¹ Only 10 examples in a count, made by me, of occurrences of the vocative in some 280 pages of miscellaneous texts, both prose and verse. In the remaining instances (79.6%) the vocative was like the nominative, although theoretically the ending can be attached to any noun or adjective. See G. Weigand, Praktische Grammatik der rumänischen Sprache², 65–7 (Leipzig, 1918), for the best available analysis of the use and distribution of the various voc. forms.

² H. Tiktin, Rumänisches Elementarbuch 86 (Heidelberg, 1905).

³ W. Meyer-Lübke, Grammaire des langues romanes 2.12 (Paris, 1890–1900).

⁴ E. Bourciez, Éléments de linguistique romane³ 579 (Paris, 1930).

⁵ In practice doamne is usually restricted to addresses to the Deity, while domnule is used for human beings. Dumnezeu < domine deus is often nominative, and has also a voc. Dumnezeule; in any case it is probably a learned word.

⁶ I confess that it might be difficult to attach -e—though not -ule—to a word already ending in -e (Lat. 3d decl.), such as frate ‘brother’. A few masc. nouns in -ăn and -e are declined like feminines and may have a voc. in -o; thus popă ‘priest’, voc. popo; bade (term of respect used in addressing an older person, esp. an elder brother), voc. sometimes badeo.

⁷ T. Gartner, Darstellung der rumänischen Sprache 150 (Halle, 1904); cf. O. Densuşianu, Histoire de la langue roumaine 1.244 (Paris, 1901): ‘Peut-être faut-il plutôt supposer que le vocatif roumain reproduit d'un côté le vocatif latin, de l'autre côté le vocatif slave; la terminaison slave se serait superposée sur celle qu'on avait héritée du latin.‘

⁸ So e.g. P. Axelrad, Elements of Roumanian, New York, 1919; J. A. Candrea-Hecht, Cours complet de grammaire roumaine, Paris, 1900; M. Gaster, Chrestomatie română, 1.cxiv (Leipsic & Bucharest, 1891); cf. Bourciez 579, ‘cette désinence [-e] qui peut se placer après l'article (domnule!).‘

⁹ Op. cit. 86–7; cf. also his statement in Gröber's Grundriss 1.451.

¹⁰ See Weigand's account referred to in note 1 above.

¹¹ R. I. Paul, Flexiunea nominală internă in limba română 163–4 (Bucharest, 1932).

¹² F. von Miklosich, in Sitzungsber. d. phil.-hist. Classe der kais. Akad. der Wissenschaften (Vienna), 99.72 (published 1882).

¹³ Loc. cit.; cf. Meyer-Lübke in Zeitschr. f. rom. Phil. 19.479.

¹⁴ See Densuşianu 244.

¹⁵ On the definite article with a vocative, cf. the very common South German Guten Tag, der Herr and the like. For Romance examples see F. Diez, Grammatik der romanischen Sprachen⁵ 778 (Bonn, 1882).

¹⁶ Distinctly so stated in the more reliable grammars, e.g. Tiktin 87. (Candrea-Hecht gives -o only for soră ‘sister’ and proper names.) Yet Meyer-Lübke uses caso ‘O house’ as an illustration, Gram. des langues romanes 2.12, and Tiktin himself gives it in Gröber's Grundriss 1.452. See also note 1 above.

¹⁷ Rum. Elementarbuch 87, cf. 39. Likewise Gartner 150, ‘unstreitig slawisch’; Densuşianu 244; Meyer-Lübke 2.12: ‘L'influence slave est certaine pour le vocatif des féminins qui est en -o: caso; car, en ce cas, le latin ne nous est absolument d'aucun secours’; Bourciez 578, ‘un vocatif en -o qui vient du slave’; A. Rosetti, Istoria limbii române 3.66 (Bucharest, 1940).

¹⁸ Thus Gartner is wrong in saying (131), ‘ein lautlicher übergang von -o in ă lässt sich . . . nicht belegen’. Tiktin 44, however, says that this change occurs ‘im Inlaut altbulg. Lehnwörter’; Densuşianu 273 says that in some words ‘après l'accent et après une labiale o a passé à a’. But they do not go far enough. Weigand has observed ‘dass ă der regelmässige Vertreter von bulg. o im Auslaut ist’, in Krit. Jahresbericht über d. Fortschritte der roman. Philologie 12 (1909), p. I.97; so also Rosetti 46–7.

¹⁹ Cf. also alcătuì ‘put together’ < Hung. alkotni.

²⁰ By Tiktin 87.

²¹ In this connection one should cite lesne ‘easy, easily’ < OBulg. lъsno, adv. Gartner 132 thinks the -e is from the analogy of Slavic and Roumanian adverbs in -e.

²² P. 87.

²³ Soră, which violates the rule, is an old word, to be sure, but in a new form—Lat. 3d decl., not 1st.

²⁴ This state of affairs was pointed out 60 years ago by Miklošič, op. cit. 60; but it has been ignored by all subsequent writers.

²⁵ See G. Nandriş, Sur la postposition du pronom personnel en roumain, Bulletin linguistique (Univ. of Bucharest) 5.200–3 (1937). Examples are given of the repetition of the object (both before and after the verb), with a proposed explanation.

²⁶ Slavia 16.5.

²⁷ Op.cit. 188.

²⁸ Op.cit. 60, with examples 72. Miklošič cites a form addo as imperative of adduce ‘bring’ (alongside addu), and derives it from addu o; I have found no other evidence that such a form exists.