¹ Meillet and Ernout, on the contrary, reject the haplologic explanation and suppose a direct formation after perīculōsus, ventōsus. If this be true, the explanation given in the text (*caritātōsus — χαριτ + -ōsus) would suffice. F. J. Cooper, Word Formation in the Roman sermo plebeius 122, points out that the -ōsus derivatives are most frequent in scriptores rerum rusticarum, e.g. in Cato.

² Gillis P:son Wetter, Charis 14 ff. (Leipzig, 1913), points out that in Philo and in Greek documents of the Roman emperors xάριs (also the plural χάριτ∈s does not mean a favorable disposition of mind, but an act of favor, benefit (e.g. χάριτ∈s is coupled with ∈ὐ∈ργ∈σίaι); he translates St. Paul's δικαιοὐμ∈νοι δΩρ∈àv τ αὐτο χάριτι διά τ άπολντρώσ∈Ωs τ v Xριστ 'Iησο as 'alle werden umsonst gerechtfertigt durch seine Gnade (ntat) vermöge der Erlösung in Jesus Christus'—this Pauline conception of the χάριs being not a Jewish but a Greek word-use in accordance with the conception of the sovereign's grace as opposed to the Jewish idea of λ∈οs 'the pity of God': 'χάριs drückt ein positives Gefühl von Liebe, Wohlwollen aus, das sich in einer wirklichen Tat Gottes, in einer Handlung äussert. ... Es ist ein Akt der königlichen Gnade Gottes, dass er uns in diesen Stand der Gerechtigkeit versetzt.' This active 'grace' is akin to cāritās, whereas the passive enduement of man with the divine grace was translated by grātia.

³ He comments (51–2): ‘Tertullian verwendet z.B. für den Begriff der christlichen Liebe das Wort agape. Die bekannten lateinischen Ausdrücke schienen ihm allzu profan zu sein [note : De oratione 28 (CSEL 20,199). De baptismo 9 (CSEL 20,208). Ad Martyres 2 (Oehler 18)]. Das Wort hat sich auch für die Liebesmahlzeiten der Christen eingebürgert und in diesem Sinne länger erhalten. Für die Gottes- und Nächstenliebe aber setzte sich agape nicht durch. Man bevorzugte das Wort caritas, das in der liturgischen Sprache ganz allgemein wurde [note: Vgl. I. Kor. 13 nach der Vulgata!]. Im späteren Mittelalter verlor das Wort amor seinen heidnischen Klang und wurde auch dort verwendet, wo man tausend Jahre vorher nur caritas hätte leiden mögen [note: Vgl. z. B. Imitatio Christi, lib. III, cap. 5].‘

⁴ Perhaps a contamination of cāritās with Greek words derived from the same root, like χαρίζομαι ‘dono, gratificor, munero’; χαριστήριον ‘donarium, donativum, donum’ (cf. CGIL), could explain an OCat. caristiós ‘caro, costoso’ (aquest temple de Deu en la ciutat de Hierusalem axi sumptuós y αxί caristiós y bell), Mod. Prov. caristie 'amour, amitié', and Mod.Fr. caristade 'aumône', FEW s.v. caritas, no. 2 (this last surely of southern origin)—unless we prefer to see in all of them carestía ‘dearth’, which von Wartburg suggests for the latter form. On carestía cf. Lang. 14.144–5.

⁵ Cf. the parallelism of semantic development of Fr. charité (caristade) and Prov. pi(a)tansa ‘alms’ (REW s.v. pietas), and of pietās used of clerics (pietas nostra says an abbot; see Du Cange s.v. pietas) and Sp. caridad used in the same way (St. Theresa, cf. Roman, Chilenismos 1.267), as in St. Augustine's charitas vestra (Chr. Mohrmann, Die altchristliche Sondersprache in den Sermones des hl. Augustin 1.53).

⁶ I do not see how humilde could be an extraction from humildad, since -ad is not a productive suffix: one finds no *ruinde from ruindade, ruindoso. Humilde must be a post-verbal adjective (with -e like Sp. alegre, firme, libre, ruín) from a verb *humildar = humilitāre (sl late verb attested in Amm. Marc, and Augustine), perhaps surviving in OSp. omildança (Menéndez Pidal, Cantar de Myo Cid I and II); cf. ondrar : ondrança (ibid.). The explanation of the replacement of the Latin adjective humilis by a semi-participle must be the same as in the case of amargo in place of *amaro, from amargar = *amāricāre; cf. Stilstudien 1.77. Ptg. impidoso cannot come from *impedit-ōsus, the perfect participle + -ōsus (which would still be haplological: *imped[id]oso), but is formed directly from the verb (impedir), like Sp. resbaloso, guardoso (Hanssen, loc.cit.), Ptg. abduroso, It. rincrescioso, Rum. lunecos, and Lat. bibōsus, censured by Nigidius in Gellius (Cooper, loc.cit. 124). Ptg. soidoso is not better explained by the substantive soidom + -ōsus: if we do not start from sōlitās : *sālitōsus (after *caritōsus), the model ambitiōsus (ambitiō,-ōnis) seems closer.

The fact that Spanish, too, has such formations—piadoso, habildoso; bondoso beside bondadoso (as Mr. Allen states); caridoso (see above); maldoso beside maldadoso; amistoso (borrowed from OProv. amistos = *amlcitās + -ōsus); vanidoso (corresponding to Ptg. vaidoso) and dioso ‘de muchos días’ (originally idoso = *aet[āt]-ōsus); humildoso (attested, in spite of Mr. Allen's denial, by Hanssen, Gram. hist. 344; cf. Oudin: humildoso ‘plein d'humilité‘)—shows that very much the same development must have taken place in Spanish as in Portuguese, only that in more recent times the type cuidadoso, bondadoso seems to prevail, perhaps because of the intervention of grammarians. (However, while caridoso, humildoso, maldoso were dropped, in these cases along with the variants in -adoso, the forms piadoso, vanidoso, amistoso, habilitoso persist.) Thus, quantitatively speaking, the type is only slightly less well represented in Spanish than in Portuguese.