SBORNÍK PRACÍ FILOZOFICKÉ FAKULTY BRNĚNSKÉ UNIVERZITY STUDIA MINORA FACULTATIS PHILOSOPHICAE UNIVERSITATIS BRUNENSIS N 12, 2007 LUCIE PULTROVÁ The LaTin adjecTives wiTh The suffix -idus The Latin adjectives with the sufix -idus belong to the most pronounced Latin adjective types. The type is relatively productive and considerably – although not absolutely – semantically homogeneous. In almost every diachronically oriented Latin grammar book, this sufix is particularly accentuated as a representative of the sufix belonging to the so-called “set of sufixes” (Sufixverband). This important linguistic term denotes a group of sufixes by which the nouns and adjectives, or verbs, are derived from one root of speciic type; these derivations then form a group of related words within which we cannot distinguish between the primary and secondary derivation (that is which words are derived from which within the given sufix family). In Indo-European linguistics, holding a prominent place among the “sets of sufixes” or “systems of sufixes” is the so-called “Caland system of sufixes”,1 comprising the sufixes *-ro-/*-mo-, stative verbs in *-eh1– (> -ē-), -es-neuters, the comparatives in *-Ôos- and the superlatives in *-is-to-. In Latin, such a prominent set of sufixes is formed – from the synchronic point of view – by the mentioned adjectival sufix -idus with the substantive -or and the verbal -ēre or -ēscere, e.g. calidus – calor – calēre – -calēscere; tepidus – tepor – tepēre – -tepēscere etc.2 1 2 The term “Caland system of sufixes” is rather a kind of abbreviation in the “jargon” of IndoEuropeanists – Holland linguist Willem Caland (1859-1932) himself is, from our today’s perspective, only remotely associated with this issue. The law he formulated and through which he made his mark in the history of linguistics (so-called Caland’s law), consists in that the adj. in -ro- in Avesta in the position of the irst member of compounds are substituted by the adj. in -i-, e.g. the compound kh3vi-dru- to the adj. khrūra-. Later, J. Wackernagel acknowledged this phenomenon to be generally Indo-European (that is why it is often called also Caland-Wackernagel’s law) and only after various scholars had gradually researched the issue, the awareness of the semantically interconnected set of sufixes developed (cf. e.g. Collinge 1996: 23ff.). Among Latin adjectives in -idus, the adjectives of this particular type (i.e. with related stative verbs in in -ēre and the abstract nouns in -or) are absolutely dominant; however, it is not the only type: on the one hand, these sets are not always complete, either the verb is missing (e.g. lepidus), or the abstract noun is not formed (e.g. āridus), and on the other hand there are other marginal types, e.g. the derivations from the verbs of other types (cupiō – cupidus) or the 88 LUCIE PULTROVÁ Despite the unquestionable importance of the sufix -idus in Latin word-formative system and in spite of many attempts made to explain its origin, the linguists have so far failed to accomplish the task. The basic problem is that Latin sufix -idus has no clear unequivocal equivalent among other IE languages and that the adjectives derived thereby are, at irst sight, quite unique in the IE linguistic area. Let us start our interpretation with an outline of the up-to-now presented opinions on the origin of the sufix -idus and on the formation of these adjectives: 1) The easiest thing to do is simply to presume that Latin sufix -dus is a direct successor of the PIE sufix *-do- or *-dho-. Trying to answer the question whether *-do-, or *-dho-, the linguists usually lean on the presumed related forms recorded in other Italic languages, which are, however, ambiguous: Umbrian KALEŘUF / calersu3 would testify rather to *-d-, Oscan Callifae4 , on the other hand, sooner to *-dh- (cf. e.g. Brugmann 1906: 471f.; Nussbaum 1999: 381f.). Benveniste (1935: 144) regards the sufix -idus to be related to the – neither suficiently explained – gerundival sufix -e/ondo-. In his opinion, the gerundival sufix is formed parallelly to the sufix of active participles in -e/ont-; i.e. he construes the element -e/on- either with -t-, or with -do-, and this very -do- is according to Benveniste identical with the -do- in our sufix -idus. The sufix -do(according to Benveniste *-dho-) indicates a state. The weak point of this whole interpretation is evident and it has already been mentioned: we can see, apart from isolated forms, no equivalent adjectives in other IE languages, and we cannot postulate a PIE sufix on the basis of a single language, or a single language branch. 2) Already in 1878, Osthoff (1878: 121ff.) came with a completely different solution, cited as possible by most of the later authors researching in the topic (e.g. also Brugmann 1906: 472). In his opinion, the adjectives in question are actually compounds (cali-dus), the irst element of which is formed by the appropriate content verbal abstract noun (according to Osthoff namely -es-abstract noun, converting in composition to the o-declination); the second element is then also verbal, that is nomen agentis from the root *deh3- (Lat. dare) or *dheh1- (in Latin only in compounds, e.g. condere). The adjectives formed in this way then would be equivalent to the other Latin compound adjectives, ending in -icus: horri-icus (to facere) ~ horri-dus (to dare, “give”, or *dare, “lay”). Balles has quite recently (2003) come with quite the same solution, citing the root *dheh1- (“lay”) in the base. 3 4 evident denominatives (herba – herbidus) – for an exhaustive list of singular types see e.g. Nussbaum (1999). Tabulae Iguvinae Ia 20, resp. VIb 19; this is assumed to be the equivalent to Latin calidus, “with a white spot on the forehead” (ISID. orig. 12,1,52). Local name known only from the citation in LIV. 8,25,4; it is usually associated with Lat. calidus, “warm” (i.e. possibly “a place with warm springs”). THE LATIN ADJECTIVES WITH THE SUFFIx -IdUs 89 3) Another hypothesis of the origin of the sufix -idus stems from the fact that it seems to have a semantic relation with the sufixes belonging to the so-called “Caland system of sufixes” (see above) – as if it stood for the sufix -ro- in Latin system. This hypothesis was researched in detail by Bloch (1954), who, having made comparison with the Greek adjectives in -r ov~ and Indo-Iranian in -ra-, comes to the solution that in regard to sence they are comparable; but, on the other hand, that only few Latin adjectives in -idus have direct equivalents among the IE adjectives in -ro- (e.g. crūdus – OInd. krūrá-, madidus – ma da r ov~; in other cases we ind only a semantic relation, e.g. adjectives frigidus – y uc r ov~, pallidus – wjc r ov~, puttidus – s a p r ov~ and other, see Bloch 1954: 24f.). From the view of phonology, the hypothesis that Lat. -dus < PIE *-ros, is only partly acceptable: the alternation testifying to phonetic closeness of r and d exists in Latin (e.g. *medi-diēs > meridiēs) and the dissimilation r – r > r – d (*crūrus > crūdus) is easy to imagine, nevertheless, two simple and weighty objections can be raised: First, only a minority of the adjectives with the sufix -dus has in the root the phoneme r, which could effect the dissimilation. Secondly, there is a number of Latin words where the sequence of two rs is maintained while no dissimilation occurs (prōsperus, properus etc.). 4) Quite recently, another interesting hypothesis was put forth by Danish researcher B. A. Olsen (1994), based on two premises: irst, the adjectives of the type calidus stand by the verbs that do not form the participles in -tus; secondly, the cases are well recorded where an occlusive in contact with the preceding laryngeal *h1 or *h2 yields an aspirate (e.g. Greek p l hquv~ < *pleh1-tu-). Olsen therefore suggests the reconstruction of the Latin sufix -idus from *-thos < *-h1-tós, with the laryngeal *h1 representing the zero grade of the sufix *-eh1- of stative verbs, which she derives the given adjectives from. One weak point of the theory can be seen at irst sight: such process evidently did not occur in the participles of the type (g)nātus (< *g’nh1-tós), and these, in my opinion, can be only with dificulty considered secondary, analogical. Concerning the element -i- in the sufix, possible variants of its origin are summed up by Sznajder (2002: 63f.): it can be a) the -i- in the compound according to the Caland-Wackernagel’s law (see above note 1), or b) the compositional vowel -i- (which prevails in Latin, disregarding the form of the irst member of the compound) or c) the reduced thematic vowel e/o of the base verbs, or d) the relex of the laryngeal *h1 of the sufix of stative verbs. To which can be added that quite simply this can also be the anaptyctic vowel, common in Latin words on the seam between the root and the sufix. Let us briely once again sum up the fundamentals of the mentioned hypotheses and their strong and weak points. The hypothesis in the point 2), considering Latin adjectives in -idus to be original 90 LUCIE PULTROVÁ compounds, is very tempting from the formal point of view. The so-formed adjectives would have equivalents in Latin adjectives in -icus and also -i- closing the irst element of the compound is easy to be explained in several ways, according to what word-formative type we assign the irst element of the compound (the orthodox Indo-Europeanist would probably lean to the “Caland’s” -i-). What causes troubles here is, however, the semantic aspect of the matter: to interpret the adj. calidus as “giving warmth” instead of “being warm” (this way, we would probably have to expect sooner *calibus (?) < *-bhu-os) seems to be rather purposebuilt; however, it cannot be excluded. Other above mentioned hypotheses presuppose a common sufixal derivation, the scholars, however, disagree on what type of derivation is in question, whether deverbative or denominative one. LIV treats the verbs of the type calēre among the primary verbs. Contrary to the traditional image of the form of stative verbs (sufix *-eh1-), it attributes to them the structure *R(z)-h1Ôé-,5 i.e. the sufix consisting of two elements, *-h1(the zero-grade of the stative sufix *-eh1-) + the sufix -Ôé-. What must be added, however, is that when going through the LIV dictionary, next to the Latin verbs of similar type we can usually ind the notes such as “with analogical R(e)”, “neologism” etc., in other words, in most cases the suggested reconstruction in fact does not, following the up-to-now deined sound laws, correspond to the Latin outcome. This interpretation, namely that the verbs of the type calēre are primary verbs, implies that the adjectives of the type calidus are deverbatives. Traditionally, these adjectives are regarded as deverbatives also by the researches who postulate the inherited sufix *-do- / *-dho- (not by all, though, see Nussbaum below); this interpretation would probably be preferred also if we accepted the hypothesis of the origin from the PIE *-ro-, which is sooner considered to be deverbative,6 and certainly this applies to the hypothesis by Olsen. Regarding the latter, we must express two critical comments, which, in my opinion, fundamentally dispute its validity: 1. Were we to attribute a verbal characteristics to Latin adjectives in -idus, then we would deinitely have to describe them as “stative adjectives”. The PIE verbal adjectives with the sufix in -tó- are, however, anything but stative: they are in principle adjectives derived from the active (transitive) verbs. Stative, i.e. perfect verbal adjectives, have in PIE the reconstructed sufix *-„es- (in Latin, we may perhaps consider even the transition to the sufix -„o-, see Pultrová 2006b: 54ff.), 5 6 “R(z)” is a symbol used in LIV to denote the root in the zero grade (× “R(e)” = the root in the full e-grade). This, however, is by no means clear, since neither the PIE sufix *-ro- has an unambiguous function; e.g. Leumann (1977: 315) introduces the interpretation of Latin adjectives in -roby the following: “soweit etymologisch ableitbar, Deverbative”; Brugmann (1906: 348ff.) evades the deinition of the function of the sufix, he offers just mere list of IE formations in -ro-. THE LATIN ADJECTIVES WITH THE SUFFIx -IdUs 91 but by no means *-to-; we cannot presuppose any analogical secondary formation here, since the very basis of Olsen’s hypothesis is that these adjectives must have been formed very early, surely in the PIE period, before the elimination of laryngeals took place. 2. The PIE verbal adjectives in *-tó- are primary verbal derivations, which is to say that they are derived by adding the sufix directly to the verbal root, not to the verbal stem. If we were, therefore, disregarding the semantics, to derive standardly the adjective in -tó- from the root e.g. of the verb calēre, the result would read as follows: *k’C-tós > Lat. **kultus (?). In contrast with LIV, Watkins (1971: 68) describes the verbs of the type calēre as “adjective-verbs” and classes them sooner with denominative statives. I consider this interpretation absolutely pertinent. If we acknowledge that in the deep structure of language there exist the categories of quality, which hardly anyone would dispute, then it is hard to imagine which word type would better conform to this particular category. To put it simply, it is certainly more natural to derive the verb with the meaning of “to be warm” (calēre) or “to get warm” (calēscere) from the adjective “warm” than vice versa (besides, modern languages clearly attest to it). All in all, I believe that we must start from the fact that neither in the case of the verbs of the type calēre, nor in the case of the adjectives in -idus we deal with deverbative derivations; they do not express the relation towards an action, but they denote the quality of a substance. From the semantic aspect then we must regard the adjectives in -idus as “adjectives of quality”. The same applies also to the related abstract nouns in -or (calor) – also these are clear cut qualitative abstract nouns, not verbal abstract nouns (nouns of action). For that matter, Nussbaum (1999) came to the same conclusion, when he presupposed the evolution from the primary adjective through the i-abstract noun to the adjective in -idus (demonstrated on the relation between the adjectives rūfus – rūbidus – p. 404). The adjectives of quality generally derive both from the concrete and the abstract nouns. Turning for help to the mother language of the author (since English is actually not very “eloquent” in discussing word-formation), we can list as the examples of the adjectives of quality derived from concretes for instance tuk – tučný (“fat – fatty”), špína – špinavý (“dirt – dirty”), piha – pihatý (“freckle – freckled”), from the abstract nouns then for example naděje – nadějný (“hope – hopeful”), štěstí – šťastný (“happiness – happy”) etc.7 In Latin, corresponding to this type is e.g. barba – barbātus (in Latin grammar books these types of adjectives usually are referred to not as “qualitative” but “possessive”, in the sense of “furnished with something”). Here, a little digression must be made: The most important types of denominative adjectives in Latin have the sufixes identical with the deverbative sufixes 7 The examples are taken from MČ, I, 370f. 92 LUCIE PULTROVÁ (in particular -nus, -lis, -tus); their main characteristics, however, is – regarded synchronically and disregarding the marginal cases – that they join the stem of their founding noun, and if the stem is vocalic, the respective vowel is lengthened. Thus we have barba – barbātus, crīnis – crīnītus, Rōma – Rōmānus, tribus – tribūnus, cūria – cūriālis etc. Apparently, there must have been another wordformative element in between the sufix itself and the founding word; naturally, the connection with laryngeal offers itself. Bader (1992: 99ff.) identiied in similar cases the denominative sufix of appurtenance *-h2-. This brings us back to the hypothesis by Olsen, to which I have expressed critical objections earlier; this, however, certainly is not meant to dispute the mentioned phonological principle, e.g. that the laryngeals *h1 or *h2 in some cases cause the aspiration of the succeeding t. On the contrary, based on that principle, the theory by Bader and, at the same time, the semantic analysis of the Latin adjectives in -idus, a new hypothesis can be put forth. Latin adjectives in -idus could have originated from the following structure: *f-h2-tos, where F stands for the founding abstract noun with the meaning of quality and the structure *-h2-tos is thus actually identical with the complex sufix occurring in the adjectives of the type barbātus (*-h2- is the Bader’s appurtenance sufix). The founding word must have been an abstract noun in the form of the root word of the type lūx (similarly also rōs – rōridus). Some more comments to be made on the hypothesis: I build on the theory that the group of phonemes with the interconsonantal laryngeal in medial syllable in Latin yields CiC (not CaC as in initial syllables), which I formulated in detail in Pultrová 2006a: 76ff. The general opinion on the evolution of laryngeals in the interconsonantal position is that what occurred here was not the “vocalization” of the laryngeal, but the insertion of the anaptyctic vowel into the consonant group (and in some languages, e.g. in Greek, the anaptyctic vowel was then coloured according to the type of the laryngeal). The development of the researched group of phonemes could then have been as follows: *-ch2t- > *-cth- > through anaptyxis *-cith- > -c-id-. To be fair we must admit that the weak point of this hypothesis lies in its dificult veriication, at least within the scope of Latin. Unambiguously parallel formations are not likely to be found in the structure of Latin language – I mean the formations where the consonant ending of the founding word was joined with the sufix of appurtenance *-h2- followed by the sufix in -t-. Suppose we were looking for the parallels on the purely phonological level, then we are sure to ind the words containing the consonant group CHt8 in Latin, 8 C ≠ R; in the group *RHt we have to take into account a different development. THE LATIN ADJECTIVES WITH THE SUFFIx -IdUs 93 although even of these there are not many (cf. e.g. Schrijver 1991: 94, 99). Even more limited list of words we shall obtain if we are interested in this consonant group in the medial syllable only. On the basis of Schrijver’s synthetic work, we can class into this ile only the preixed derivations from the verbs patior (< *ph1t-) and fateor (< *bhh2-t-)9, verbal adjectives satus (< *sh1-t-), status (< *sth2-t-), passus (< *peth2-t-), to which must be added datus (< *dh3-t-), ratus (< *rh1-t-), and according to LIV also missus (< *mith2-t-) and fossus (< *bhedhh2-t-), and perhaps Iūpiter < pater (< *ph2t-). No derivation from the above listed founding words shows the phonological development *Ht > *th, but in fact they cannot be considered real parallels to our hypothetical phenomenon, as in all the present cases the laryngeal is the part of the root. The works concerned with laryngeals in general mechanically research the development of the laryngeals in speciic phonological environments, that is of speciic groups of phonemes (e.g. CHC, CRHC etc.), regardless of to which morpheme the single phonemes of the group belong. In my opinion it is evident that the development of phoneme groups depends on the morphematic structure of the word and in this respect the sound laws should be speciied, above all when deining the development of laryngeals (but also of other phonemes, e.g. syllabic resonants). Thus for example the PPP from the verbs with the root HeC- would have to look quite differently from the reality should we use the “mechanical phonological” interpretation (*HC- > C-), e.g. edō: *h1d-tos > **dtos > through anaptyxis **dVtos, or even **ttos > **ssos > **sus (?). Language, or at least Latin, however, will not allow such development – actually eliminating the root – and opts for an alternative. Yet another example of a different kind: so-called ā-intensive (the type facere – -icāre, see Leumann 1977: 534, 549) is formed from the verb stāre in quite a different way than from the other verbs: to “help” with the formation, the n-inix is needed (-stināre), since by undergoing the standard derivation the sufix would completely cease to exist, the derivation would be homonymous with the base verb and thus unclear as to the meaning. In a word, phonological and word-formative processes, at least in Latin, in principle occur in the way as not to obscure the meaning of the word – as if there always were more variants of how the language can “deal with” unsustainable group of phonemes. Should the standard way bring obscuring of the word’s sense, the alternative is chosen. What must be always maintained is a clear (syllabic) root and an unambiguous ending; an unambiguous word-formative sufix then where it is “grammaticalized” (i.e. it became the means to express a grammar category; in Latin for example the deverbative sufix -tus). Our example falls right within these theoretical relections. To sum up, I suggest that we add one more hypothesis to the up-to-now presented interpretations of the origin of Latin adjectival -idus, one which presupposes the 9 The same reconstruction applies also to the adj. fātus (to fārī), which has in Latin a problematic (dificult to explain) long ā. 94 LUCIE PULTROVÁ structure of “root qualitative abstract noun + *-h2- + *-tos” and which is in fact an attempt to synthesize the theories by Nussbaum (on denominative character of the adjectives in -idus), Bader (on derivation of denominative nouns and adjectives by the means of the sufix of appurtenance *-h2-) and Olsen (on the development *h2t > th > d). To either conirm or disprove this hypothesis will only be possible after further research in Latin phonological phenomena, above all the development of laryngeals, against the background of morphematics and word-formation. RefeRences Bader, F. (1975). La loi de Caland et Wackernagel en grec. In Mélanges linguistiques offerts à Émile Benveniste. Paris, Société de linguistique de Paris, 19–32. Bader, F. (1992). Problématique du génitif thématique. Bulletin de la société de linguistique de Paris, 87 (1), 71–119. Balles, I. (2003). Die lateinische Adjektive auf -idus und das Calandsystem. In e. TIchy – d. s. WodTko – B. IrslInger (eds.). Indogermanisches Nomen (Akten der Arbeitstagung der Indogermanischen Gesellschaft, Freiburg, 19. bis 22. September 2001). Bremen, Hempen Verlag, 9–29. BenvenIsTe, e. (1935). Origines de la formation des noms en indo-européen. Paris, Librairie Adrien-Maisonneuve. Bloch, a. (1954). Zur Herkunft der lateinischen Adjektive auf -idus. In sprachgeschichte und Wortbedeutung (Festschrift Albert debrunner). Bern, Francke, 19–32. Brugmann, k. (1906). Grundriss der vergleichenden Grammatik, II,1 (Lehre von den Wortformen und ihrem Gebrauch). Strassburg, K. J. Trübner. ernouT, a. (1957). Metus – timor. In Philologica II. Paris, Librairie C. Klincksieck, 7–56. ernouT, a. (1971). Les mots en -eō, -or, -idus. In Notes de philologie latine. Paris – Genève, Droz, 1–18. krIsch, Th. (1992). Analogische Prozesse in der lateinischen Sprachgeschichte. In o. Panagl – T. krIsch (eds.). Latein und Indogermanisch (Akten des Kolloquiums der Indogermanischen Gesellschaft, Salzburg, 23.–26. September 1986). Innsbruck, Institut für Sprachwissenschaft der Universität Innsbruck, 155–181. leumann, m. (1977). Lateinische Laut- und Formenlehre, Lateinische Grammatik I. München, C. H. Beck. LIV = rIx, h., eT al. (2001). Lexikon der indogermanischen Verben. Wiesbaden, Reichert. MČ = PeTr, J., eT al. (1986). Mluvnice češtiny. Praha, Academia. nIedermann, m. (1899). Studien zur Geschichte der lateinischen Wortbildung. Indogermanische Forschungen 10, 221–258. nussBaum, a. J. (1999). *Jocidus: An Account of the Latin Adjectives in -idus. In h. eIchner – h. ch. luschüTzky – v. sadovskI (eds.). Compositiones Indogermanicae in memoriam Jochem schindler. Praha, Enigma Corporation, 377–419. olsen, B. a. (1994). The stages of IE Aspiration by Laryngeal. In g. e. dunkel – g. meyer – s. scarlaTa – ch. seIdl (eds.). Früh-, Mittel-, spätindogermanisch (Akten der Ix. Fachtagung der Indogermanischen Gesellschaft vom 5. bis 9. Oktober 1992 in Zürich). Wiesbaden, Reichert, 267–277. osThoFF, h. (1878). das Verbum in der Nominalkomposition im deutschen, Griechischen, slavischen und Romanischen. Jena. PulTrová, l. (2006a). The Vocalism of Latin Medial syllables. Praha, Karolinum. PulTrová, l. (2006b). The Indo-European Verbal Adjectives and their Relexes in Latin. Listy ilologické CxxIx (1–2), 51–70. THE LATIN ADJECTIVES WITH THE SUFFIx -IdUs 95 schrIJver, P. (1991). The Relexes of the Proto-Indo-European Laryngeals in Latin. Amsterdam – Atlanta (GA), Rodopi. sIhler, a.l. (1995). New Comparative Grammar of Greek and Latin. New York/Oxford, Oxford University Press. sznaJder, l. (2002). Les adjectifs en -idus, -ida, -idum. In ch. kIrcher-durand (ed.). Grammaire fondamentale du latin, Tome Ix: Création lexicale: la formation des noms par dérivation sufixale. Louvain – Paris – Dudley (MA), Peeters, 55–65. WaTkIns, c. (1971). Hittite and Indo-European studies: the denominative statives in -ē-. Transactions of the Philological society, 51–93. ResuMÉ Článek se vrací k mnohokrát řešené, a přesto dosud uspokojivě nevyřešené otázce původu latinských adjektiv na -idus. Shrnuje dosavadní teorie a nabízí i hypotézu novou, která je pokusem o syntézu teorií Nussbauma (o denominativním charakteru adjektiv na -idus; 1999), Baderové (o odvozování denominativních jmen a adjektiv prostřednictvím příslušenského suixu *-h2-; 1992) a Olsenové (o vývoji *h2t > th > d; 1994). Podle této hypotézy jsou lat. adjektiva na -idus denominativní adjektiva odvozená prostřednictvím příslušenského suixu *-h2- a suixu -tus, jímž se odvozují jakostní adjektiva. Lucie Pultrová Ústav řeckých a latinských studií FF UK (Lucie.Pultrova@ff.cuni.cz)