¹ Charles Bally, Traité de stylistique française 1.23 (Heidelberg, 1909).

² Jules Breitmeyer, Le suffixe latin -ïvus; Genève, 1933.

³ Diet. Gén., Traité de la formation de la langue française 39.

⁴ Wilhelm Meyer-Lübke, Italienische Grammatik 289 (Leipzig, 1890).

⁵ The development in Greek has been outlined by E. Fraenkel, Beitraege zur Geschichte der Adjektiva auf -τικόs, Zeitschrift für vergleichende Sprachforschung 45.205–24 (1913), whose explanation has been subsequently endorsed by Albert Debrunner, Griechische Wortbildungslehre 197–200 (Heidelberg, 1917). Ever since recourse was had to the suffix, it served to express relation, but, unlike -ivus, chiefly that of an inanimate object to a person, e.g. Tρωïκós Trojan' (Homer); παιδικόs ‘belonging to a child’ (preclassical), μαντικόs ‘characteristic of a soothsayer’ (classical). A notable increase in the use of the element paralleled the growing predilection for complicated definitions among the Sophists. From them, it was taken over by Athenian society and ultimately adopted by Plato and Aristotle. By that time, the suffix, in addition to its traditional neuter signification, began to develop an active one. Thus, νθουσιαστικάs, aside from ‘being in raptures’, came to mean ‘enrapturing’; κπληκτικόs meant ‘amazed’ and ‘amazing’, κινητικόs ‘movable’ and ‘motive’, *παθητικόs ‘sensitive’ and ‘arousing a feeling’, φυλακτικόs ‘cautious’ and ‘preserving’, δνψητικόs thirsty' and ‘making thirst’, ὐπνωτικόs ‘drowsy’ and ‘producing sleep’, άφροδισιαστικόs ‘given up to love’ and ‘exciting love’, πιστ∈υτικόs ‘confident’ and ‘awakening confidence’. The transition to the active connotation may have been achieved through the instrumentality of such ambiguous adjectives as ὐγι∈ιvόs ‘healthy’, as shown by ὐγι∈ιvòv σῶμα ‘body being healthy’ beside ὐγι∈ιvόs τόπο* ‘place making healthy’. It should be observed that the substantivized forms κριτική (τχνη) ‘criticism’, μουσική (τχνη) ‘music’, ρητορική (τχνη) ‘rhetoric’ split off before the turning point in the syntactic development of -ικόs, in contradistinction to what happened in the case of -ivum, -iva.

⁶ Fortunately, we are in a position to recognize each single advance and retreat in this slow progress. Gellius, who lived in the 2d century, goes along the new path when in need of grammatical terms: connexivus ‘tending to connect’, disiunctivus ‘expressive of disjoining’, frequentativus ‘denoting repetition’, plurativus ‘designating the plural’, 'privativus 'causing privation', vocativus ‘used in calling’. He clings to the old function when hampered by an antonym: infestivus ‘not agreeable’ after festivus ‘pleasant, gladsome’. Apuleius seeks to reconcile both shadings in his well-balanced style. He conforms to the new rule in forming declarativus ‘elucidative’, dedicativus ‘affirmative’, illativus ‘conclusive’, negativus ‘negative’, while he is tolerant of substitutivus ‘subordinate’ and suffers passivus to retain its neuter connotation. On the evidence of comminativus ‘threatening’ and confirmativas ‘confirmative’, Tertullian seems to adhere to the modernists, but for all that he does not discontinue to employ words built up after the old pattern : substantivas ‘substantial’, inerrativus ‘unutterable’, putativus ‘supposititious’. It is only in the writings of the physician Caelius Aurelianus that the factitive function has been generalized: corporativus ‘strengthening’, districtivus ‘relaxing’, mitigativus ‘soothing’, mordicativus ‘stinging’, evaporativus ‘producing evaporation’, interfectivus ‘causing death’, resumptivus ‘fit to restore’, vexativus ‘making suffer’. See Frederic T. Cooper, Word-Formation in the Vulgar Sermo Plebeius 105–9 (New York, 1895).

⁷ No translation is given of these philosophic terms, as they would require a lengthy paraphrasis; but they are linguistically quite transparent.

⁸ The following abbreviations are used in this article: Braz. = Brazilian; Camp. = Campidanese; Cors. = Corsican; E = English; Fr. = French; Gall. = Gallegan; It. = Italian; Leon. = Leonese; Log. = Logudorian; OFr. = Old French; OPr. = Old Provençal; Ptg. = Portuguese; Rhaet. = Rhaeto-Romance; Rum. = Rumanian; Sard. = Sardinian; Sp. = Spanish.

Dialectal material has been taken from Antonio Tiraboschi, Vocabolario dei dialetti bergamaschi² 1351 (Bergamo, 1873); L. Carré Alvarellos, Diccionario Galego-Castelán (A Cruña, 1933); Luis Carlos de Moraes, Vocabulário Sul-Rio-Grandense (Pôrto Alegre, 1935); Giacomo Schaad, Terminologia Rurale di Val Bregaglia 30 (Belinzona, 1936); C. Martin Lutta, Der Dialekt von Berguen 191–2 (Halle, 1923: ZRPh. Beiheft 71); Das neue Testament, erste rhaetoromanische Übersetzung von Jakob Bifrun 1560, ed. Th. Gartner (Dresden, 1913); Max Leopold Wagner, Das ländliche Leben Sardiniens im Spiegel der Sprache (Heidelberg, 1921: Wörter und Sachen Beiheft IV); F. D. Falcucci, Vocabolario dei dialetti della Corsica (Cagliari, 1915); Giorge Pascu, Sufixele Româneşti 221 (Bucureşti, 1916); Edward L. Adams, Word-Formation in Provençal 322–3 (New York, 1913); José de Lamano y Beneite, El dialecto vulgar salmantino (Salamanca, 1915), and others.

⁹ Gall, vadío ‘vagabond’ had its d preserved through the influence of the words in -adío. Leon, cansío ‘weary’ developed from the now antiquated participial form canso. Rhaet. descendente of fugitivu had their initial consonant changed into another labial : musdif, muštadif. Ptg. sadio ‘healthy’ includes an active shading like its counterparts in other languages. Apart from *laborat-ivu, note *labor-ivu as implied by the Corsican place-name Casalabriva (beside Casalta, Casablanca, Casanova, Casamozza). Camp, trigađiu, Log. regađiu ‘backward in ripening’ may postulate the base *tricativu, with -ativu added to a nominal stem, a formation made possible by the incorrect analysis of *laborat-ivu as *laborativu.

¹⁰ That manantío must have originally extended also to Castilian is illustrated with Sp. manantial ‘brook’. Laschantiv is attested as early as 1560. In Corsica, the representative of *vacantivu applies to the uncultivated part of the vineyard, whereas in Sardinia two separate forms have developed, one relating to the field, the other to men and beasts.

¹¹ The phonetic restriction, however, does not apply to Gallo-Romance to the same extent as elsewhere. Provençal, for instance, apart from agradiu ‘agreeable’, braidiu 'eager, cridiu 'tearful', faidiu ‘banished’, regardiu ‘waiting’, all of them characterized by their d, contains many more formations in which phonetic conditions no longer have any evidential value: adomniu ‘dominating’, adoniu ‘generous’, cargiu ‘onorous’, celiu ‘hidden’, penhoriu ‘pledgeable’, remembriu ‘mindful’. This, in addition to other semantic and syntactic criteria (such as abstract meaning and partly active function) would strengthen the assumption of a semi-learned origin of the whole set.

¹² G. Schaad 30.

¹³ In a similar way, *sub umbriva > sumbregva ‘shadow’.

¹⁴ OFr. jolif ‘merry’, haïf ‘hateful’ and other derivatives from Germanic stems have been quoted as evidence of the vitality of -ivu as late as the 5th century. In almargio the suffix is added to an Arabic stem; this takes us back to the 8th century at the earliest.

¹⁵ Other examples are frötif ‘fructiferous’, gerif ‘icy’, grassif ‘fat’, guastif ‘fallow’, magrif ‘meager’, montif ‘mountainous’, ortif ‘full of gardens’, morunif ‘planted with mulberries’, pascolif ‘consisting of pastures’, pratif ‘covered with meadows’, ripif ‘on the slant’, ronchif ‘with vineyards on the hill’, sabiuní ‘sandy’, salegif ‘planted with willows’, sgrösif ‘dry’, sulif ‘sunny’, zerbif ‘untilled’.

¹⁶ Vestiges of this nuance can be found in Portuguese also, e.g. alvadio ‘whitish’.

¹⁷ In Leonese, contamination of -íο with -(er)ía was conducive to the formation of the suffix-chain -erío: ascu-erío, bras-erío ‘heap of red-hot coals’, astill-erío ‘heap of splinters’, αν-erío ‘poultry’.

¹⁸ Of particular significance are the Ibero-Romance examples. They show that the theory advanced by Meyer-Lübke, It. Gr. 289, that this variety of -io goes back to -erium, is untenable. His contention rested on the comparison of It. lavorio ‘work’, diavolio 'devilry with Milanese lavoreri, diavoleri.

¹⁹ Señorío may have interfered with the shift of stress suffered by albedrío < arbitrium ‘free agency’.

²⁰ No major significance should be attached to the occasional alternation between -io and -ia. This latter variety, derived from the plural of the neuter, gives an even stronger prominence to the collective shading. The starting point in Latin is sativa ‘tilled land’ as against silvestria ‘woodland’ (Pliny). Note the indiscriminate use, in Portuguese, of baixia and baixio ‘shallow’; in Leonese, of solombría and solombrío ‘shadow’, manantía and manantío ‘brook’. Compare Fr. lessive ‘lye’ and Rhaet. glischiva, altschiva ‘washing’ to OPr. leisiu ‘lye’ < lixiva, lixivum. Interesting formations in -iva are also Log. nađia children, progeny' < nativa and bolađia ‘dust soaring up when corn is winnowed’ < *volativa.

²¹ The Italian formations alluded to are, for instance, addormentativo ‘producing sleep’, ammaestrativo ‘instructive’, appiccativo ‘contagious’, impacciativo ‘embarrassing’, infreddativo ‘refreshing’, nettativo ‘purgative’, proseguitivo ‘persecuting’, and the like; see Breitmeyer 342.

²² Cf. sua autem voluntas est efectiva rerum ‘his will brings about everything’ (Thomas Aquinas), almost synonymous with sua voluntas efficit res.

²³ Jointiz > jointif ‘joint’, faitiz > faitif (1) ‘produced’, (2) ‘serving to produce’, hostil > hostif ‘hostile’, soutil > soutif ‘subtle’, orphelin > orphelif ‘orphan’; in Middle French, subrepticiu > subreptif ‘clandestine’, obrepticiu > obreptif ‘obtained by craft’, authenticu > autentif ‘authentic’, erraticu > erratif ‘erratic’.

²⁴ The material presented in the following paragraphs is based on the dictionaries of Godefroy and Tobler-Lommatzsch, for Old French; of Huguet, for French in the 16th century; of Bloch - von Wartburg and Hatzfeld-Darmesteter-Thomas, for Modern French; of Tommaseo-Bellini, for Old and Modern Italian. The NED has been consulted for English.

²⁵ Though Jehan de Meung is credited with having been the first to introduce scholastic terms in -if, -ive into French, the termination as such existed prior to him in substantival derivatives, e.g. ois(d)ive ‘idleness’, in all probability a contamination of oisif ‘idle’ and bois(d)ie ‘mischievousness’ (Wace, Marie de France, G. de Coinci, Rendus de Molliens). Though lacking a clearly circumscribed function, the ending was known sufficiently well to infringe upon the domain of -ie, as in estoutive instead of estoutie ‘stupidity’ (Benoît, Ducs de Normandie); in a similar way, Jehan Petit (1360–1411) used maladive in place of maladie ‘illness’.

²⁶ English also is rich in plural formations, e.g. disjunctives ‘disconnected things’ (1624–77), incentives ‘beginnings’ (1728–34), perceptives ‘perceptive faculties’ (Herbert Spencer), in keeping with the numerous words in -ics; compare It. le trattative ‘negotiations’, Fr. les préparatifs ‘preparations’, les figuratifs ‘figurative aspects’ (Pascal).

²⁷ In Middle French, -if and -ive repeatedly served as stylistic resources. Christine de Pisan is reputed to have introduced words ending in this learned suffix to imitate the ‘style clergial’ and to express ‘choses estranges en plus grand soubtilite’; her enthusiasm for monastic Latinity was sincere and genuine. As for Villon and Rabelais, there are reasons enough to suspect both writers of having intended ridicule of the scholastic tradition they are known to have abhorred; for them, -if and -ive were symbolic of the atmosphere of the Sorbonne.

²⁸ Les préparatifs does not count because of its plural form.

²⁹ Jacques Damourette and Édouard Pichón, Des mots à la pensée 1.354–423 (Paris, 1930).

³⁰ One such original formation is ahorrativa ‘economy’ (Polo de Medina).

³¹ For instance, el destrueque ‘la remise’, el desquite ‘la revanche’, el desmogue ‘la mue’, el destemple ‘la dissonance’, el descargue ‘la décharge’, el desgaje ‘la déchirure’.

³² Compare un encuentro ‘une rencontre’, un anuncio ‘une annonce’, and many others.

³³ Another instance of the same phenomenon is seen in -ica, even more closely related to -ική. At the outset, it was used elliptically: (ars) musica; cf. Einar Loefstedt, Syntactica, vol. 2, chapter 10: Ellipse (Lund, 1933). At a later date, it became independent enough to compete with -ia, cf. Fr. diplomatie beside diplomatique (subsequently differentiated), théorie beside théorique, pédagogie beside pédagogique. In the cases of dialectique, éthique, métaphysique, clinique, esthétique, the adjective followed on the noun. Even more reminiscent of -iva is -oria: here, the noun omitted in clerical style seems to have been littera ‘letter’. Thus, in the individual Romance languages we come across descendants of executoria ‘executive’, comminatoria ‘comminatory’, declaratoria ‘declaratory’, compulsoria ‘compulsory’, dimissoria ‘dimissory’, confessoria ‘confessing’, dedicatoria ‘dedicatory’, eliminatoria ‘eliminating’, used as ‘execution’, ‘menace’, ‘declaration’, ‘compulsion’, etc. Note It. deploratoria ‘elegy’ (F. Belcari, 15th century), Sp. escapatoria ‘escapade’, with its humoristic effect (presented as the sentence of a judge). In Middle French, the gender was fluctuating: la revocatoire ‘revocation’ beside le robatoire 'theft.

³⁴ I am profoundly indebted for criticism of this paper to Professor Leo Spitzer, who set forth his own views on -ivus in Vox Romanica 2.204. Professors Eugen Lerch, Urban T. Holmes Jr., and Mario A. Pei gave me valuable suggestions. Dr. Marion Carroll and Miss Adele Wolkin kindly polished up the style. I understand that Professor Lerch has recently published an article on ‘Das Motiv’ in Studia Neophilologica. To my regret, this study is so far not available here in New York.