Journal of Latin Linguistics 2018; 17(1): 93–110 Lucie Pultrová* Periphrastic comparison in Latin https://doi.org/10.1515/joll-2018-0004 Abstract: Comparison of adjectives (and adverbs) is a grammatical category that has passed practically unremarked upon by generations of Latin linguists. Latin grammar books (with few exceptions, cf. Kühner and Stegmann, 1955: 565– 566The Oxford Latin syntax. Volume I: The simple clause, 47. Oxford: Oxford University Press) omit entirely the question of which adjectives can be compared and which cannot. Nevertheless, the data from modern languages show that the category of comparison of adjectives (and adverbs) is actually highly limited, making it essential to address this question for Latin, too. One of the issues comprehended within the extremely complex area of (non-) gradability of adjectives is periphrastic comparison. Latin grammar books explain, based on the assertions of ancient Latin grammarians, that it applies to adjectives ending in -eus, -ius and -uus, implying, or even explicitly stating, that the reason for this type of comparison is phonetic incompatibility of the word-formative suffix with the comparative suffix. However, two facts call for reinterpretation of the matter: (i) periphrastic comparison also occurs in other adjectives for which there is no phonetic incompatibility; (ii) by contrast, some adjectives in -eus, -ius and -uus actually do have simple forms. In the light of these facts, this paper aims to map the real situation of periphrastic comparison in Latin. The employed corpus comprises all the words marked as adjectives in the Oxford Latin Dictionary (more than 10,000 items) and all their occurrences throughout the database Bibliotheca Teubneriana Latina III. Keywords: Latin, adjectives, gradability, comparison, periphrastic comparison 1 The category of comparison The grammatical category of comparison is the Cinderella of Latin linguistics: studies treating it from other than a syntactic point of view are scarce.1 The 1 Two semantic studies of the relative and absolute use of comparative and superlative forms were written by Fugier (1971; Fugier 1972). Purely formal aspects of the formation of superlatives and some comparatives were studied by Szemerényi (1976), and superlative suffixes were *Corresponding author: Lucie Pultrová, Institute of Greek and Latin Studies, Faculty of Arts, Charles University, Prague, Czech Republic, E-mail: Lucie.Pultrova@ff.cuni.cz Unauthenticated Download Date | 1/5/19 10:44 PM 94 Lucie Pultrová medium-sized Latin grammar books used by students and most influential on general knowledge of Latin usually confine their comments merely to stating that adjectives and adverbs can be compared and to describing degree forms.2 However, the fact that comparison does not apply to every adjective is rarely mentioned.3 The comprehensive Latin grammar books are not much better in this respect, either – only very exceptionally is comparison treated as other than a universal category: Kühner and Holzweissig write that “eine nicht geringe Zahl von Adjektiven entbehrt der Steigerungsformen” (Kühner and Holzweissig 1912: 565),4 and Pinkster writes that “in the case of quite a few adjectives, derivation of comparative and superlative adjectives is excluded for semantic and/or formal reasons” (Pinkster 2015: 47).5 But neither Kühner and Holzweissig’s “nicht geringe Zahl” nor Pinkster’s “quite a few” reflect the true extent of the phenomenon. For example, according to the Czech National Corpus, in Czech – a language very similar to Latin in its range of inflections – just 6 % (!) of adjectives have degree forms, and only 3 % have both degrees of comparison.6 treated by Bartoněk (1955). From the syntactic perspective, by contrast, comparative constructions in Latin have been covered in great detail, e.g. Bertocchi and Orlandini (1996), Cuzzolin (2011). 2 For example, Panhuis asserts, as if with universal validity, “[t]he adjective has three degrees of comparison: …” (Panhuis 2006: 25). Similarly, Oniga goes straight in medias res, i.e. to the way of formation: “Adjectives decline in three possible degrees …” (Oniga 2014: 88). Touratier makes a more adequate formulation, employing potential modality: “Die Adjektive können mit besonderen Morphemen verbunden werden, die den Grad einer durch das Adjektiv ausgedrückten Eigenschaft anzeigen” (Touratier 2013: 75). 3 We are facing here a more general drawback of teaching dead languages: students are drilled to manufacture grammatical forms mechanically, without regard to their actual usability. In Czech grammar school textbooks, we can find many forms that are contextually difficult to imagine (and actually not attested), such as carpebaris or cremabimini, to choose some examples at random. 4 In Kühner and Holzweissig (1912: 565–572), they give a list of the adjectival types concerned. 5 Pinkster follows on the same page with a very brief outline of the adjectives in question. “Semantic reasons” are cited for the so-called ungradable adjectives, namely (according to Pinkster) adjectives of function or purpose, substance, origin or provenance, time, location and adjectives with a negative meaning. 6 See Cvrček et al. (2015: 205). Even in general linguistics it seems not to be appreciated that the number of ungradable adjectives could be as large as the Czech National Corpus suggests; cf. e.g. the statement in the respected handbook Morphology/Morphologie: “One of the defining characteristics of adjectives in languages with appropriate inflection is the propensity to be gradable, in particular to have comparative and superlative forms. However, there are several stative or absolute adjectives that normally lack these forms” (Booij et al. 2000: 650). This situation is undoubtedly influenced by the fact that in general linguistics the primary working language is English, which employs considerably fewer adjectives than for example Latin or Czech. As for the typology of gradable adjectives, cf. e.g. Dixon (2004). For Czech, cf. the inspiring monograph by Marvan (2008). Unauthenticated Download Date | 1/5/19 10:44 PM Periphrastic comparison in Latin 95 It is difficult to provide similarly exact statistics for Latin: given that the surviving textual corpus of a dead language such as Latin represents a fragmentary view, the fact that an item is not attested obviously does not mean that it did not exist. Based on the excerption of approximately 10,000 forms labelled in the Oxford Latin Dictionary as adjectives7 and their presence in the database Bibliotheca Teubneriana Latina III, a very rough estimate gives that comparative or superlative forms (synthetic and/or periphrastic) are attested in Latin in approximately 15 % of adjectives, while both degrees are found in about 8 % of adjectives.8 It is therefore clear that this grammatical category has limited scope in Latin, too. Generally speaking, we can compare adjectives that denote a quality that can be expressed on a scale; these are often called “scalar” adjectives. While this fact is commonly known in linguistics and is relatively trivial, it is problematized when working with concrete language material as in fact there are many instances when – some apparently ungradable adjectives allow degrees; – some apparently gradable adjectives do not allow degrees. And it is these instances of discrepancy that are interesting from a linguistic perspective. The comparison of apparently ungradable adjectives can be easily explained at a general level by semantic shifts: ungradable adjectives frequently become gradable when bearing a metaphorical meaning (e.g. the bluest song, the most wooden actor, etc.). The absence of degrees in adjectives that seemingly present no obstacle to comparison is theoretically more complicated to explain. To be 7 I write “labelled as adjectives” and not simply “adjectives” because to make a clear distinction between adjectives and participles and between adjectives and substantives is not at all trivial. The choice of criteria for such a distinction affects both what counts as an adjective and what counts as a gradable adjective, with a consequent effect on the observed statistics. Criteria to distinguish between substantival, adjectival and verbal uses of present participles were set by Eklund (1970: 14–30); one of the criteria to mark the forms in -nt- as adjectives is precisely whether they have degrees. I presume that the same criterion was applied in the OLD when assessing the forms in -tus/-sus and -ndus: if a degree form exists then it is without doubt classed as adjective with its own entry. If a degree form is not attested, however, interpretation is more problematic. Consistency of interpretation is essential when collecting statistical data on how many adjectives form degrees and how many do not. 8 These data must be considered provisional at the moment: The author of this study is currently preparing a monograph on the topic of comparison in Latin; as work progresses, all the preliminary data are individually assessed, with total counts adjusted accordingly (the trend appears largely to be downward due to the fact that upon deeper analysis some originally excerpted instances of comparison end up being excluded as irrelevant). Unauthenticated Download Date | 1/5/19 10:44 PM 96 Lucie Pultrová able to address these questions satisfactorily, a thorough and complex analysis of the semantics of Latin adjectives and adjectival types is needed. (This need will be met in the form of a monograph currently being prepared by the author – see note 8.) The present study has the more modest aim of examining a single subtopic of the second instance of discrepancy above, namely periphrastic comparison. Latin adjectives that are compared periphrastically are from the semantic point of view evidently gradable, but for some reason comparison is not achieved by means of inflection.9 The question is for which adjectives this is true and why. 2 Periphrastic comparison in Latin grammar books Latin grammar books habitually offer clear-cut answers to the questions of which adjectives are compared periphrastically and why. For example, Leumann: “Steigerung von Adjektiven auf -uus und -ius erfolgt normalerweise durch Umschreibung mit magis und maxime […]. Im Komparativ wird damit die Folge von drei Vokalen vermieden; und dem magis folgt maxime. Soweit -ior -issimus gebraucht sind, wird bei Adjektiven auf -ius das erste i von -i-ior übergangen. […] Die Folge -uior bei -uus ist weniger anstößig” (Leumann 1977: 498). Hofmann and Szantyr: “Es [sc. magis] tritt zunächst dort umschreibend ein, wo lautliche und formale Gründe die Bildung der Steigerungsgrade nicht empfehlen; so bei den Adj. auf -ius, -uus (und -eus …) zur Vermeidung der Folge -iior, -uior” (Hofmann and Szantyr 1965: 165). Also Pinkster, referring to adjectives in which derivation of the usual comparative and superlative is excluded for formal reasons,10 cites the adjectives ending in -ius, -eus and adds: “As substitutes for the missing comparatives and superlatives, combinations of these adjectives with the degree adverbs magis ‘more’ and maxime ‘most’ are used” (Pinkster 2015: 47). As is apparent from the above set of quotations, the suffixes -ius, -eus and -uus are considered phonetically incompatible with degree suffixes, or, 9 From the synchronic point of view, there is no functional difference between the two forms of comparison; cf. Cuzzolin (2011): “the way the marker of comparison is expressed, whether it occurs as a morphological element […] or as a lexical one […] is not as important as the standard of comparison, because the process by which the comparative is formed does not affect the nature of comparison” (Cuzzolin 2011: 559). 10 On semantic reasons, see note 5 and the quotation it remarks upon. Unauthenticated Download Date | 1/5/19 10:44 PM Periphrastic comparison in Latin 97 specifically, with the comparative suffix -ior- (thus Leumann and Hofmann & Szantyr). For this reason, supposedly, periphrastic comparison with the degree adverbs magis and maxime was introduced in the corresponding adjectives (according to Leumann and Hofmann & Szantyr this first took place in the comparative with magis and only secondarily, analogically, in the superlative with maxime, since the superlative suffix would not itself present such a phonetic barrier). Modern grammar books in this regard are undoubtedly echoing formulations repeated in the texts of Late Latin grammarians. For example, Priscian: Inveniuntur quaedam, quae quamvis sint accidentia [id est adiectiva] et eorum significatio exigat, ut faciant comparativa, tamen non habentur in usu frequenti. Sunt autem ea plerumque, quae vocales ante ‘us’ habent, ut ‘pius’, ‘arduus’, ‘egregius’, ‘dubius’, ‘strenuus’ … Et puto hanc esse rationem, quod oportet comparativum una syllaba vincere genetivum positivi, nisi sint anomala … Si ergo haec, quae ante ‘us’ vocalem habent, assumant genetivo ‘or’, necesse est inter duas vocales positam i transire in vim consonantis, quod in Latinis dictionibus semper fere patitur, cum inter duas vocales invenitur vim suam servantes … Plerique assumunt igitur ‘magis’ adverbium et usum comparativi complent, ut ‘magis pius hic quam ille’ … (Priscian GL II 86,21–87,15) In fact, however, matters are more complicated: periphrastic comparison with magis and maxime in Latin does not only apply to adjectives that “have a vowel before -us”, i.e. the adjectives in -eus, -ius and -uus. This can even be seen in the list given by Kühner and Holzweissig (1912: 565–572), referred to earlier in note 4: the set of adjective suffixes that avoid degrees “aus Wohllautsrücksichten” also includes adjectives in -icus, -alis, -aris, -ilis, -atus, -itus, -ndus, -inus and -ivus. Why these suffixes in particular should be phonetically incompatible with the comparative, or even the superlative suffix, remains unclear. Moreover, the inflected graded forms (-ior, -issimus) actually do exist in many of these adjectives.11 And let us add a more general note: in principle, to cite purely phonetic reasons for a grammatical differentiation is dubious, as language practically always has the phonetic means (assimilation, epenthesis, etc.) to avoid such differentiations being required. In the following section we give an overview of the empirical results found by extensive excerption of material, in which we deal with more than 10,000 adjectives contained in the Oxford Latin Dictionary. The forms for these adjectives were found in the database of Bibliotheca Teubneriana Latina III, both synthetic comparative and superlative (which are not the subject of this study) 11 Thus also Priscian in the text immediately following on from the passage quoted earlier (GL II 87,14): vetustissimi tamen comparativis etiam huiuscemodi sunt usi; then featuring examples of non-standard (in his view) synthetic comparatives and superlatives (e.g. arduius, arduissimus, egregiissima, etc.) in Cato, Pacuvius, Juvenal and others. Unauthenticated Download Date | 1/5/19 10:44 PM 98 Lucie Pultrová and periphrastic comparative and superlative (i.e. with magis and maxime). Section 3.1 summarizes the data concerning the forms with magis, and Section 3.2, the forms with maxime. We conclude in Section 3.3 with some interpretations of these data. 3 Primary sources analysis The adverb magis occurs 8,108 times in the BTL III database; maxime, 5,197 times. The functions vary, however: A. As an independent adverb/particle, modifying the whole predication; magis in the sense of ‘rather’ (often linked with sed following negation, i.e. ‘not X, but rather Y’), maxime meaning ‘above all’, ‘in particular’: (ENN. ann. 272) non ex iure manum consertum, sed magis ferro rem repetunt (CIC. fam. VII,18,1) quae ego paulisper in te ita desideravi, ut non imbecillitate animi tui, sed magis ut desiderio nostri te aestuare putarem (CIC. nat. deor. II,25) id maxime fieri temporibus hibernis B. magis linked with quam: (CIC. Att. X,1a,4) disertus magis est quam sapiens (SALL. Iug. 20,2) metuens magis quam metuendus There is no relation of comparison (i.e. one of higher or lower intensity of quality) between the two coordinate words; on the contrary, they stand in contradiction, one excluding the other.12 (Semantically, this is actually a very similar type of phrase to that in A: non est sapiens, sed magis disertus; non est metuendus, sed magis metuens.) C. As an adverb modifying a modal or action verb, or participle/gerundive (maxime often comes after quam, ‘as much as possible’): (CIC. Att. IX,1,4) cui magis etiam Caesar irascetur 12 On this, see Bertocchi and Orlandini (1996: 222–223). Unauthenticated Download Date | 1/5/19 10:44 PM Periphrastic comparison in Latin 99 (CIC. Att. VIII,14,1) omnino intellego nullum fuisse tempus post has fugas et formidines, quod magis debuerit mutum esse a litteris (LIV. XXXIX,35,6) ne victi magis timendi forent, quam bellantes fuissent (CIC. Verr., actio secunda, II,150) statuae … quae populi Romani oculos animosque maxime offendunt (CIC. Scaur. 33) voluit eum quam maxime offensum Even though these examples express higher intensity, this is not one of quality but rather one of “quantity” of action, or its necessity/suitability. D. As an adverb modifying a stative verb: (ENN. ann. 372) magis que viri nunc gloria claret (SEN. nat. IVb,11,4) qui dicit altiorem montem, quia solem propius excipiat, magis calere debere… (NEP. Milt. 1,1) sua modestia unus omnium maxime floreret In these cases, magis expresses a higher intensity of quality (our general focus); however, unlike the case of adjectives, in verbs there is no competition between periphrastic and synthetic comparison, which is the actual subject of the present study. (Nevertheless, the question of verb comparison is per se very interesting and deserves an individual study.) E. As an adverb modifying an adjective or other adverb, i.e. as a means of periphrastic comparison (both in the absolute and relative sense).13,14 13 Absolute and relative meanings of Latin degree forms were treated in detail by Fugier in her two studies (Fugier 1971; Fugier 1972). This distinction is highly pertinent to the framework of studies of the category of comparison in general; nevertheless, in relation to the specific subject of the present partial study, it is irrelevant, so we will lay it aside for now. 14 This is only a very rough enumeration. Moreover, individual types of phrases with adverbs magis and maxime represent a kind of continuum. As we have observed, the examples featured under A and B are semantically very close. The same applies to D and E. There are overlaps among other pairs as well, and it is not always easy (or indeed possible) to decide unequivocally into which group a particular instance belongs. For example, the phrase with magis/maxime and a verb form composed of a participle or gerundive and the verb esse is practically always on the borderline between C and E; likewise, in adjectives in -bilis it is modality that is compared rather than the quality itself, etc. See also the case of quanto magis aurei coloris in note 25. Unauthenticated Download Date | 1/5/19 10:44 PM 100 Lucie Pultrová 3.1 Magis Magis linked with an adjective or adverb, i.e. type E in the list above, appears in the whole BTL III corpus approximately15 660 times, comprising 8 % of the attested examples of the adverb magis. Regarding the number of adjectives and adverbs modified in this way, in the whole corpus we can find approximately 400, i.e. a mere 4 % of the excerpted adjectives. Of these 400 adjectives there are about 70 for which the comparative with magis is attested in more than two instances. The rest have just one or two records of a periphrastic comparative, with single occurrences being predominant. The majority of adjectives with an attested periphrastic comparative (280 out of 400) also have attested examples of a synthetic comparative. This leaves 120 of these adjectives (30 %) with no attested synthetic comparative. If we take into account only the adjectives with a statistically significant number of occurrences of a periphrastic comparative (= more than two examples), there are 17 with no attested synthetic comparative, i.e. about 24 %. They are as follows: anxius, comicus,16 contrarius, decorus, dubius, idoneus, mirandus, noxius, obnoxius, par, perspicuus, pius,17 ridiculus, rutilus, vacuus, varius, vulgaris.18 These figures are summarized in Table 1. Among the adjectives with examples of both periphrastic and synthetic comparatives there are only five in which the former predominates to any significant degree over the latter. They are as follows: assiduus (assiduior 2x in Varro, magis assiduus 4x in other authors), fidus (fidior 3x, magis fidus very 15 The reason why I give a round number here with the modifier “approximately” is that while carrying out my analysis I often reassess some cases originally regarded as periphrastic comparatives as collocations with the adverb magis that perform another function, and thus as being irrelevant. The exact figures are therefore still being adjusted – but the final count differs only by single units and is always rounded down. Included in the total are examples of the “distant” periphrastic comparative; e.g. magis decorum et insigne (PLIN. epist. I,17,4) counts as two periphrastic comparatives: magis decorum and (distant) magis insigne. 16 In the case of the adjective comicus we can find three examples, all attested in Aelius Donatus – this could be a matter of idiolect. 17 The synthetic comparative pior is attested only in the texts of grammarians, with the commentary claiming that it is not used and that the expression magis pius is used instead. However, beside these instances magis pius is attested only twice, and one of these examples is quo magis aptum pium que est (PLIN. paneg. 1,6), where the connection of magis and pius is loose. On the other hand, the synthetic superlative piissimus is quite common – see Section 3.2. 18 To make the list complete: There are 14 adjectives with only two records of a periphrastic comparative and no record of a synthetic comparative: admirandus, canorus, conspicuus, elaboratus, exitialis, gnarus, ignarus, impiger, infamis, miserandus, patulus, praeceps, pudendus, rubens. Unauthenticated Download Date | 1/5/19 10:44 PM Periphrastic comparison in Latin 101 Table 1: Attestations of periphrastic comparative in BTL III. Adjectives with attested periphrastic comparative Adjectives with more than two occurrences of periphrastic comparative Total No attested synthetic comparative     common), mirus (mirior only 2x in fragments, magis mirus common), necessarius (necessarior only 1x + 3x in the text of grammarians with the commentary claiming the form is not used, magis necessarius common), popularis (popularius 1x in Livy, magis popularis 6x, of which 3x in Livy).19 In general, most adjectives with attested examples of a periphrastic comparative also have records of a synthetic comparative, which moreover occurs more frequently; periphrastic comparatives are thus usually less frequent as a variant than synthetic comparatives. Quantitatively, adjectives with a given number of examples of a periphrastic comparative have proportionately many examples of a synthetic comparative.20 Of the adjectives with solely or mainly periphrastic comparatives attested in more than two instances, the following are the adjectives in -ius, -eus or -uus: -ius: anxius, contrarius, dubius, necessarius, noxius, obnoxius, pius, varius -eus: idoneus -uus: assiduus, perspicuus, vacuus21 The adjectives with other types of suffixes in this group are as follows: comicus, decorus, fidus, mirandus, mirus, par, popularis, ridiculus, rutilus, vulgaris.22 19 A not quite statistically convincing predominance of periphrastic comparatives is to be found also in the following adjectives: arduus (arduius in a Cato fragment in Priscian x magis arduus twice in Pliny the Younger), communis (4 times synth. x 6 times periphr.), generalis (synth. once in Consentius x periphr. once in Quintilian, once in Consentius), informis (synth. once in Seneca x periphr. once in Lucan, once in Pliny the Elder). 20 E.g. utilis, tutus, inanis, verus, gratus, durus, aptus, and many others. 21 Should we wish to add the less statistically conclusive adjectives (see notes 18 and 19), then also arduus and conspicuus. 22 And potentially another 16 listed in notes 18 and 19. Unauthenticated Download Date | 1/5/19 10:44 PM 102 Lucie Pultrová Let us now shift perspective and analyze the adjectival types in -ius, -eus and -uus with respect to the existence of degree forms.23 3.1.1 -ius In total, 944 adjectives in -ius were excerpted from OLD,24 of which only 15 adjectives (i.e. 1.6 %) have any attested comparatives (synthetic and/or periphrastic). Of these 15 adjectives: – in eight adjectives periphrastic comparatives predominate (anxius, contrarius, dubius, necessarius, noxius, obnoxius, pius, varius – already considered above); – two adjectives have a single occurrence of a periphrastic comparative and no example of a synthetic comparative: amatorius, nefarius; – three adjectives have ambiguous records: sobrius (one instance of synthetic comparative in a fragment from Decimus Laberius and a commentary by Charisius claiming that magis sobrius is used instead of the synthetic comparative; however, the periphrastic comparative only appears in Charisius and his followers); egregius (one example of a synthetic comparative in the late grammar work Ars Bernensis with a commentary claiming that the synthetic comparative is not used, but no occurrence of a periphrastic comparative); caesius (a note in Varro that the synthetic comparative is not used, unlike the superlative; but, at the same time, no periphrastic comparative is attested); – the adjective industrius has one instance of a synthetic comparative (in Plautus) and no instance of a periphrastic comparative; 23 Particularly in the case of adjectives ending in -ius, here we are not dealing with a homogeneous word-formative type: Adjectives with the complex suffixes -acius, -icius, -arius and -orius are also included. Among adjectives ending in -eus, there are also found those with the complex suffixes -aceus and -aneus and the Latinized Greek adjectives in -ēus (or -aeus < Gr. -ειος / -αιος), e.g. boreus, colosseus, draconteus and others. Concerning adjectives in -uus, compounds with second element -fluus are also included, despite the fact that their first -u- is part of the root so that by no means can this be a case of derivation by the suffix -uus; nevertheless, these compounds do belong here taking into account the way Priscian defined this class ( = adjectives having a vowel before -us). 24 We count in only appellatives here, even though the suffix -ius is also found in a considerable number of proper adjectives, which represent almost one-fifth of the 10,000 adjectives excerpted from the OLD. However, proper adjectives, with a very few exceptions, do not form degrees; if they were included in our statistics, the percentage of adjectives with the suffix -ius with attested instances of degree forms would decrease even more dramatically. Unauthenticated Download Date | 1/5/19 10:44 PM Periphrastic comparison in Latin – 103 the adjective proprius has 10 instances of the synthetic comparative, while editors tend to correct the form to propior (e.g. Varro, rust. II,9,8); there are dozens of instances of the periphrastic comparative. 3.1.2 -eus In total, 304 adjectives, of which only five have any attested comparative (i.e. 1.6 %); however, only one of these five is statistically relevant: idoneus. The other four adjectives (aureus, caeruleus, consentaneus, purpureus)25 have only a single instance of an attested comparative, and this is always periphrastic. There are no examples of synthetic comparatives for adjectives in -eus. 3.1.3 -uus In total, 67 adjectives,26 of which 12 form comparatives (i.e. 17.9 %). Of these 12 adjectives, there are five with an attested instance of a synthetic comparative; 11 adjectives have attested examples of a periphrastic comparative (five adjectives have just one occurrence; three adjectives, two occurrences; three adjectives, more than two occurrences: assiduus, perspicuus, vacuus). Synthetic comparatives are attested in the following adjectives: arduus (see note 19), assiduus (2x in Varro vs. magis assiduus 4x in other authors), exiguus (2x in Frontinus, Columella vs. 1x periphrastic comparative likewise in Columella), perpetuus (perpetuius in a fragment from Cato cited in Priscian [GL II 88,9–10] vs. magis perpetuus not attested), strenuus (strenuior 1x in Plautus, 1x in a fragment from Lucillius cited in Priscian in the same place as perpetuius). We will attempt an analysis of these results later in Section 3.3; first, however, we give the corresponding statistics for the superlative with maxime. 25 The comparative magis consentaneum is attested in Cicero, in whose works we also find the only instances of maxime consentaneus (see Section 3.2) – this is therefore probably a matter of Cicero’s idiolect. All the other given adjectives, denoting colors, are attested in the comparative in Pliny the Elder; in the case of the adjective aureus, the particular example reads quanto magis aurei coloris quantoque minus plumbosa, which clearly expresses comparison; formally, however, the modifier magis could be linked to an elliptical esse. 26 Including nine compounds with the second element -fluus – see note 23. Unauthenticated Download Date | 1/5/19 10:44 PM 104 Lucie Pultrová 3.2 Maxime The adverb maxime in its use for periphrastic comparison appears in the corpus of texts edited in BTL III approximately27 520 times, comprising 10 % of its attested occurrences (other occurrences of maxime having different functions). There are approximately 350 adjectives and adverbs modified this way in the whole corpus. Of these 350 adjectives, about 60 have the superlative with maxime attested in more than two instances. The rest only have one or two appearances of this type of comparison, while, as in the case of the comparative with magis, most such comparatives are attested in just one instance. The majority of adjectives with records of a periphrastic superlative have concurrently attested instances of a synthetic superlative: of around 350 adjectives there are 127 with no attested example of a synthetic superlative, i.e. about 36 %. Of the adjectives with more than two instances of a periphrastic superlative, 25 do not have any attested synthetic superlative, i.e. about 20 %. They are as follows: adversarius, consentaneus, contrarius, favorabilis, idoneus, infensus, insignis, laudabilis, lubricus, memorabilis, memorandus, mirabilis, mirus, miserabilis, naturalis, necessarius, obnoxius, popularis, probabilis, proprius, repentinus, rubens, salutaris, terribilis, varius.28 See the figures in Table 2. Table 2: Attestations of periphrastic superlative in BTL III. Adjectives with attested periphrastic superlative Adjectives with more than two occurrences of periphrastic superlative Total No attested synthetic superlative     There are again only five adjectives with attested examples of both periphrastic and synthetic superlatives in which the periphrastic superlative exceeds the synthetic by a relatively clear margin. These are cognitus (synthetic superlative 2x, periphrastic 4x), decorus (synthetic 1x in Apuleius, periphrastic 3x), neglectus 27 The remarks in note 15 apply mutatis mutandis. 28 Another 18 adjectives and one adverb have two appearances of a periphrastic superlative and no appearance of a synthetic superlative: cognatus, conspicuus, credibilis, imbellis, invius, mediterraneus, multiplex, mutabilis, notabilis, noxius, optabilis, sabulosus, spongiosus, sterilis, supervacuus, tempestivus, tolerabilis, vicinus and the adverb procul. Unauthenticated Download Date | 1/5/19 10:44 PM Periphrastic comparison in Latin 105 (synthetic 1x, periphrastic 3x), rudis (synthetic 1x, periphrastic more times in the texts of grammarians, who often tend to list the adjective rudis, together with the adjective sobrius and proprius, as examples of adjectives with periphrastic comparison; beside these instances in the texts of grammarians, there are two more appearances in other texts), suspectus (synthetic 2x, periphrastic 4x). Of the adjectives that form exclusively or predominantly periphrastic superlatives (attested in more than two instances), the following belong among the adjectives in -ius, -eus, -uus. -ius: adversarius, contrarius, necessarius, obnoxius, proprius, varius -eus: consentaneus, idoneus -uus: none Adjectives of other types in this group are more numerous: cognitus, decorus, favorabilis, infensus, insignis, laudabilis, lubricus, memorabilis, memorandus, mirabilis, mirus, miserabilis, naturalis, neglectus, popularis, probabilis, repentinus, rubens, rudis, salutaris, suspectus, terribilis. Let us again now restrict our attention to the adjectival types in -ius, -eus and -uus. 3.2.1 -ius In total, 944 appearances were excerpted (see notes 23 and 24), of which 15 adjectives have a form of superlative attested (i.e. 1.6 %; while this number coincides with the corresponding number of attested comparative forms, the adjectives in question do not always coincide). Of these 15 adjectives: – in six adjectives the periphrastic superlative clearly predominates (adversarius, contrarius, necessarius, obnoxius, proprius, varius – we have seen already that these also have a predominantly periphrastic comparative form); – in another two adjectives the periphrastic superlative occurs twice and there are no synthetic superlatives (invius, noxius); – four adjectives have only a single occurrence of the periphrastic superlative and none of the synthetic superlative: conscius, nefarius, pervius, voluptarius; – two adjectives have ambiguous records: sobrius is only found in the texts of grammarians, where it is commented that this adjective does not form any synthetic superlative (see earlier with magis); egregius has one example of a synthetic superlative (in a fragment from Pacuvius) and one instance of a periphrastic superlative, but a “distant” one (see note 15): maxime gravem et egregiam (VITR. V,4,3); Unauthenticated Download Date | 1/5/19 10:44 PM 106 – Lucie Pultrová the synthetic superlative form of the adjective pius clearly predominates (piissimus – the earliest example in Cicero vs. maxime pius only two appearances, of which one as early as Cato). 3.2.2 -eus In total, 304 adjectives, of which only three have an attested superlative form (i.e. 1 %): idoneus has dozens of occurrences of a periphrastic superlative; the superlative maxime consentaneus is attested three times, all the examples being found in Cicero (cf. already earlier magis consentaneum in 3.1); and maxime mediterraneus appears twice in Cicero’s Verrinae (a colloquial way of saying ‘one which is the furthest inland’). There are no synthetic superlatives for adjectives in -eus. 3.2.3 -uus In total, 67 adjectives (see note 23), of which seven have an attested superlative form, as follows: the adjectives conspicuus and supervacuus have two occurrences of a periphrastic superlative and no occurrences of a synthetic superlative; the adjectives ingenuus and perspicuus each have only a single occurrence of a superlative, in both cases periphrastic; by contrast, exiguus has two appearances of a synthetic and one of a periphrastic superlative, assiduus three instances of a synthetic and one of a periphrastic superlative, and strenuus has as many as 15 occurrences of a synthetic superlative and only a single periphrastic superlative. 3.3 Interpretation of excerpted material We now attempt to analyze the data given in 3.1 and 3.2. The initial thesis we wish to test is that periphrastic comparison occurs in adjectives in -ius, -eus and -uus on account of phonetic incompatibility of these suffixes with the degree suffixes, either with them both or with the comparative suffix alone (see Section 2). For adjectives in -uus, the excerpted material does not support the widespread assumption that periphrastic comparison is the standard and dominant form. Synthetic and periphrastic forms are equally attested. Only three adjectives have more than two records of the periphrastic comparative; by contrast, the synthetic comparative (-uior-) is attested in five adjectives. In the case of the Unauthenticated Download Date | 1/5/19 10:44 PM Periphrastic comparison in Latin 107 superlative, the synthetic form is generally dominant (the form of strenuissimus is attested as many as 15 times). The adjectives in -eus are virtually all ungradable. Exceptions are absolutely marginal: only forms of the adjective idoneus occur with any frequency, and this adjective is etymologically obscure,29 being an oddity among the otherwise relatively homogeneous group of adjectives in -eus of Latin origin.30 Other, more relevant exceptions include the adjectives consentaneus and mediterraneus, which are gradable only in Cicero’s idiolect.31 Adjectives in -ius may also be described as essentially ungradable: 98 % of adjectives with this ending have no attested comparative or superlative. Only in a mere 1 % of them are there a statistically relevant number of comparatives and superlatives. Synthetic forms are few, but they do occur; phonetic barriers are not the issue here (in the comparative, the -i- sound in both suffixes simply merge, i.e. the suffix -ius is “absorbed” by the comparative suffix; nevertheless, ambiguity of meaning – apart from in the genitive plural – never occurs; paradoxically, in these adjectives it is the form of synthetic superlative that is phonetically less favorable, although in the case of the adjective piissimus this form of the superlative is very common). It must be further pointed out that the researched corpus also contains adjectives with various other suffixes that have records of periphrastic comparison only – among such adjectives these other adjectives are twice as numerous as those ending in -ius, -eus and -uus. Moreover, periphrastic comparison is common for some adjectives from this group (e.g. fidus, mirus, and others). It is impossible to delimit this group of adjectives phonologically in any way – the endings of adjectives vary and periphrastic comparison cannot be explained here by phonetic incompatibility. Resulting from the complete work with the excerpted material is the need to amend the generally given basic information on comparison in classical Latin more or less as follows: 1. A substantial majority of Latin adjectives do not form degrees for semantic reasons. 29 For example, De Vaan (2008), s. v. idoneus, says simply, “[n]o etymology.” It is not even clear whether to segment the adjective as idon-eus or as ido-neus. The IEW does not list the adjective idoneus at all. 30 Cf. note 23. 31 The suffix -aneus derives adjectives from various bases (both deverbatives and denominatives), which do not constitute any semantically homogeneous type. The original meaning of the adjective consentaneus is the non-scalar ‘in accord with sth’, ‘corresponding to sth’; by a small semantic shift it acquired a scalar meaning of ‘natural’, ‘appropriate’. Unauthenticated Download Date | 1/5/19 10:44 PM 108 2. 3. Lucie Pultrová Gradable adjectives may be compared by adding comparative and superlative suffixes, or – less frequently – periphrastically, by linking with the adverbs magis and maxime. A small group of formally and semantically heterogeneous adjectives (44 in total) have attested instances of periphrastic comparatives and superlatives in texts, while there are no forms of synthetic comparison. Why it is this group of adjectives in particular that forms periphrastic comparatives and superlatives is not obvious from the synchronic point of view. One suggestion is that the development of the whole category of comparison in Latin could have been as follows: In prehistoric times suffixes had precise meanings, hence derived adjectival types were semantically homogeneous, and, therefore, as a whole either scalar or non-scalar. The historic period saw various semantic shifts at the level of suffixes or individual adjectives, due to which some originally non-scalar adjectival types or individual adjectives became scalar, and thus new need to grade them arose; this, however, took place at a time when the original intensifying/grading suffixes had ceased to be productive. Consequently, alternative means of gradation emerged (e.g. suppletive comparison, or periphrastic comparison, or intensification by prefixes). However, in the course of time the suffixes -ior- and -issim- became productive again and grammaticalized, while other forms of expressing gradation had been concurrently established. Comparison using magis/maxime was probably more common in Vulgar Latin32 and it must have also gradually grammaticalized, as is illustrated by later Romance languages (the same applies to comparison using plus). Nothing of the kind is, however, detectable in Latin literary texts: for the whole period from beginnings to Late Antiquity synthetic and periphrastic comparatives and superlatives appear as variants,33 but not as complementary variants, i.e. most gradable adjectives allow both forms: suffixes and magis/maxime.34 It is evident that the more highly regarded the author, the fewer periphrastic forms he uses, 32 Cf. also Don. Ter. Eun. 2,227: Animadverte, ut amet Terentius ‘magis’ addere positivo quam comparativum facere. Nam neque ‘severior’ neque ‘continentior’ tam voluit dicere quam ‘magis severum’ et ‘magis continentem’. 33 According to Cuzzolin (2011: 597 ff.) the constructions with magis and plus started to replace morphological comparatives very early on. However, the construction with plus is only very sporadically attested in literary texts: in the whole corpus there are fewer than 50 instances. 34 Thus also Cuzzolin (2011): “However, it must be pointed out that magis and plus were not used only when they were mandatory, such as when the synthetic comparative forms were ruled out on phonological grounds. It is difficult, however, to envisage other reasons with certainty. Certainly the analytical forms were used in poetry for stylistic reasons or to achieve Unauthenticated Download Date | 1/5/19 10:44 PM Periphrastic comparison in Latin 109 which accords with the assumption that periphrastic comparison was one of the distinctive features of Vulgar Latin. These are hypotheses. What is outside the realm of hypothesis, however, is that the reason why there are no attested synthetic comparatives and/or superlatives in those 44 adjectives is not – or at least not primarily – phonological. An element of arbitrariness as to which texts have come down to us probably cannot be ruled out. But the fact is that at least some of these adjectives were certainly ungradable in their original meaning and acquired scalarity only after semantic shifts. Nevertheless, this is not a sufficient explanation, as there are many other such adjectives that went through similar development and still are compared synthetically in Classical Latin. In many concrete examples of attested periphrastic comparatives and superlatives, sentence rhythm or use of other modifiers may have played a role (see note 34). If there are any inner, semantic reasons why an adjective in principle adopts periphrastic comparison, these will only be discovered after a complex semantic analysis of all Latin adjectives, including those compared synthetically. Funding: The article was written as part of the grant project GAČR 17-11247S “Comparison of adjectives in Latin.” References Bartoněk, A. 1955. K problematice latinského superlativu na -issimus. Listy filologické 78. 1–10. Bertocchi, A. & A. Orlandini. 1996. Quelques aspects de la comparaison en latin. Indogermanische Forschungen 101. 195–232. Booij, G. E., CH. Lehmann & J. Mugdan (Eds.). 2000. Morphologie/Morphology. Ein internationales Handbuch zur Flexion und Wortbildung/An international handbook on inflection and word-formation. Berlin & New York: Mouton de Gruyter. Cuzzolin, P. 2011. Comparatives and superlatives. In PH. Baldi & P. Cuzzolin (eds.), New perspectives on historical Latin syntax, vol. 4: Complex sentences, grammaticalization, typology, 549–659. Berlin & New York: Mouton de Gruyter. Cvrček, V. et al. 2015. Mluvnice současné češtiny, 1. Jak se píše a mluví, 2nd ed. Praha: Karolinum. De Vaan, M. 2008. Etymological dictionary of Latin and the other Italic languages. Leiden & Boston: Brill. Dixon, R. M. W. 2004. Adjective classes in typological perspective. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Adjective classes: A cross-linguistic typology, 1–49. Oxford: Oxford University Press. mimetic effects through alliteration, for instance; or in prose for pragmatic effect, by separation of the parameter marker from its parameter” (Cuzzolin 2011: 600–601). Unauthenticated Download Date | 1/5/19 10:44 PM 110 Lucie Pultrová Eklund, S. 1970. The periphrastic, completive and finite use of the present participle in Latin. Stockholm: Almqvist & Wiksells. Fugier, H. 1971. Le superlatif latin: étude fonctionnelle. Revue des études latines 49. 314–330. Fugier, H. 1972. Le système latin des comparatifs et superlatifs. Revue des études latines 50. 272–294. Hofmann, J. B. & A. Szantyr. 1965. Lateinische Syntax und Stilistik. Lateinische Grammatik II. München: Beck. Kühner, R. & F. Holzweissig. 1912. Ausführliche Grammatik der lateinischen Sprache. Erster Teil: Elementar-, Formen- und Wortlehre, 565–566. Hannover: Hahnsche Buchhandlung; H. Pinkster. 2015. The Oxford Latin Syntax. Volume I: The Simple Clause, 47. Oxford: Oxford University Press. Leumann, M. 1977. Lateinische Laut- und Formenlehre. Lateinische Grammatik I. München: Beck. Marvan, J. 2008. České stupňování – Slovanská podoba evropské univerzálie. Praha: Karolinum. Oniga, R. 2014. Latin. A linguistic introduction. Oxford: Oxford University Press. Panhuis, D. G. J. 2006. Latin grammar. Ann Arbor: University of Michigan Press. Pinkster, H. 2015. The Oxford Latin syntax. Volume I: The simple clause. Oxford: Oxford University Press. Szemerényi, O. 1976. Problems of the formation and gradation of Latin adjectives. In A. Morpurgo-Davies & W. Meid (eds.), Studies in Greek, Italic and Indo-European linguistics (offered to Leonard R. Palmer on the occasion of his seventieth birthday, June 5, 1976), 401–424. Innsbruck: Institut fur Sprachwissenschaft der Universität Innsbruck. Touratier, CH. 2013. Lateinische Grammatik. Linguistische Einführung in die lateinische Sprache, übersetzt und bearbeitet von Bianca Liebermann. Darmstadt: Wissenschaftliche Buchgesellschaft. Unauthenticated Download Date | 1/5/19 10:44 PM