general, however, I might state that in the Cuentos there is no clear line between the two forms ande and onde in usage. Donde and adonde are also frequently used even in texts that abound in dialect forms. As a general rule it might be stated that ande stands more frequently for adonde but is also used for donde, while onde usually represents donde but is also frequently used for adonde. As a matter of fact in Old and Classic Spanish the line between donde and adonde was not well defined22 and the clear distinction between donde expressing rest and adonde expressing motion is quite modern.

Ande seems to be found in all parts of Spain judging from the numerous examples in the Cuentos. I give only a few: Santander 63 (‘¿Ande anda tu madre?'); Burgos 225 ('Y del pozo ande estaba enterrado. . .'), 404, 414, 510; Soria 101, 249, 250; León 420; Palencia 253; Zaragoza 156, 322; Toledo 44, 45, 71, 91; Cuenca 204; Ciudad Real 199, 259; Córdoba 42, 43, 472; Sevilla 422, 476; Granada 46, 48, 113, 145, 212, 244.

There are examples from Santander, Burgos, Palencia, Soria, Toledo, Cuenca, Zaragoza, Granada, Sevilla, Córdoba and León, from all parts of Old and New Castile, from La Mancha, Estremadura and Andalucía. It is also a form found in Salamanca (Lamano y Beneite, s.v.) and Murcia (Alberto Sevilla, s.v.). It is therefore fundamentally and originally an old and modern Castilian phenomenon and not only Andalusian and American-Spanish as Menéndez Pidal seems to believe.23 It is certainly not from Toledo in origin as Mugica states.24 The form ande is also found in American-Spanish, for example in Argentina,25 Puerto Rico,26 and New Mexico.27

The form onde is not as common as ande but it is also found in the

Cuentos from all parts of Spain: Santander 343 ('. . y les dió de beber en la misma jarra onde estaba el dinero.'), 367, 372; Burgos 224 ('Y onde quiera que vía balcones. . . .'), 497; Valladolid 403, 436; Soria 57; Segovia 302; Avila 87; Astorga 393, 395; Zamora 181; Zaragoza 156; Cáceres 210; Cuenca 111, 142, 187, 205; Granada 46, 53, 81, 215, 240; Córdoba 469; Sevilla 298, 317.

Onde = donde, adonde is also found today in Salamanca (Lamano

22 Cuervo, Diccionario.

23 Gramática histórica española 70 note 1, Madrid, 1925.

24 ZRPh 30. 351.

25 Martin Fierro 375, (ed. Tiscornia, Buenos Aires, 1925).

26 Teófilo Marxuach 61.

27 Studies in New Mexican Spanish II § 85.

y Beneite,), is also recorded for La Montaña by García-Lomas (I have given examples from Santander) and is a well known phenomenon in American-Spanish.28

The question of the origin of the modern dialectic forms ande and onde is quite complicated. In Old and Classic Spanish there are a variety of forms meaning 'where,' 'whence,' 'whither': obe, o <ŭbi; do<de-ubi; onde <unde; donde, dond, donde unde; adonde, adond, adón <ad-de-unde. In Old Spanish all these forms were used and with little or no distinction in meanings. At first practically all the forms could mean 'where,' 'whence' or 'whither' and could be used with any preposition that expressed rest or motion. That gave rise to many reduplications of prepositions especially to express the idea of separation or the of or from case.29 Latin unde came to mean in Vulgar Latin 'where,' 'wherefore,' 'therefore. '30 In Old Spanish onde derived from it has all these meanings. Cuervo in his Diccionario has given us numerous examples. In the Crónica General, for example, onde is frequent with the meanings donde, de donde, adonde and also por lo cual. In the XIIIth century, however, the most frequent meaning of onde seems to be de donde. The idea of separation was soon forgotten and a new form, donde, was made, but we cannot say whether this new form is as old as onde or a later formation.31 In Classic Spanish onde was still commonly used with the old meanings, although the meanings 'wherefore' and 'therefore' appear to be more general.*2

The modern dialectic onde, therefore, is probably an archaism, the regular Old Spanish onde<Latin unde. As a matter of fact, however, it could also be explained phonetically as a development of the form donde with apheresis of initial d, a common Spanish phenomenon. Ande may also be explained as an archaism, although we have no examples of it in literature, derived from a+onde <ad-unde, or it may represent a-donde <ad-de-unde. In either case, assuming that intervocalic d

28 In New Mexico (Studies in New Mexican Spanish I § 32 and II § 83); in Mexico (Marden, The Phonology of the Spanish Dialect of Mexico City 15, Baltimore 1896); in Columbia (Cuervo, Apuntaciones 752); in Chile (Ramón A. Laval, Folklore de Carahue 70, 88, Madrid, 1916); in Puerto Rico (Marxuach 87).

29 Cuervo, Diccionario, s.v. donde.

30 Grandgent, Vulgar Latin §§ 73, 84.

31 Dond, don are the only forms found in the Cid. In other early monuments, however, ond and onde are found. In the Reyes Magos (XIIth century) we find dond de donde.

32 Acad. 15, s.v. onde merely states that it is an archaism equivalent to por lo cual, por cuya razón, en donde, de donde.

falls in adonde, we have first a shift of accent to the more sonorous vowel and then do>a by regressive assimilation as in the case of áo>a in aunque (aonque)> anque discussed later, and ae>a in adelante> aelante >alante.

It is interesting to note that in the Cuentos the new formations de ande, por ande, de onde, are also commonly used. When the preposition has once been joined to another word its primary meaning is easily lost and a new preposition is required to express the necessary meaning. The case of de ande is especially noteworthy.33

4. an guar de, an eguar de

Each case of these interesting linguistic phenomena occurs only once in the Cuentos. Both cases are from Córdoba: 165 ('Y otro día por la mañana fué a vel la vela y en la mesa vido que an guar de la vela staba una piesna de un muerto.'); 346 ('Y cuando ya la agüela se había ido va la chica y saca mantequilla der pucherete y se unta debajo e lo sobaco y an eguar de decí, Sin Dios ni Santa María, de villa en villa, dijo:-Con Dios y Santa María de viga en viga').

=

The source is evidently en lugar de en vez de, both phrases being quite common in literary and popular Spanish since early times.34 Of the above two forms from Córdoba an guar de is the older form. A still older form is en guar de which I find in the Corvacho of Alfonso Martínez de Toledo (XVth century prose):35 'Ay sin ventura de mí. non ove yo ventura como mi vezina, que en guar de medrar desmedro; en guar de fazerme paños nueuos empeñasteme estos captiuos que en la boda me distes, .' (121-2); ‘E por aquí se pierden ynfinidos e muchos que en guar de conoscer senorío e otorgar mejoría. . .' (140). For the XVth century I have also a case of guar de from the Cancionero de Montoro3 (a native of Córdoba) that Dr. Karl Pietsch has called to my attention, but in view of the fact that the previous verse ends in

33 Astorga 328: 'Y ya el hijo del rey no pudo resistir más y salió de ande estaba escondido y le declaró su amor.' Burgos 413: 'Y a otro día cogió Juan y llevó las cabras a pacer a un prao que estaba cerca de la huerta de ande vivían unos frailes.'

34 One of the oldest examples in my notes is the following from the Documentos linguísticos de España of Menéndez Pidal, Madrid, 1919, p. 76 (a Castilian document of the year 1225): e si otra bestia ouiere hi qe uala mas qe el cauallo,

qe la tomedes en logar del cauallo; . . .'

35 Madrid, 1901.

36 Cancionero de Antón de Montoro, ed. Emilio Cotarelo y Mori 175–6, Madrid,

1900.

an accented syllable I believe that we have here a simple, ordinary case of compensación entre versos37 and that the faulty text should be corrected to en guar de:

'Si le tocan de la espuela

con gran priesa y turbación,

<en> guar de salir con cautela,

da los dientes a la ación.'

Aside from the above cases for Córdoba I know of no examples of the phenomena in question for the modern Spanish dialects except the New Mexican en gual de (Studies in New Mexican Spanish I § 214).

The development was probably as follows: in locale de>Old Spanish en logar de (cf. French en lieu de) > en lugar de. Old Spanish logar became lugar very early although the reason for the change of o to u is not clear. There is also an Old Spanish form logal a more primitive form from locale. En lugar de through metathesis becomes *en gular de, and then the pretonic vowel u although tending to disappear, following the old rule in Spanish, remains as consonantal u after the fall of intervocalic 1, the result being *en gular de>*en gu(l)ar de>en guar de, the form found in the XVth century Spanish of Corvacho. The fall of the intervocalic 7 in this case is not without parallel in Spanish phonology. I shall cite now only one case with a similar development, namely, eguar <egualar (modern Spanish igualar) <Latin aequalāre, a form found in Berceo, Milagros 67, c. The Córdoba an guar de is exactly the same phenomenon that we find in Corvacho except for an instead of en, an opening of the vowel e to a before a nasal +consonant, a very common linguistic change in Spanish dialectology.38 The New Mexican Spanish en gual de is also the XVth century phenomenon but with final instead of r. It could be explained as a more archaic development with a final 7 preserved from the Old Spanish logal, en logal de> en lugal de>*en gulal de>*en gu(l)al de>en gual de. As a matter of fact such a development would also explain the tendency of the intervocalic to fall through dissimilation. Since Old Spanish oá>tends to become uá (cf. Joan>Juan), especially in popular speech,39 it is not even necessary to assume that logal or logar went through the

37 See my two articles La sinalefa entre versos en la versificación española (The Romanic Review 16. 103–121) and La compensación entre versos en la versificación española (Ibid., 306-29).

38 Studies in New Mexican Spanish I § 23.

39 Ibid. § 85. The phenomenon is common in Vulgar Latin.

stages lugal or lugar in the development of en guar de, en gual de, for after the fall of intervocalic l *en go(l)ar de would regularly become *en goar de en guar de.

The second example from Córdoba, an eguar de with epenthetic e between n and g is not easy to explain, but the presence of the epenthetic vowel between a nasal and a velar is not entirely different from the numerous cases between a liquid and a velar that are found in the Spanish dialects.40 It would also be possible to explain the epenthetic vowel through analogy with the phrase igual de < Old Spanish egual de, but I have no cases of en igual de = en guar de or en lugar de either from literary or dialectic sources.

5. anque, onque, enque

.

These forms of the literary Spanish aunque, adverb and conjunction, < Latin ad-huc (cf. Aragonese adú)+que with final n in aun from analogy to bien, non, sin, are very rare in the Cuentos, although according to some of the authorities cited below at least anque is quite common in the Spanish dialects. I have only the following examples: anque, Granada 52 ('Y er de la rueda e molino coge la botella y anque la bruja había partido ante sale é como un rayo .'), Córdoba 458 ('Ese sapo me vi a pillá y anque sea sapo me lo vi a comé.'); onque, Granada 334 ('Y dijo:-Unos caballos como éstos los llevaría yo al palacio onque me costaran un ojo de la cara. Y cuando vido la carroza que estaba también convertida en alabastro dijo:-Una carroza como ésta la llevaría yo al palacio onque me costara un ojo de la cara.'); enque, Palencia 253 (-No se corte usté nada que enque sea con trabajo yo le llevaré.').

The form anque is recorded for Salamanca by Lamano y Beneite, for Galicia by Cuverio Piñol,41 for Asturias by Munthe,42 and further for Andalucía by Machado y Alvarez.43 In American Spanish it is found in New Mexico, Mexico, Chile, Puerto Rico and Columbia.44 Onque and enque are also found in New Mexico. In New Mexico there is also a form unque.

The forms anque, onque are quite commonly used in XVIth century

40 Ibid., § 193.

41 Diccionario gallego, Barcelona, 1876, s.v.

42 Anteckningar om formålet i en trakt af vestra Asturien 58, Upsala, 1887.

43 Studies in New Mexican Spanish I, § 34 and notes.

44 Ibid.; also II § 89; Cuervo, Apuntaciones § 764, and Marxuach 61.