered severe losses, the verb levar (> Sp. llevar) shows a greater deviation ('to carry away') from the original meaning ('to support [something light]') than do its cognates in French and Italian ('to lift').

³⁰ REW³ No. 279b would benefit from a thorough revision. The entry should be subdivided into *aggrauiāre (> Sp. agraviar ‘to offend’, with the back-formation agravio 'offense') and *aggreuiāre (OFr. agregier, OProv. agreujar, Burg. [] 'to favor, to spoil a child by caresses', see A. Thomas, Notes étymologiques et lexicographiques, Rom. 39.222–3 [1910]; supply the star to this base, listed as recorded by Meyer-Lübke, and delete agregier in No. 3853). An even more radical revision is in order for REW³ No. 4428; the entry should read *ingreuiāre, and to OFr. engregier may be added Fr. rengréger ‘to make an evil worse’, inadvertently listed under No. 3853. *Ingreuicāre 'to weigh(t)' is a peculiarly Rumanian type: Mac.-Rum. angricari, Megl. angrecare 'to load', Rum. îngreca ‘to make pregnant’ (with the sense development of grauidus), and must be sharply separated from OFr. engregier (the confusion goes back to Puşcariu's dictionary). A parallel eastern formation in -ināre is Rum. îngreuna ‘to load’.

³¹ While Sp. gravedad, It. gravità may easily be learnèd, Rum. greutate ‘load’ and OFr. grieté testify to the wide spread of *greuitāte. The entry 3856 in REW³ needs careful reorganization.

³² Contrast It. gravezza, OSp. graveza with Prov. greveza, OArag. greveza (Fueros de Aragón, ed. Tilander, §192.8). On Rum. greață ‘sickness’ (of the stomach), Mac.-Rum. greața ‘weight’, see S. Puşcariu, Etymologisches Wörterbuch der rumänischen Sprache 63. Raeto-Romance makes use of the same derivational device: cf. Eng. greivetsa. Further formations, involving different suffixes, include (REW³ No. 3855) Sic. greviu ‘stupid’, grevia ‘bad humor’; ORum. grecioşa ‘pregnant’. Descendants of greuis, exempt from the pressure of Church Latin, characteristically perpetuate those semantic shades of grauis in which this adjective was near-synonymous with molestus.

³³ Once again, the extraordinary dependence of Ibero-Romance (specifically of Castilian as against Portuguese and Aragonese) on the ecclesiastic tradition is evident: grave held in check the pressure of grieve from the northeast even as the suffix -(i)encia, a perfect replica of Lat. -(i)entia, having overlaid vernacular -iença around the year 1200, hindered -ança from speading to the -er and -ir verbals, as it did in Northern France. See UCPL 1.41–188 (1945).

³⁴ Pleonastically (at least in the historian's perspective), sū;sum was, upon occasion, reinforced by uorsum: cf. the sequence deorsum uersus, in Cato's Agricultura; Ernout-Meillet, Dictionnaire² 1007, opportunely recall G. heutzutage (heul < OHG hiu tagu), Sp. hoy día, pop. Fr. au jour d'aujourd'hui, It. conmeco; add Sp. conmigo, Fr. dedans, Sp. de donde. Sū;rsum and deorsum were paralleled by seorsum beside sorsum 'separate, apart' (both forms occur in Lucretius) < sē + uorsum.

³⁵ The abundant bibliography includes studies by Latinists: E. Seelmann, Die Aussprache des Lateins nach physiologisch-historischen Grundsätzen 330 (Heilbronn, 1885); W. M. Lindsay, The Latin language: An historical account of Latin sounds, stems, and flexions §104 (Oxford, 1894); F. Sommer, Handbuch der lateinischen Laut- und Formenlehre² 557 (Heidelberg, 1914), with the companion-volume Kritische Erläuterungen 36, where dorsum and sorsum are characterized as ‘Allegroformen oder schwachtonige Dubletten’; M. Niedermann's review of Walde's Lateinisches etymologisches Wörterbuch, in Anz. f. indog. Sprach- und Altertumskunde 18.79 (1905–6), with documentation of iū;sum from the Mulomedicina Chironis and the Itinerarium Antonini Piacentini; Roland G. Kent, The sounds of Latin §181.6; Ernout-Meillet, Dictionnaire² 261; Walde-Hofmann, LEW³ 1.342 with references to F. Solmsen and M. Leumann. Romance contributions are equally scattered: F. Diez, Etymologisches Wörterbuch s.v. giuso; H. Schuchardt, Vokalismus des Vulgärlateins 1.434; G. Gröber, Vulgärlateinische Substrate romanischer Wörter, ALLG 2.100 (1885), where the forms with -o- and -u- are not yet adequately distinguished; W. Meyer-Lübke, Die lateinische Sprache in den romanischen Ländern, in G. Gröber's Grundriss der romanischen Philologie² 1.472 (Strassburg, 1904–6), with references to susum CIL 1.199.2 (117 b.c.), rusus ibid. 10.5670.2 (a.d. 107), extrosum CGL 5.287.44; REW³ No. 2567 (where the base, surprisingly, is spelled deōrsum and the Rumanian offshoots are not classified with accuracy); FEW 3.44a (1928), with clean-cut separation of the families of iosum and iū;sum, a wealth of information on Gallo-Romance dialect forms, and some valuable data on Italian—but notice the confusion between Castilian and Portuguese products of iū;sum, and the disregard of the discernible remnants of iosum in the ancient dialects of Aragon, Castile, and Portugal. The old MS readings of deo(r)sum include deosum, diosum, iusum, iosu(m), iosso. There is a similar wavering between ios(s)anus and iusanus; see A. Thomas, Notes lexicographiques sur la plus ancienne traduction latine des œuvres d'Oribase, in Philologie et linguistique: Mélanges offerts à L. Havet 514–5 (Paris, 1909).

³⁶ The picture, in reality, is not nearly so simple as this short list implies. Notice the multifarious reflexes in the branches and sub-branches of Raeto-Romance (sei, sai, se, sε, sœ, su, šu, etc.); see T. Gartner, Die rätoromanischen Mundarten, in Gröber's Grundriss² 1.624, and, almost simultaneously, J. Luzi, Lautlehre der sutselvischen Dialekte §68 (Diss. Zurich, 1904), who thinks of a blend with uia. On compounds peculiar to peninsular and insular Catalan, see L. Spitzer, Lexikalisches aus dem Katalanischen und den übrigen iberoromanischen Sprachen 123 (Geneva, 1921). In Old Italian, suso was sporadically transformed into soso, on the pattern of gioso, which represents the opposite of the standard development gio(so) > giu(so); cf. Dante, Inferno 10.45; see P. E. Guarnerio, Fonologia romanza 278 (Milan, 1918). The REW³ also quotes (from an unstated source) Ptg. sosso as an outgrowth of sū;(r)sum, a derivation which should be ruled out on account of the voiceless s, incompatible with intervocalie VLat. -s-. The Sardinian reflex, omitted by Meyer-Lübke, has, over a period of centuries, invariably been (in)susu [sū;⋅u, insū;⋅u], according to Wagner's Lautlehre.

³⁷ S. Puşcariu, Etymologisches Wörterbuch 80. I have no access to Bogrea's article in DR 4.824, from which REW³ transcribes în al gios ‘upside down’, a face în al gios ‘to treat injustly’, yet I fail to see in these data any clue to Meyer-Lübke's analysis of gios as a product of iū;sum rather than a mere variant of jos. For the phraseology of jos (a pune jos ‘to lay down’, pe jos ‘on foot’, etc.), see H. Tiktin, Rumänisches Elementarbuch 208, and the same writer's dictionary.

³⁸ P. E. Guarnerio, I dialetti odierni di Sassari, della Gallura e della Corsica, in AGI 14.134 (1898); and particularly M. L. Wagner, Historische Lautlehre des Sardischen §135 and §276 (Halle, 1941), with full utilization of such texts as the Condaghe di San Pietro di Silki, Condaghe di Santa Maria di Bonarcado, Condaghe di San Nicolà di Trullas, Condaghe di San Michele di Salvennor, Carte volgari dell'archivio arcivescovile di Cagliari, Carta de Logu de Arborea.

³⁹ G. Rohlfs is noncommittal in his Historische Grammatik der italienischen Sprache und ihrer Mundarten 1.141 §71 (1949). He too uses the inaccurate spelling deōrsum; note his material from Apulia and Calabria in Unteritalienische Beiträge, AR 9.156 (1925). B. Wiese, Altitalienisches Elementarbuch² §30, §34, §43, §63, and §86 (Heidelberg, 1928), consistently treats gioso and apocopated giò as the basic forms. R. R. Bezzola, Abbozzo di una storia dei gallicismi italiani nei primi secoli (Heidelberg, 1925), does not consider the possibility of the influence of OFr. jus (across the interjacent Provençal territory) on It. gio(so); if there has been a measure of influence of iū;sum on OFr. jusarme, Olt. giusarma, of non-Latin provenience, the process began on French soil (183).

⁴⁰ Within Western Raeto-Romance, centered in the upper Rhine valley; cf. the testimony of M. Grisch, Die Mundart von Surmeir (Ober- und Unterhalbstein) 102 (Zurich, 1939), who quotes [] from Sursés, [zaw, , ziw] from Sotsés, with Vaz standing apart: []; Vieli, Deutsch-romanisches Wörterbuch: Surselvisch (Chur, 1944), lists s.v. unten: ‘giudem, giusut, giudapeis’; hier unten ‘cheu giu’; dort unten ‘leu giu’; nach unten ‘engiu’; weiter unten ‘plinengiu’ (744b), as against oben ‘sisum’ (i.e. sū;su summu), ‘sura, sisura’; dort oben ‘leu si’; hier oben ‘cheu si’. East of Sotsés, in Mittelbünden, we find a great diversity of forms, all of them traceable to deosu; cf. C. M. Lutta, Der Dialekt von Bergün und seine Stellung innerhalb der rätoromanischen Mundarten Graubündens §80, §148 (Halle, 1923): dju, djeú, djù, zew, zo, djò, jo. For the continued contrast between sū;sum and deosum farther east in the Engadin, see A. Velleman, Dicziunari scurznieu de la lingua ladina pustüt d'Engadin' Ota 746 (Samaden, 1929): sü e giò ‘auf und nieder’; R. R. Bezzola and R. O. Tönjachen, Deutsch-romanisches Wörterbuch: Ladinisch (Chur, 1944): nach oben 'insü', hier oben 'qui sü, cusü', dort oben 'là sü, lo sü', ganz oben 'sü som, sü dim' (702), as against unten ‘giò, , giodim’ (958); note especially von oben nach unten ‘da süsom fin giosom’. East of the Inn valley, in the adjoining Münster valley, A. Schorta, Lautlehre der Mundart von Müstair (Zurich, 1938), distinguishes neatly between deo(r)su > (§64, §125) ‘hinunter’ and sū;su-summu > (§75) ‘zuoberst’.

Yet in the dwindling Central Raeto-Romance communities, reflexes of deosu and iū;su sometimes do rime; T. Gartner and H. Fezzi, Ladinische Wörter aus den Dolomitentälern 82, 114, 165, 189 (Halle, 1923) quote su ‘oben’ and žu, ‘unten’. The pressure of Lombardic may in large part be responsible for this situation. The contrast between Raeto-Romance and Lombardic is brought out clearly by the record of K. von Ettmayer, Lombardisch-Ladinisches aus Südtirol, RF 13.555–7 (1902): Sulzberg , ; Monsberg zọ, , , Rendena , dzọ beside , ; Bona dẓü, ẓ ^ie; Brescia .

⁴¹ Relics of sus include courir sus, en sus; and dessus with its ramifications. Note also suzerain, which recalls OSp. susero ‘upper’.

⁴² See A. Thomas, Mélanges Havet 514–5, quoting data contributed, in part, by A. Longnon. Iū;sānus, *sū;sānus pertain to the same Vulgar Latin series *dēretrānus, *forānus, *propiānus, *sübtānus, *subterānus, *superānus. At a late date, the name Suzanne was fancifully associated with some of the descendants of *sū;sānus.

⁴³ Here is the breakdown of characteristic forms: suso en la mannana ‘early in the morning’ (Fuero de Teruel §318.4, §698.2); de suso (ibid §6.6, 8; §17.3; §20.2; §32.3; §34.2; §45.4); en suso ‘and more’ (after figures: ibid. §718.2); a iuso (ibid. §338.3); mas a iuso (ibid. §301.3); the same source contains examples of the prepositional use of dius (§354.2, §448.3, §453.3, §494.2, §592.3, §628.3), deius (§430.4), de ius (§427.4, §445.2, §600.2), dius de (§728.3), deius de (§31.3); and of the adverbial use of diuso, without apocope: de iuso (§338.2), diuso (§430.4), de diuso (§335.2, §336.2, §341.3, §344.2). Many more examples have been assembled by G. Tilander, Fueros desconocidos de 1348, RFE 22.126 (1935), and Fueros de Aragón 362–4, with ample documentation of diuso from the Ordinaciones de Zaragoza. Diosmetido beside soçmetido: Fueros de Aragón §261.10, §266.23.

⁴⁴ V. R. B. Oelschläger, A Medieval Spanish Word-List 8b, 9a, 113b (Madison, 1940), has assembled the following instances of deosum forms: a iosa (sic; contamination with iuxtā, as occasionally in Gallo-Romance, see FEW): Caparroso, a.d. 1102, apud T. Muñoz y Romero, Colección de fueros municipales y cartas-pueblas (Madrid, 1847); ioso: Sahagún, a.d. 1166, apud V. Vignau, Glosario y diccionario geográfico (1874); goso: San Juan de la Peña, a.d. 1084, apud Menéndez Pidal, Orígenes del español 53, and Sos, Cartulario, a.d. *1066 (ibid. 349).

Suso was currently used in all medieval peninsular dialects: Cuento de una enperatriz, ch. 12: suso en los çielos; Estoria de Merlin, fol. 293vo: suso en el ayre; contrasted with juso (or yuso) in Estoria del Rrey Guillelme, fol. 36ro: tanto andó por ventura suso e juso, e acá e allá, similarly in Old Riojan: Signos 56c, Milagros 83c. Oelschläger, op.cit. 113b, quotes the older forms and spellings of yuso (Ptg. juso): iuso (Becerro gótico de Cardeña, a.d. 935); iusso (ibid., a.d. 956), guso (San Juan de la Peña, a.d. 1084), juso (ibid., a.d. 1084 and Tudela, a.d. 1127), yuso (Silos, a.d. 1085; San Pedro de Arlanza, a.d. 1135). Note the prepositional use of de ius in Old Riojan (clearly a branch of Old Aragonese): de jus las pennas (San Millán 28b); de yus el techo (Santo Domingo 598c). The occurrence of jus as far west as Burgos (a.d. 1146) is isolated.

There existed numerous compounds: asuso (Leon, a.d. 983; Caparroso, a.d. 1102; Docum. bilingüe de Alfonso VII, a.d. 1143; Toledo, a.d. 1215; Fuero de Teruel (see glossary, with examples of en asuso ‘and more’, after figures; de ... asuso ‘from ... up‘); Libro de Josep ab Arimatía, fol. 257vo) as against atuso, aiusso, ayuso, adius(s)o. Oelschläger quotes examples from Leon, a.d. 951; Santillana, a.d. *967; Leon, a.d. 983; Rioja Alta, a.d.* 1044; Salvador, a.d. 1074; Silos, a.d. 1098; Lerma, a.d. 1148; Burgos, a.d. 1220; cf. Cantar de Mio Cid, lines 354, 446; San Millán 384a (de media noch a iuso); Libro de buen amor 967b, 975a, 990b, 1007b; Barlán e Josaphá, fol. 132vo [var. abaxo]; Libro de Josep ab Arimatía, fols. 257ro, 271vo; Estoria de Merlín, fol. 290ro; Toledo Glossary 2393 [de dos annos ayuso]; Confisión del amante, fols. 28ro, 81ro. Asuso and ayuso gave way to arriba and abaxo in Spanish, to acima and abaixo in Portuguese.

De suso ‘above’ (Libro de buen amor 412d, 472a) was contrasted with de yuso ‘below, down’; for a fitting example of the clash, see the Libro de Josep ab Arimatía, fol. 276ro. Similarly, para suso and para ayuso are symmetrically arranged in the Vida de Santa Catalina, fol. 21ro; para suso also occurs in the Cuento de una enperatriz, ch. 4. Rarer expressions for ‘upward’ include fazia suso (Libro de buen amor 412c) and contra suso (Carlos Maynes, ch. 25). Finally, compare the corresponding adjectives susano (Fuero de Teruel §338.1, 2) and yusano (Juan de Mena, La Coronación, str. 25).

⁴⁵ See A. J. Carnoy, Adjectival nouns in Vulgar Latin and Early Romance, RR 8.172 0017); Meyer-Lübke, REW³ No. 7141 and, earlier, Einführung³ 181. Note that at later stages of the development reddere was weaned from the tutelage of pr(a)e(he)ndere: Sp. rendir ˜ prender. The original state of affairs in Ibero-Romance (prender, render) has been preserved in Portuguese. Notice the coincidental loosening of semantic ties between the two verbs in Ibero-Romance.

⁴⁶ Throughout the history of Catalan, retre has been used extensively in a wealth of meanings. The Diccionari Aguiló 7.117–8 (1931) quotes retre ‘to bring back’ from the Chronica del Rey En Jacme Primer, Tirant lo Blanch, Curial y Guelfa, B. Boades' Libre dels feyts d'armes de Catalunya, the first edition of the Flos Sanctorum, Mossén Guillem Serra's historical compendium of the Bible, the Manual de Novells Ardits, and other sources; reflexive retre-s as a verb of movement (Fr. 'se rendre à') occurs in the same chronicle and in Tirant lo Blanch, besides J. Esteves' Catalan-Latin dictionary of 1489. Notice retre ‘to make, to effect, to change into’ (ordinarily in conjunction with an adjective) in Franc. Eximenis' Dotzén libro del crestiá and Libre de les dones; retre ‘to produce’ in Mallorca; retre les armes in Tirant lo Blanch; retre compte in Flos Sanctorum and the Documentos del Archivo general de la Corona de Aragón. Some of these combinations may have been suggested by French or Provençal, but the form retre seems autochthonous; peculiar is the use of retre as a prefix (confusion with retrō- ?) as in retreguardar ‘to reward, to compensate’ in R. Lull's Blanquerna (Valencia, 1521), fol. 24vo. Rendre (now obsolete; replaced by retornar and retre) was somewhat less widely used in the same literary genres; the Diccionari Aguiló 7.87 offers documentation from the Chronica d'En Jacme, the Vida de St. Antiogo, the 1524 ed. of Flos Sanctorum, and the Ordinari Lleyda of the year 1567. Rendir and arrendir, which may have radiated from the center of the Peninsula, share the meanings of Sp. rendir; for examples from Empordà, Mallorca, and Menorca, see Alcover-Moll, Dicc, català-valencià-balear 2.18 (1935).

⁴⁷ See my monograph in UCPL 1.187–211 (1945) and the reactions of the reviewers: V. Buben, Čas. 31.321 (1948); G. Gougenheim, BSLP 43.2.71–2 (1946); F. Lecoy, Horn. 69.119–20 (1946–7); L. Mourin, RBPhH 26.840–1 (1948); K. S. Roberts. HR 14.85–6 (1946); G. Rohlfs, ASNS 186.185–6 (1949); J. A. van Praag, Mus. 53.159 (1948); and, in particular, H. F. Williams, MLJ 33.157-S (1949).

⁴⁸ See my article La derivación de rebelde, rebeldía y las fuentes del grupo de consonantes -ld- en ibero-románico, in Estudios dedicados a Menéndez Pidal 1.91–124 (1950), and the forthcoming note on peldaño < pedāneu, in AGI.

⁴⁹ The standard expression for ‘every day’ in Old Spanish was cada día (cf. cada mes del año: Castigos e documentos, ed. Gayangos, BAE 51.81b): Martirio de San Lorenzo 90d (ed. Marden); Milagros 220c (ed. Solalinde); Libro de los caballos 12, 25, 27 (ed. Sachs); Don Juan Manuel, Libro de la caça, fol. 211vo (ed. Baist); Conde Lucanor 7 (ed. Knust); P. López de Ayala, Libro de las aves de caça 108 (ed. Gayangos); C. Sánchez de Vercial, Libro de los enxenplos por a.b.c. (ed. Gayangos), BAE 51.504a; A. Martínez de Toledo, Fl Corbacho (ed. Simpson, fols. 4ro, 5ro, 16ro (twice), 28ro, 37ro, 37vo, etc.; Libro de cetrería de Evangelista (ed. Paz y Melia), ZRPh. 1.230 (1S77). Similarly in Old Portuguese, cf. BF 6.444; RFE 23.262 (1936). A variant construction, de cada día (quite infrequently de cada un día: El Corbacho, fol. 35vo), had a vogue ca. 1300–1450: Don Juan Manuel, Libro de la caça, fols. 202vo, 214ro; Ruy Yáñez (?), Poema de Alfonso Onceno 2243a (ed. Janer); prologue to the Dança de la muerte; López de Ayala, Libro de las aves de caça 9; El Corbacho, fols. 2ro, 3ro, 3vo, 6vo, 8vo, 11vo, 13ro, 15vo, etc. Cf. desta vez, dellos ... dellos, discussed in UCPL 1.363–4, 403–8 (1948); and de cada rato ‘every moment’: El Corbacho, fol. 27ro.

The rival form cadaldía (again including the variant de cadaldía, especially in Don Juan Manuel and in Juan Timoneda) has an equally notable record in Old Spanish: Libro de Apolonio 522d (ed. Marden); Purgatorio de San Patricio (ed. Solalinde), ch. 18 [editor's emendation]; Primera crónica general 628b (ed. Menéndez Pidal); Libro de buen amor 345c (ed. Ducamin), MSS T, S (G has cada dia; so has S 147a, 151c, 176b, 228a, 245a, 283a); Don Juan Manuel, Libro de los estados (ed. Gayangos), Part 1, chs. 22, 25, 59, 67, 91, 94, 97, and Part 2, ch. 41; Libro infinido (ed. Gayangos), ch. 2 (three times); Libro del cavallero e del escudero (ed. Gräfenberg), chs. 20, 38; Libro de la caça, fol. 214ro; Poema de Alfonso Onceno 81a; J. Timoneda, El Patrañuelo 11 (ed. Domínguez Bordona). For examples from Lucas Fernández and Gil Vicente, see D. Alonso's note to his edition of the Tragicomedia de Don Duardos 174; excerpts from the Crónica de San Fernando and the Cancionero de Baena are offered by the Dicc. hist. 2.475b-476a; Cejador y Frauca's Vocabulario medieval castellano contains a cryptic reference to a translation of Valerius Maximus. A. Morel-Fatio, L'Espagne au XVI^e et au XVII^e siècle 502b (Heilbronn, 1878) quotes cadaldía from Juan de Coloma's translation of the Trionfo di morte, included in the Cancionero general de obras nuevas (1554). Cadaldía has persisted in Western Asturias (Acevedo and Fernández) and among Moroccan Jews; see J. Benoliel, BRAE 15.47 (1928).

⁵⁰ E. Gessner's cautious attempt to separate cadaldía into cada al dia (Das spanische indefinite Pronomen, ZRPh. 19.157 [1895]) and A. Par's more recent claim that ECat, cadaldia (a.d. 1560, 1687) indubitably continues cada el dia (RFE 16.412 [1929]: 'caso curioso de fonética sintáctica; sin duda, la inserción de l obedece a la construcción cada el dia, cada'l dia?; cf. also H. R. Lang, Contributions to Spanish grammar II: Indefinite pronouns, MLN 2.185 [1887]) are quite infelicitous. The perfect parallelism, semantically and syntactically, of (de) cada día and (de) cadaldía; the observable coalescence of similar groups involving cada (primarily of cadaguno, patterned on alguno, nenguno, and of cadauno, compounded of cada and anno, with derivatives in -ero and, dialectally, in -ego; see my forthcoming study of the suffix -iego); and the notorious unwillingness of speakers, over a period of two thousand years, to combine cada with the definite article or a preposition merging with the definite article (inquiries into Greco-Roman cata, documented since the Glosas silenses, were initiated by P. Meyer, Quisque et cata dans les langues romanes, Rom. 2.80–5 [1873]; cf. notes 45, 241, 315, 336 to my study of alguien, UCPL 1.9.357–442) are telling arguments against Gessner's, Lang's, and Par's not wholly concordant views. The l of cadaldía is epenthetic; cf. OPtg. escoldrinhar 'to scrutinize', from scrū;tinium, Jud.-Sp. doldze < duo decim, and related developments, including in all likelihood peldaño ‘step of a staircase’ < pe(d)año < pedāneu. The fluid character of the construction of OSp. cada is neatly exemplified by the Poema de Fernán González: cad un(o) (319a, 559d) beside cada unos (277d).

⁵¹ If more were known about the infix -al- in Spanish, one could possibly analyze -al-dad as patterned on -al-eza (comunaleza, naturaleza). The Spanish infixes are among the most neglected chapters of historical grammar; on -eg-, see Lg. 25.139–81 (1949); on -ar- preceding -(i)ego, see my forthcoming monograph on the latter suffix; cf. polv-ar-eda, hum-ar-eda, llam-ar-ada (which one finds, upon occasion, inaccurately analyzed llama-rada).

^51a See M. Vey, BSLP 45:2.198 (1949), with a reference to J. Tyxa, Čas. 31.287–91 (1948).

⁵² On the syncope of frīgidus, see E. Richter, Chronologische Phonetik des Französischen bis zum Ende des 8. Jahrhunderts §46 (Halle a/S, 1934); V. Väänänen, Le latin vulgaire des inscriptions pompéiennes 93 (Helsinki, 1937); A. Labhardt, review of Väänänen, ZRPh. 61.357 (1941); P. Fouché, De quelques changements de quantité dans le latin parlé, Mélanges de philologie romane et de littérature médiévale offerts à Ernest Hoepffner 18 (Paris, 1949). Richter and Väänänen interpret frǏda as reflecting [frijida], an opinion impugned by Labhardt, who sums up his own view thus: 'Les dérivés romans (it. freddo, fr. froid etc.) supposent l'évolution frīgidus > frigdus > *, sans doute par analogie avec rǏgǏdus > *rẸgdus.' Labhardt completely disregards the evidence of Ibero-Romance and of Provençal.

Meyer-Lübke, Einführung³ 155, stresses the fact that It. freddo, Fr. froid (OFr. freit, froit) presuppose early syncope, the former through the loss of the post-tonic vowel, the latter through the fall of the final vowel. If the interpretation of OSp. fri(d)o here proposed is tenable, three independent arguments, in addition to Latin testimony, favor the assumption of early syncope.

Rigidus matches frīgidus in regard to syncope only in Italian and in Old French (roit, roide ˜ froit, froide); remains at a considerable distance in Old Provençal (rege ˜ freǧ, frei, freit;) and follows a radically different course in Ibero-Romance (OSp. rezio, OPtg. reijo, OGal. regeo, Ptg. rijo, Gal. rejo besides arrijarse ‘to lift oneself up, to rise’ ˜ OSp. frido, frío; sporadically in archaic texts fredo: Menéndez Pidal, Orígenes del español 270). The problematic -z- of rezio may be due to a blend with reziente ‘fresh, new’; cf. Lg. 23.393 (1947). Coincident with this varying degree of similarity in regard to syncope is the extent of the influence of the stressed vowel of rigidus on frīgidus (or at least the identity of the stressed vowel in the two words): It. freddo again sides with OFr. freit, -de and clashes with OSp.-OPtg. fri(d)o.

⁵³ The normal development is represented by acidu ‘sour’ > Gal. acio, flaccidu > Sp. lacio ‘faded, withered’, īnsipidu > Jud.-Sp. [∫ebdo] ‘tasteless’ (Morocco) beside recomposed * īnsapidu > Ptg. enxabido, Guadal. enjabio (of a child) ‘whining, annoying’ (RDTP 2.139 [1946]), Jud.-Sp. [∫avdo] (Balkans), Arag.-Rioj. jaudo, jauto (P. Bénichou, NRFH 2.265–8 [1948], reckons with substitution of -itu for -idu; yet cf. the Mozarabic and Algarve forms quoted by H. Janner, Apuntes para la toponimia española, Stud. Neoph. 21.94–5 [1948–9]); lū;cidu > OSp. Ptg. luzio ‘glaring’ (Proverbios morales, line 612); murcidu ‘idle’ > Gal. murcio ‘moisture exuded by corned meat’, Ptg. murcho ‘weak’ (REW³ No. 5752); nitidU > Ptg. nédio, Leon. nidio (A. Castro, RFE 9.66 [1922]; REW³ No. 5929) beside the Gallicism neto; pū;tidu > ORioj. pudio (Berceo, Duelo de la Virgen 87a), cf. pudio ‘mata de río’ in Libro de la montería 2.21 and the toponyms Sant. Repudio < rīuu pū;tidu and Ampudia (Palencia) < Fontepúdeda (a.d. 938), also Fuempudia (Borja, Zaragoza) and Ampudia (Burgos, a.n. 1206); cf. Moz. puṭdo in the botanic treatise made available by M. Asín Palacios; P. Sánchez Sevilla, RFE 15.138 (1928), quotes Leon. topon. Fonte Iū;liānī > El Antillen, Fontānellās > Las Antanillas, which show the same change of fonte. Further examples of characteristic development: rancidu > Sp. rancio (Meyer-Lübke's query: learned?); rōscidu ‘wet’ > rocio, rucio (later > rocío 'dew'), see Menéndez Pidal, Etimologías españolas, Rom. 29.369 (1900); squālidu 'dirty' > Sp. escalio ‘land abandoned for tillage’ (REW³ No. 8198); sū;cidus appears in the forms OSp. OLeon. suzio (Fuero de Ledesma §81), OSp. susyo (El Corbacho, fol. 9ro), çusio (ibid., fol. 3ro), yet Ptg. sujo, cf. rijo, murcho, percevejo ‘bedbug’ (if it contains foetidus, as supposed in J. I. Louro's recent note on L. Spitzer's divergent explanation; cf. Algarv. fèto quoted by J. J. Nunes, RL 7.124, and by Janner, loc.cit.); tepidu > Sp. Zam. Ptg. tibio (F. Krüger, Lautgeschichte westspanischer Mundarten 200), OLeon. tebio (the -d- disappeared early enough to have prevented the diphthongization of -Ě-; the change tebio > tibio represents an independent second phase; the assumed variant *tēpidus (Menéndez Pidal, Manual de gramática histórica⁷ §41.2 is unnecessary); turbidu > Sp. turbio ‘muddy.‘

⁵⁴ On frīgidus in Ibero-Romance, see UCPL 1.5.187–211 (1945) and my forthcoming monograph on the suffix -(i)ego. Meyer-Lübke's remark, REW³ No. 3512: ‘Ob sp. frío wirklich auf frīgidus beruht, ist fraglich, da asp. fredo besteht (... Orígenes del español 270), frío aus freo entstanden sein and asp. frido eine literarische Mischung sein kann’, shows the conjectural method at its weakest. The author bases his entire hypothesis on a non-existent variant (freo) of a rare by-form (fredo), calling frido without independent proof a learned form simply to fit it into a ready-made construction. F. Hanssen, Spanische Grammatik auf historischer Grundlage 31 (Halle a/S, 1910), unconvincingly divorced frīgdu > frido from frīgidu > frío.

⁵⁵ The influence of rigidus on frīgidus with respect to the stressed vowel was assumed for Italian, Sardinian, Western and Eastern Raeto-Romance, French, Provençal, and Catalan by Meyer-Lübke, Gröber's Grundriss 1.508 and REW³ No. 3512; more cautiously by C. H. Grandgent, An outline of the phonology and morphology of Old Provençal §31 (Boston, 1905); and without qualification by A. Labhardt, loc. cit., with a reference to Schwan-Behrens; it has become a commonplace of our historical grammars. Bàrtoli thought of contamination by Germ, frisk. O. Bloch and W. von Wartburg's Dictionnaire étymologique lists several hypotheses but remains noncommittal. O. Schultz-Gora, Altprovenzalisches Elementarbuch⁴ §30 (Heidelberg, 1924), soberly traces freit to * frǏgdu.

⁵⁶ The earliest phases of the development fall into the Latin period: note ualidus beside ualdē, and solidus beside *soldus, presupposed by It. soldo, Fr. sou, Sp. sueldo; see Meyer-Lübke, Einführung³ 154, REW³ No. 8069. In close proximity to -lǏd- > -ld- was -rid-> -rd-: lardum < laridum ‘fat of an animal’, *uirdis < uiridis ‘green’ and *uirdia < uíridia < uiridia, posited by Romance reflexes; in the last case, the stress shift may have accelerated, if not entailed, syncope. Cf. E. Seifert, Die Proparoxytona im Galloromanischen 71–80 (Halle a/S, 1923).

⁵⁷ The early syncope of calidus had two perceptible effects: one in Latin, the homonymic clash with callidus ‘callous, sly’, which prompted speakers to discard that word; the other in Ibero-Romance: structurally isolated caldo, without any connection with the series of adjectives in -io < -idu, was substantivized, and caliente or caluroso (from calor, calura 'warmth, heat') was used instead for 'hot'. Characteristically, OSp. rabdo (Primera crónica general 714b) > raudo ‘swift, impetuous’, likewise separated from its series, has also become obsolete, while the derivative raudal ‘torrent, stream; plenty’ (the last meaning under the concomitant influence of caudal?) continues current. One way of maintaining the adjectival character of a word subject to syncope was the addition of an appropriate suffix; cf. the products of rauidu + ānu > OPtg. roudão, Sp. roano ‘sorrel’: Menéndez Pidal, Rom. 29.367 (1900); REW³ No. 7100. The early coinage of the derivative, the disappearance of the primitive, and the loss of -d- after the monophthongization of -ou- < -au-are all interlocking processes. Cuervo, Corominas (Problemas por resolver, AILC 1.176 [1941–2]), and Bénichou (NRFH 2.266 [1948]) assume influence of rapere.

⁵⁸ After the perfect balance between caldu and frī(g)du had temporarily been attained through polarization, the two adjectives developed along divergent lines in individual parts of the Empire: the -d- was more resistant in the group ld- than between vowels.

⁵⁹ Observe that in the gen. sing. alone, in contrast to all other cases, a consonant cluster follows upon the stressed vowel in both words.

⁶⁰ On the Sardinian groups a sa. di ˜ a su notte (outside of this combination notte is invariably feminine) and on the Old French sequences tote di, tote jorn, mie di (persistence of fem. diēs in Northern Gaul is unlikely), see W. Meyer[-Lübke], Die Schicksale des lateinischen Neutrums im Romanischen 11–2 (Halle a/S, 1883). Rum. zi, Camp. di, both restricted to isolated areas, either continue fem. diēs or have belatedly and independently become feminine under the pressure of descendants of nocte. A. Tobler's interpretation of toute di < tōtum ad diem in ZRPh. 2.628 (1878) elicited a friendly comment from Meyer-Lübke in 1883, yet was rejected on good grounds in REW³ No. 2632. These formations, selected at random from Sardinian, Old French, and German, have one trait in common: the change of gender occurs within stereotyped adverbial groups treated as inseparable lexical units, with the article (if used at all) divested of its primary functions.

⁶¹ Meyer-Lübke, Grammatik der romanischen Sprachen 2.380 (Leipzig, 1894), REW³ No. 5349; cf. J. E. Matzke, MLN 2.334–5 (1887), with a reference to H. Sachs's Göttingen thesis (inaccessible to me), Geschlechtswechsel im Französischen (Frankfurt a/O, 1886).

⁶² Cf. the similar pressure of Tusc. il fiore on the feminine descendants of flōre in Northern Italy, whose effects are described in K. Jaberg's masterful essay, Aspects géographiques du langage 55–7 (Paris, 1936).

⁶³ This view of the separate status of Northern France and its ties with Germany has lately been developed with particular forcefulness by von Wartburg.

⁶⁴ Among the various explanations of the gender of Fr. mer, surely few are so devoid of substance as that of J. Damourette and E. Pichon in Des mots à la pensée: Essai de grammaire de la langue française 1.371–2 (1927): ‘La mer est d'aspect changeant comme une femme, journalière, d'humeur mobile comme une jolie capricieuse, attirante et dangereuse comme une beauté perfide. Le citadin qui lui consacre ses vacances est amoureux d'elle; elle est l'amante et la meurtrière du marin ...‘ Most of these associations are the consequence of, not the cause of the feminine gender; the neuter gender of G. Meer, Russ. more, Lat. mare creates a different imagery. Nor were mer and mère homophones at the time when mare changed its gender in Gallo-Roman. The only useful statement by Damourette and Pichon is that iū;mentu > (la) jument represents the lone parallel to mare > la mer; there are many more instances in Spanish (la sal, la miel, la hiel).

⁶⁵ It would be fruitless, then, to try to apply a generally valid label to the lexical character of Rumanian. An archaic strain prevails in the development of nurus and socrus, while the evolution of mare is independent, no doubt because in pastoral societies the terminology of kinship is upheld with tenacity, while the sea and its accessories are outside the purview of socio-linguistic traditions.

⁶⁶ Meyer-Lübke, Grammatik 2.380 and REW³ No. 9134.

⁶⁷ Cf. the semantic disintegration of Alpine words for ‘mountain rock’ and ‘grass-covered lane around a rock’ (Jaberg, op.cit. 47–0).

⁶⁸ Lomb. valeğ(a) ‘wild torrent’ (REW³ No. 9133) < uallicula; OIt. vallone ‘bay’, Abruzz. vallone̥ 'brook in a valley,' Occit. vala(t), valon 'rivulet' (REW³ No. 9134).

⁶⁹ Examples from medieval texts are offered by Meyer-Lübke, Die Schicksale des lateinischen Neutrums 12. H. Gröhler, Über Ursprung und Bedeutung der französischen Ortsnamen 2.100–4 (Heidelberg, 1933), lists numerous relics of the old feminine gender: La Vau (Vienne), Lavan (Yonne, Loire Inférieure), Lavault (Allier), Bonneval (Eure-et-Loir, Haute-Loire, Savoie, Drôme), Bonnevaux (Doubs, Haute-Savoie, Vienne), Valbonne (Alpes-Maritimes, Gard), Bellevaux (Haute-Savoie), Valbelle (Basses-Alpes), Sommeval (Aube), Vaucluse (Hautes-Alpes, Doubs, Vaucluse), Vaujuraine < ualle iū;sāna (Aube), Francheval (Ardennes), Sourdeval (Calvados, Manche), Fréteval < frācta ualle (Loir-et-Cher), Plainevaux (Liège); note the prevalence of the south, southeast, and east over the west, north, and center. A. Vincent, Toponymie de la France 207–10 (Brussels, 1937), offers additional documentation of Benoitevaux, Valbenoite, Bonneval, Valbonnette, Droiteval, Valfanjouse, Froidevaux, Vaugrigneuse, Longueval, Malleval, Vaunoire, Parfondeval, Sècheval, Torteval, beside numerous toponyms involving masculine val; vall is invariably feminine in Grisons; see A. Kübier, Die romanischen und deutschen Örtlichkeitsnamen des Kantons Graubünden 210–2 (Heidelberg, 1926).

⁷⁰ Jaberg, op.cit. 57–63, makes a thorough inquiry into the competition between the intruder testa and native capo in Northern and Central Italy (note the appended linguistic map); G. Rohlfs, Das Fortleben der vierten lateinischen Deklination in Italien, ASNS 177.93–7 (1940), reviewed by A. Lângfors, Rom. 69.128, and by W. von Wartburg, ZRPh. 62.148–9 (1942).

^70a An interesting attempt to account for changes of gender in French through the assumption of synonymic pressure (OFr. aé, eded < aetāte influenced by eage; esme modeled after vis; œuvre partially attracted by mestier, travail) is made by J. Orr, On homonymics, Studies presented to M. K. Pope 291–2 (Manchester, 1939).

⁷¹ Meyer-Lübke's cryptic entry (REW³ No. 7315) is less helpful than E. Gamillscheg's discussion in Romania Germanica 1.226, 375–6 (Berlin-Leipzig, 1934): Prov. ric, Cat. rich, Sp.-Ptg. rico are attributed to Goth, reiks; It. ricco to Lang. rîhhi; Fr. riche (and, inconsistently, Prov. ric) to Frk. rîki; all three kindred types apparently overlay ricus as used in Vulgar Latin by Germanic legionaries. The frequently expressed view that Sp. rico may be of Gallo-Romance background is disproved by the final vowel (contrast rico with OSp. duc, duque or Enrique). On the sense development of riche, see G. Gougenheim, RPh. 3.270–2 (1949–50).

⁷² See REW³ No. 6307 (Meyer-Lübke fails to qualify pobredad as distinctly medieval). OFr. povérte reflects the straight case paupertās. OProv. paupriera, paubreira (REW³ No. 6306) < pauperie, *pauperia (cf. facie ˜ Late Lat. facia): the near-synonym miseria probably did yeoman service; cf. OSp. lazeria, OPtg. lazeira ‘misery, wretchedness’ < lazerar, laz(d)rar (the by-form lazerio presupposes attraction by cat-, cautiverio).

⁷³ Rictad: Cantar de Mio Cid, lines 688, 1399; Berceo, Milagros 158d; ritad: Cantar de Mio Cid, lines 1189, 1245. Menéndez Pidal, in his etymological vocabulary, quotes Ausias March: Si com rictat no porta bens ab si, and refers the reader to OFr. richetet (modeled on povretet).

⁷⁴ This influence is, for instance, suggested by the use in the Cantar de Mio Cid of the two variants riquiza (lines 481, 1269, 2663) and riqueza (lines 811, 1200, 1648; see pp. 242, 831 of the standard edition), and by the general intervention of Gallo-Romance in the formation of Old Spanish abstracts; see UCPL 1.4.41–187 (1945) for data on -ança.

⁷⁵ Pobredad (pobredat) clearly prevailed down to 1400 and held its own in the 15th century: Vida de Santo Domingo 322c, 446b (ed. Fitzgerald); Milagros 655d (ed. Solalinde); Sacrificio de la misa 224a (ed. Solalinde); Martirio de San Lorenzo 64d (ed. Marden); Calila e Digna (ed. Allen), ch. 5, lines 274, 278, 290; Primera crónica general 269b (ed. Menéndez Pidal); Libro del cavallero Zifar 8, 24, 27, 124, 328, 337 (ed. Wagner); Libro de buen amor (ed. Ducamin), p. 5 and quatr. 727a, 1384d; Biblia medieval romanceada, MS Esc. I-j-8 (ed. Battistessa), Deut. 15.7; Vida e costumbres de los viejos filósofos (ed. Knust) 79, 159, 259, 383; Rimado de palacio N 132b, 196d, 984c, 1114b, 1133a (ed. Kuersteiner, checked by Zeitlin); E. de Villena, Arte cisoria 95 (ed. Navarro); A. Álvarez de Villasandino, Cancionero de Baena 60 (ed. Pidal-Ochoa); Marqués de Santillana, Proverbios, NBAE 19.458a; Frey Íñigo de Mendoza, Vida de Cristo, ibid. 11a, 17a. Probedat: Cancionero de Baena 315.

Pobreza: Fueros de Aragón §100.2 (ed. Tilander); Las siete partidas 1.315, 2.35, 461 (ed. 1807); Primera crónica general 36b, 122a; General estoria 1.584a (ed. Solalinde); Flores de filosofía 13 (ed. Knust); Proverbios morales (ed. González Llubera), lines 824, 1060; Vida e costumbres de los viejos filósofos 187, 225, 259, 341; A. Álvarez de Villasandino, NBAE 22.343b; Frey Íñigo de Mendoza, NBAE 19.15b. Probeza (in late Old Spanish also proveza): A. Álvarez de Villasandino, NBAE 22.349a; D. Hurtado de Mendoza, Cartas 127 (ed. Rose and Vázquez), distinct from proveza ‘provencia’ in Roncesvalles, line 65, see Menéndez Pidal, RFE 4.121–2 (1917).

⁷⁶ See my study of Spanish abstracts, NM 46.171–91, 47.13–45 (1945–6).

⁷⁷ Including Alonso Garrote and Lamano y Beneite (the latter also offers an alternate explanation). Covarrubias grasped the connection between sombra, sombrero, sombrío, asombro(s), and asombrar.

⁷⁸ F. Diez, Etymologisches Wörterbuch 535 (Bonn, 1853). In revisions, Diez added OPtg. soombra, Sp. sombrero, OLeon. solombrero.

⁷⁹ Sombre, which appears only in late Middle French, has puzzled etymologists; cf. Bloch and v. Wartburg, Dict, étym.² (Paris, 1949).

⁸⁰ H. Rönsch, Romanische Etymologien 2, ZRPh. 3.104 (1879): New Testament MS of Verona (4th or 5th cent.), Marc. 9.7 nubs subumbrāns ; Old Testament Cod. Ottobonianus (8th cent.), Ex. 25.20 subumbrantēs, corresponding to συσκιάζουτες.

⁸¹ The Spanish Academy Dictionary, in its latest editions, followed suit (through the efforts of J. Alemany Bolufer?).

⁸² J. Leite de Vasconcelos, Estudos de filologia mirandesa 2.217 (Lisbon, 1900–1); J. J. Nunes, Crestomatía arcaica² 599 (Lisbon. 1921).

⁸³ The proposal of ON sumbla as a base by H. Wedgwood, French etymologies, Rom. 8.439 (1879), failed to satisfy A. Scheler; see his supplement to Diez's Etymologisches Wörterbuch⁵ 777. Note Bloch and v. Wartburg's restraint.

⁸⁴ Solombra: Libro de Alexandre (ed. Willis) O 816c (= P 844c: sonbra), 892a (= P 921a: sonbra), 1179c (= P 1208c: sonbra), 1817d (= P 1958d: sonbra); cf. E. Gessner, Das Altleonesische: Ein Beitrag zur Kenntnis des Altspanischen 34 (Berlin, 1867); A. Morel-Fatio, Recherches sur le texte et les sources du Libro de Alexandre, Rom. 4.50 (1874); Poema de Alfonso Onceno 1683d (Janer: 1680d) ‘fancy’, 2266d (Janer: 2263d) ‘shelter’, see J. P. ten Cate, Poema de Alfonso Onceno: estudio preliminar y vocabulario 110 (Amsterdam, 1942); Ferrara Bible, Jud. 9.15, see L. Wiener, The Ferrara Bible, MLN 11.102 (1896); Constantinople Bible (printed 1873), Gen. 19.8, Jud. 9.15, Is. 4.6, Job 40.22; see M. Gaspar Remiro, Vocablos y frases del judeo-español (segunda serie), BRAE 5.361 (1918); J. Alemany [Bolufer], Voces de Maragatería y de otra procedencia usadas en La esfinge maragata, BRAE 3.61 (1916); S. Alonso Garrote, El dialecto vulgar leonés hablado en Maragatería y tierra de Astorga² 323 (Madrid, 1947); B. Acevedo y Huelves and M. Fernández y Fernández, Vocabulario del bable de occidente 203 (Madrid, 1932), beside sombreiro, somòriya, possibly-imported, cf. also sombrero in El Bierzo, hardly autochthonous (V. García Rey, Vocabulario del Bierzo 146). The dialects of Segovia, Aragón, Álava, and Murcia, on the basis of available word lists, show no trace of solombra. Finally, note the villancico in Gil Vicente's play Comedia do viuvo: Arrimárame a ti, rosa, / no me diste solombra, probably taken from a folk song. Significantly, in conservative Central Asturias solombra not only, as a substantive, corresponds to ‘shade’, but is also equivalent to the adverbial groups ‘bajo la sombra, a la sombra’, which tends to strengthen the argument in favor of its division into so l'ombra; see M. J. Canellada, El bable de Cabranes 340 (Madrid, 1944).

Solombría, solombrío, solumbrío ‘shady place’ are listed and in part illustrated by J. de Lamano y Beneite, El dialecto vulgar salmantino 634–5 (Salamanca, 1915); cf. Maestro Gonzalo Correas, Vocabulario de refranes 214: Ni trigo de valle, ni leña de solombrío lo vendas a tu amigo; ibid. 396: Por San Matías cantan las cotovías y entra el sol por las solombrías. Solombroso: Ferrara Bible, Is. 18.1 (Wiener, MLN 11.102); Constantinople Bible, Is. 18.1, Ezek. 31.3, Job 40.21, 22 (Gaspar Remiro, BRAE 5.361). Note that sonbroso in MS P of Alexandre 348a seems an error for sabroso; see J. Keller, Contribución al vocabulario del Poema de Alixandre 172 (Madrid, 1932). Solonbrero: Alexandre O 245d (= P 252d: sonbrero); solombrera: ibid. 1713c (= 1854c: sonbriella). Ptg. sorumbático ‘sad, gloomy’, associated by M. L. Wagner with solombra, RFE 12.81–2 (1925), is independent, according to J. Corominas. Garcilaso in his Third Eclogue used umbroso beside sombrío ‘shady’ (lines 73, 82); Galician favors umbrío (Saco Arce, Gramática gallega 291). Umbroso and umbrío easily represent the original derivatives from umbra.

⁸⁵ M. L. Wagner, Observaciones sobre el judeo-español de Oriente, RFE 10.241 (1923); idem, Caracteres generales del judeo-español de Oriente 98 (Madrid, 1930); J. Leite de Vasconcelos, Estudos de filologia mirandesa 2.217; S. Alonso Garrote, op. cit. 318–20; Crónica troyana, ed. A. Martínez Salazar, see Nunes, Crestomatía arcaica² 118; Ferrara Bible, Dan. 4.9 (Wiener, MLN 10.39 [1896]); Constantinople Bible, Dan. 4.9 (Gaspar Remiro, BRAE 3.507 [1916]). M. C. Casado Lobato reports that selombra, selumbreira (note the wavering between the back vowels) are in use in La Cabrera Alta (§7).

⁸⁶ See HR 14.130–7 (1946) and RPh. 1.209–34 (1947–8) for illustration and bibliography.

⁸⁷ Cf. J. Brüch, ZRPh. 41.757 (1921); M. L. Wagner, loc.cit.; G. Rohlfs, ASNS 187.167 (1950). Brüch's earlier assumption of a cross between sub umbra and sublū;stris ‘dim, half dark’, ZRPh. 40.652–3 (1919–20), is infelicitous. Meyer-Lübke's attitude lacks precision: REW³ No. 8405 s.v. subumbrāre (unjustifiably starred) lists OFr. sombre, Iber.-Rom. sombra, Vegl. sombreya, Eng. sumbriva; No. 9046 records Fr. sombre, Prov. sozombrar, solombrar, OSp. solombra, Leon, selombra as compounds of umbra. Some overlapping is visible, particularly in French; apparently Meyer-Lübke was intent on separating sombra from solombra. The transfer of sozombrar, which had served Diez to infer subumbrāre, to the entry umbra causes surprise: sozombrar continues subumbrâre in a direct line, with soz- replacing so- even as des- encroached on de-.

⁸⁸ L. Wiener, op.cit., MLN 11.102 (1896).

⁸⁹ See J. Corominas' brilliant argument against W. D. Elcock's attempt to identify the s- of Arag. sobago ‘shady place, slope open to the north’ (Venasque) < opācu with the definite article (based on ipse in the Tyrrhenian area), in Los nombres de la lagartija y del lagarto en los Pirineos, RFH 5.6–7 (1943). Corominas contrasts Ven. sobago with Arag. obago, paco, Cat. obac, and assumes influence of sol on both sombra and sobago. A problem not yet touched upon is the possibility of an ancient bridge between WSp. solombra and Occ. solombrar (see J. Ronjat, Grammaire historique 3.438); cf. a similar configuration of areas in the case of trīcārī, UCPL 1.7.244–7, 258–60, 283 (1947), where I have belatedly discovered an intermediate link in Asturian. Ronjat himself formulated the problem differently in his review of REW¹ (RLaR 59.136 [1916–7]) and in his Grammaire historique 3.438, 453 (1937): like Dauzat (RPhF 28.156 [1914]) he was much impressed by subumbrâre, but he was also under the spell of an early note by J. Haust, who traced Occ. souloumbra to sous l'ombre. Ronjat's oscillation between the two hypotheses was intensified by the discovery of asoumbra < adumbrāre beside Aq. aloumbra ‘ombrager, mettre à l'ombre’, which seemed to suggest the late insertion of the definite article.

⁹⁰ Bloch and von Wartburg, Dictionnaire étymologique² 568b: ‘Comp. encore l'anc. prov. solombrar “mettre à l'ombre,” issu du même verbe [subumbrāre] sous l'influence de sōl “soleil”.‘ The return to pre-Diezian etymologies is characteristic of our times, cf. Phil. Quart. 29.151–71 (1950); it does not imply the rejection of the effort of four generations of scholars, but shows the progressive liberation of our minds from some suggestive but not invariably correct hypotheses formulated by distinguished pioneer comparativists whose direct influence is less strongly felt at present than it was decades ago. Post-Renaissance etymologies, if revived, require a rigorous new formulation which usually changes them beyond easy recognition.

⁹¹ G. A. García-Lomas y García-Lomas, Estudio del dialecto popular montañés 323 (San Sebastián, 1922); Lamano y Beneite, op.cit. 634.

⁹² On hībernum (tempus) in Ibero-Romance, see Lg. 25.172–3 (1949). Add vierno from the Old Leonese Fuero de Ledesma §34.

⁹³ Enverano is attested by Wagner, Caracteres generales 77. Examples of the flexible use of en- in Old Spanish—dēbile > endeble ‘weak’, *īnsalsu (for īnsulsu) > soso ‘tasteless‘—are provided by P. Bénichou, NRFH 2.265 (1948). In the case of these adjectives, co-existing verbs in en- had their distinct share in facilitating the addition and subtraction of the prefix.

⁹⁴ For some special applications of these principles, see my forthcoming studies on the etymology of Sp. piara and manada ‘herd’, on the derivation of mentira(s) ‘lies’ and veras ‘truth’ and on the history and prehistory of Sp. pihuela, tro(m)pezar, and maña.

⁹⁵ A much mooter question is: What guided speakers in distributing the active and the passive role between the two partners of the process? Only in a few cases can a tentative answer be provided with any degree of certainty. The prevalence of dexter over sinister seems to parallel the stronger development, hence greater usefulness, of the right arm and hand. The predominance of sū;(r)sum over deo(r)sum, iosum may go back to the observable rhythm of the movements of the human body: arms and legs are usually first raised, then lowered; the basic motion is the lifting, while the letting down is associated with drooping and relaxation. In other instances, social organization may be involved: it is inconceivable, in the light of our knowledge of the family structure among the original speakers of Indo-European, that, all other conditions being equal, the word for ‘daughter-in-law’ should have exerted influence on the word for ‘mother-in-law’: this would have implied complete reversal of the accepted hierarchy. The analysis of other contrastive pairs does not yield equally satisfactory objective results, although the explorer, more than once, is tempted to provide an intuitive answer. Any statement had better be delayed until after the collection of richer and more adequately spread material. In particular, it may be fruitful, at some later date, to study the point in conjunction with the broader problem of the preferred sequence of words in formulaic groups like land and sea, io and fro, the rights and wrongs.