Hostname: page-component-74656f9884-rdgjk Total loading time: 0 Render date: 2026-11-06T21:15:15.906Z Has data issue: false hasContentIssue false
  • English
  • Français

A NOTE ON THE STYLE OF THE BELLVM HISPANIENSE. COLLOQUIALISMS AND ALLUSIVE SELF-FASHIONING?

Published online by Cambridge University Press:  17 October 2016

T.S. Allendorf
T.S. Allendorf*
Affiliation:
University of Oxford
*
tobias.allendorf@classics.ox.ac.uk
Rights & Permissions [Opens in a new window]

Extract

The Bellum Hispaniense (= BHisp.), often considered alongside the two other pseudo-Caesarian commentarii, the Bellum Alexandrinum and the Bellum Africum, contains the account of Caesar's final battles with Pompey's forces in Spain during 46 and 45 b.c. In terms of its language and style, the BHisp. is certainly the most un-Caesarian of the pseudo-Caesarian Bella. The work abounds in phenomena belonging to different linguistic varieties, and is an important piece of evidence in the process of the standardization of written Latin.

Information

Type
Research Article
Information
The Classical Quarterly , Volume 66 , Issue 2 , December 2016 , pp. 547 - 555
Check for updates
Copyright
Copyright © The Classical Association 2016 

The Bellum Hispaniense (= BHisp.), often considered alongside the two other pseudo-Caesarian commentarii, the Bellum Alexandrinum and the Bellum Africum, contains the account of Caesar's final battles with Pompey's forces in Spain during 46 and 45 b.c. In terms of its language and style, the BHisp. is certainly the most un-Caesarian of the pseudo-Caesarian Bella. The work abounds in phenomena belonging to different linguistic varieties, and is an important piece of evidence in the process of the standardization of written Latin.Footnote 1

The most recent scholarly treatment of the BHisp., Jan Felix Gaertner's painstaking consideration of the style of the auctor Belli Hispaniensis, attempts a reclassification of the work.Footnote 2 Before Gaertner's study, the style of the BHisp. had been classified as ‘vulgar’, ‘substandard’ and ‘non-literary’, with considerations of genre not allowing for verdicts beyond ‘hastily written soldier's diary’ or similar.Footnote 3 Questioning the communis opinio, Gaertner sets out to reclassify the style of the BHisp. as a Latin ‘pre-standard’,Footnote 4 and provides a new answer to the question of the work's genre, confidently placing it within the Roman historiographical tradition of the late annalists. Gaertner's re-evaluation of the BHisp. is motivated by three initial considerations: (1) some stylistic features and quotations from Ennius proving the educational status of the writer, ‘obviously not … an uneducated soldier’;Footnote 5 (2) the mutilation of the text due to its transmission;Footnote 6 (3) a methodological point, according to which there was not yet a clearly established ‘standard’ of Latin at the time when the author of the BHisp. was writing—hence describing his style as ‘substandard’ is based ‘on a retrojection of the classical norm’, which emerged only in the writings of Caesar and Cicero.Footnote 7 In the bulk of his article, then, Gaertner provides a valuable re-examination of many phenomena in the language of the BHisp. which were hitherto attested as ‘colloquial’, ‘vulgar’ and ‘substandard’. He variably finds the style and language of the author not to be substandard and not as colloquial as previously thought. Moreover, he argues that some linguistic features point to a conscious archaising tendency, with some features even being ‘poeticizing’ and ‘effecting casualness’, which is, according to Gaertner, in line with the style of the Roman late annalists (such as Coelius, Quadrigarius and Sisenna).Footnote 8

The present paper aims to complement Gaertner's study; it takes its impetus from two main initial queries.Footnote 9 First, some of the colloquialisms that are briefly noted but otherwise granted little attention in Gaertner's study are significant not only for the language of the BHisp. but also for the development of Latin more generally and need to be given due attention.Footnote 10 Secondly, I harbour some scepticism whether the linguistic evidence presented by Gaertner is sufficient for assuming conscious imitation of the style of the late annalists in the BHisp.

The aim of the present study will thus be twofold: it will trace significant but underrepresented colloquialisms in the style of the BHisp. (the use of quod after verba dicendi et sentiendi, the use of coepisse plus infinitive as a substitute for the perfect, and the use of bene with adjectives), and will then return to a methodological point about the classification of the likely genre of the BHisp.

‘Colloquial language’ is a fuzzy term in scholarship, describing a rather broad continuum, and usually refers, as Adams suggests, to ‘a multitude of phenomena with different distributions and different degrees of acceptability’, which is why ‘many assertions on the subject are valueless’.Footnote 11 In the following analysis, I will adhere to a method that favours the criterion of the distribution of phenomena in passages over psychological considerations and explanations: the latter are only helpful for understanding a specific colloquial phenomenon, but not to prove that it is one in the first place. Thus, I will use the term ‘colloquial’ to refer to phenomena if their distribution before, after and (if possible) roughly at the time of the composition of the BHisp. shows them to be non-literary.Footnote 12

QUOD-CLAUSES INSTEAD OF THE ACCUSATIVE + INFINITIVE

One of the most intriguing syntactic features in the language of the auctor Belli Hispaniensis is the use of quod after verbs that would require the infinitive construction in classical Latin. The phenomenon, appearing for the first time in a Latin prose text at BHisp. 10.2 (suo loco praeteritum est quod equites ex Italia cum Asprenate ad Caesarem uenissent,Footnote 13 ‘it was omitted in its proper place that the cavalry had come to Caesar from Italy with Asprenas’), is of utmost importance with regard to the Romance languages. It gains ground in later authors and becomes the dominant construction in Late Antiquity and Christian texts.Footnote 14 It is plausible to infer from both older and later passages that the phenomenon was not unusual in spoken Latin.Footnote 15 The colloquial character of the construction is particularly evident at Plaut. Asin. 52–3, where it features in an utterance of Demaenetus, a destitute and immoral old man, providing his opinion about his son's love-affair with what, in his view, is a less-than-respectable woman (equidem scio iam filius quod amet meus istanc meretricem, ‘Well, I already know that my son is in love with that prostitute’), and at Petr. 131.7, where we have the direct speech of a little old woman excitedly commenting on the success of her aphrodisiac witchcraft (‘uides’ inquit ‘Chrysis mea, uides, quod aliis leporem excitavi?’, ‘“Do you see, my dear Chrysis, do you see that I have started a hare for others?”’). The colloquial character of the quod-construction in the Petronian passage can be corroborated further by noting that it occurs in close proximity of another colloquialism, coepi + infinitive (admotisque manibus temptare coepit inguinum uires, ‘having moved her hands there, she began to test the strength of my groin’, Petr. 131.5), a feature to be discussed in some detail below.Footnote 16

The occurrence of the quod-clause instead of the infinitive construction in the BHisp. hints, moreover, at the construction having emerged independently from any influence of Greek ὅτι in Latin.Footnote 17 Occasionally, we find quod instead of accusative + infinitive in Cicero as well: after a preparatory demonstrative pronoun, addere, mittere and verbs of silently passing over things (such as praeterire, transire).Footnote 18 Since there is one piece of evidence for praeterire quod in Cicero's Pro Cluentio, the phrase is normally classified as ‘Ciceronian usage’.Footnote 19 This is, however, not quite right. A comparison of BHisp. 10.2 praeteritum est quod … uenissent and Cic. Clu. 188 praetereo quod … sibi domum sedemque delegit, ‘I pass over the fact that she … chose for herself as her house and residence …’, reveals an essential difference: while Cicero has perfect indicative (delegit), the author of the BHisp. has pluperfect subjunctive (uenissent). In the use of the oblique subjunctive he clearly indicates that he understands the content of the quod-clause as a reported opinion: it is a substitution of the infinitive construction normally required after verba dicendi et declarandi. If we accept Davies's conjecture, as Klotz does, this is the first time we encounter this phenomenon in Latin prose. At this point of the BHisp., we are on the threshold of the development towards the Romance languages.Footnote 20

The author's usage of the quod-construction after verbs of saying can be corroborated further. At 36.1, a textually undisputed passage, he writes, legati Carteienses renuntiauerunt quod Pompeium in potestatem haberent, ‘envoys from Carteia reported that they had Pompeius in their hands’. Note that, in the clause dependent on renuntiauerunt, the author again uses the subjunctive of reported opinion. The quod-clause is not facultative but required. If 10.2 were not accepted, BHisp. 36.1 would be the first textually secure instance of quod after a verb of saying.Footnote 21

COEPISSE + INFINITIVE INSTEAD OF THE PERFECT

Another striking linguistic feature in the BHisp. is the writer's predilection for using coepisse + infinitive instead of the perfect. The distribution of the phenomenon throughout extant Latin prose shows that occurrences of the phenomenon cluster in two works, the BHisp. and Petronius’ Satyricon Libri.Footnote 22 For the practice in Petronius, Hubert Petersmann has provided an exhaustive discussion, bringing out the colloquial character of the construction.Footnote 23 It is generally agreed that the construction usually indicated an ingressive aspect in Latin, by functioning in the same way as the Greek aorist.Footnote 24 In Greek, aoristic periphrasis is particularly common in New Testament Greek.Footnote 25 The BHisp., however, shows that the phenomenon in Latin is not a Graecism.

Moreover, the construction is merely phraseological in the BHisp.: it is used interchangeably with the simple perfect and without any marked difference in meaning. Take, for example, the following instances, following each other in quick succession (by no means representing a comprehensive list): Caesar munitionibus Ateguam oppugnare et bracchia circumducere coepit, ‘Caesar began to assault/assaulted Ategua with siege works and surrounded it with fortified lines’ (6.3); castellum oppugnare coepit, ‘he began to attack/attacked the fort’ (9.1); iactus facere coeperunt, ‘they began to launch missiles/they threw missile weapons’ (9.2); cum ex castello repugnare coepissent, ‘when they had begun to fight back from the fort’ (9.3); iter facere coepit, ‘he proceeded to march’ (10.3); cum eruptionem facere coepissent, ‘when they had begun to make a sally/had made a sally’ (12.6).  It is reasonable to assume that in the course of time mere pleonastic usage interfered with the original ingressive function. This is apparent in the BHisp.: passages such as qui fidens uiribus ex aduersariis sibi parem esse neminem agitare coepit, ‘and he, confident in his strength, and that there was nobody a match for him on the enemies’ side, began to taunt again and again’ (25.3), are especially striking, since the frequentative verb and the ingressive function of the periphrastic coepisse seem to contradict each other.Footnote 26 Such a usage is uncommon in Caesar: where he has coepisse plus the infinitive instead of the simple perfect, the ingressive aspect tends to be important.Footnote 27 In the Bellum Hispaniense, the construction is ubiquitous—unlike in Caesar, Cicero, and, as far as one can tell, the first-century historiographers cited by Gaertner, who argues for a conscious usage following the precedent of the ‘late annalists’.Footnote 28 The distribution, moreover, makes a conscious imitative usage unlikely in general: the construction of coepisse plus infinitive occurs quite often throughout Latin.Footnote 29 Thus, Gaertner's view, arguing for a conscious usage in imitation of annalistic practice, is open to doubt: whereas he points to the late annalists, one may wish to give prominence to the occurrences of the phenomenon in Plautus (and Terence), and claim a colloquial origin of the construction.Footnote 30 What is certainly significant, however, is the unreflected usage of coepisse plus infinitive in clusters in the BHisp.: such clusters are only to be found here, in the BHisp., in Petronius and in very late Latin.Footnote 31 The ubiquitousness and the occurrence in clusters can be taken as indication that the author of the BHisp. does not employ coepisse plus infinitive for stylistic variatio or in conscious imitation of earlier practice. Rather, coepisse + infinitive may already have become merely phraseological in spoken Latin,Footnote 32 and from there made its way into the written medium.

BENE AS AN INTENSIFIER INSTEAD OF THE SYNTHETICAL COMPARATIVE

The colloquialisms in the BHisp. also include the practice of using bene as an intensifier with adjectives or participles, in constructions such as bene magnum tempus, ‘a good long time’ (12.4) and bene multis uulneribus adfecti, ‘having sustained a good many wounds’ (12.6). This usage of bene had a colloquial ring throughout Latin (unlike the use of magis and maxime in this function):Footnote 33 its distribution shows particularly high frequency in Plautus and in letters.Footnote 34 In addition, bene in itself seems not to have been considered a high-register word: apart from its occurrences in the common collocations bene mereri and bene gerere, it does not feature in Caesar at all.

In the BHisp., bene is very common, especially in its use as an intensifier of adjectives. It occurs most frequently in the phrases bene magnus and bene multus.Footnote 35 In addition, the author uses uehementer as an intensifier of adjectives; uehementer seems to be preferred with adjectives expressing something undesirable, such as in uehementer periculosus, ‘extremely dangerous’ (30.2), and uehementer erat saucius, ‘he was seriously wounded’ (38.2), perhaps indicating a tendency that has survived in its reverse in French, where bien seems to be almost exclusively used with verbs and adjectives expressing some virtue or something desirable.Footnote 36

While it is true that the auctor Belli Hispaniensis is also capable of using the superlative,Footnote 37 it must be noted that the construction with bene does not have any function of marked language. This becomes clear when one isolates individual examples: without any consideration of the context, it is, for instance, always bene magnus and bene multus that are used, and never the corresponding synthetical forms maximus and plurimus. Moreover, the distribution of the phenomenon does not contribute to a stylistic embellishment of the work: consider passages such as BHisp. 12, where bene magnus is almost immediately followed by another bene (this time, bene multi). Rather than viewing this phenomenon as an attempt at achieving ubertas, or as an imitation of the ‘casual’ practice of the ‘late annalists’,Footnote 38 one should take it is a colloquialism—one that had become so frequent in spoken language that it also found its way into the written medium with less accomplished authors.

ENNIAN REMINISCENCES AND LINGUISTIC REGISTER: ALLUSIVE SELF-FASHIONING?

Having considered some of the linguistic evidence that is underrepresented in Gaertner's treatment, I would now like to revisit the second of his initial considerations outlined above. There, Gaertner argues that the Ennian quotations in the BHisp. point towards the writer's higher educational status and, generally, square well with the archaising tendency he detects in the work. Clear Ennian reminiscences occur twice in the BHisp., at 23.3 and at 31.7. While the quotation at 31.7 (hic, ut ait Ennius, pes pede premitur, armis teruntur arma ,Footnote 39 ‘Hereupon, as Ennius says, foot was pressed on foot, weapons were grinding against weapons’) at least occurs in a significant battle-scene, the citation in the BHisp. may have been facilitated by the proverbial status of the polyptotic phrases. One can, with Pascucci, refer to similar proverbs such as manus manum lauat at Petr. 45.13.Footnote 40 At BHisp. 23.3, it is even more evident that the Ennian reminiscence does not contribute to stylistic embellishment. The citation (hic tum, ut ait Ennius, nostri cessere parumper , ‘Hereupon, as Ennius says, our men gave ground a little’) draws too much attention to a rather insignificant passage: at the beginning of section 23, Caesar's men are engaged in an everyday military operation of no particular importance, ‘moving a line of fortifications to the river Salsum’ ([Caesar] … bracchium ad flumen Salsum ducere coepit, 23.1). Note also that the author of the BHisp. does not change the Ennian hexametrical line-ending (cessere parumper). This would be awkward at the end of a sentence in polished prose. By contrast, the author of the Bellum Africum and Livy change the phrase to parumper cesserunt (BAfr. 52.2) and parumper cessere (Livy 26.43.3) respectively.Footnote 41 The issue is clear: since one might expect a higher level of compositional awareness in historiography, can Ennian reminiscences such as the hexametrical quotation at 23.3 be taken as part of an aspiring historiographer's self-fashioning strategy, or do they rather bear witness to the writer's memories of a standard school curriculum and his lack of polish?

On a more general level, the mixture of different varieties and the abundance of significant colloquial features in the BHisp. raise questions about Gaertner's classification of the work as a piece of historiography in the tradition of Roman late annalists, a view that Gaertner supported mainly with linguistic evidence.Footnote 42 What Gaertner argues for is a type of self-conscious allusion by an author who ‘consciously placed himself in the tradition of (mildly) archaising and poeticizing historiography’.Footnote 43 Can a certain linguistic phenomenon, however, especially in a type of prose that does not seem to be very polished, confidently be considered as part of allusive self-fashioning within a tradition? Methodological problems need to be acknowledged on at least three levels. First, it is questionable whether one can infer conclusions about the genre of a work from the level of linguistic register alone. Second, even if we were to accept such a methodology for detecting an allusive practice, we face a further problem on the other side of the allusive coin, that is, the texts the writer of the BHisp. is said to allude to—our knowledge of the Roman late annalists is fragmentary at best, and we have no way of knowing about the distribution and frequency of phenomena in their writings.Footnote 44 There is a third methodological query: any decision about which linguistic features of the BHisp. are to be considered relevant for the question of genre and which are not could arbitrarily distort the full picture. In statistical terms, the totality of tokens under scrutiny is referred to as the ‘population’. Just as philologists do, statisticians may sometimes wish to study only part of the ‘population’, a so-called ‘sample’, but, for the results to be representative, this must be assembled with great care from the totality of tokens. If we follow the rules of statistics, Gaertner's conclusions about the genre of the BHisp. are questionable, since his sample does not fulfil the criteria of being representative and balanced.Footnote 45 I will not attempt to draw stronger conclusions for the question of genre from my sample size (the three colloquial phenomena discussed above), but hope to have highlighted further complexities.

CONCLUSIONS

In sum, Gaertner's new genre classification of the BHisp., essentially based on the writer's educational status (quoting Ennius) and on the archaising linguistic features shared in some fragments of the late annalists, is too confident. While the archaisms may or may not imitate the practice of the Roman late annalists,Footnote 46 they certainly occur within a linguistic register that is more colloquial than Gaertner allows: some of the neglected features—the use of quod-clauses instead of the accusative + infinitive, of coepisse + infinitive instead of the perfect, and of bene as an intensifier of adjectives—have been discussed in the present article. Many scholars have taken the BHisp. quite literally as a commentarius, one that, unlike Caesar's commentarii, was not primarily fashioned as literary art.Footnote 47 I would be loath to abandon the scholarly communis opinio because of the evidence presented in Gaertner's study, however much his thorough analysis has advanced our understanding of the style of the auctor Belli Hispaniensis.Footnote *

References

1 On which, see Petersmann, H. and Petersmann, A., ‘Standardisierung und Destandardisierung der lateinischen Sprache in diachroner Sicht’, in Hornung, A., Jäkel, C. and Schubert, W. (edd.), Studia Humanitatis ac Litterarum Trifolio Heidelbergensi dedicata. Festschrift für Eckhard Christmann, Wilfried Edelmaier und Rudolf Kettemann (Frankfurt am Main, 2004), 235–54Google Scholar, and Neumann, G., ‘Sprachnormung im klassischen Latein’, Sprache der Gegenwart 2 (1968), 88–97 Google Scholar. For a brief linguistic overview, cf. also Clackson, J. and Horrocks, G., ‘The background to standardization’, in id., The Blackwell History of the Latin Language (Oxford, 2004), 77–89 Google Scholar.

2 Gaertner, J.F., ‘The style of the Bellum Hispaniense and the evolution of Roman historiography’, in Dickey, E. and Chahoud, A. (edd.), Colloquial and Literary Latin (Cambridge, 2010), 243–54CrossRefGoogle Scholar.

3 Gaertner (n. 2), 243, esp. nn. 2 and 4, for an overview of earlier scholarly positions that were not in favour of the style of the BHisp., beginning with Scaliger's indictment (‘horrida … dictio’). Twentieth-century classicists ranging from Löfstedt, E., Philologischer Kommentar zur Peregrinatio Aetheriae. Untersuchungen zur Geschichte der lateinischen Sprache (Uppsala, 1911), 8Google Scholar to Henderson, J., ‘XPDNC / Writing Caesar (Bellum Civile)’, ClAnt 15 (1996), 261–88Google Scholar, at 274 have also agreed on the non-literary character of the BHisp.

4 Gaertner (n. 2), 251.

5 Gaertner (n. 2), 244.

6 As Gaertner himself notes, the mutilation of the text can explain only very few oddities in the language of the BHisp.: in effect, it is only responsible for rather isolated phenomena such as the genitive-absolute constructions at 14.1 and 23.5 and some fragmentary sentences. See Gaertner (n. 2), 244–5, with n. 12.

7 Gaertner (n. 2), 245.

8 See esp. Gaertner (n. 2), 252–3 for his examples of stylistic affinities.

9 There is one further, but non-essential, query; it pertains to Gaertner's suggested terminology of a Latin ‘pre-standard’. While I agree that a classification of the style of the BHisp. as ‘substandard’ is methodologically problematic, I am nevertheless hesitant to endorse Gaertner's alternative option: his term ‘pre-standard’ all the same implies the subsequent existence of a ‘standard’. I am not sure whether Caesar's and Cicero's standard actually became the accepted norm of writing Latin afterwards (before Renaissance Humanism): the majority of post-Ciceronian Latin prose deviates from the alleged standard, with Seneca and Tacitus epitomizing the development. See also n. 1.

10 Gaertner (n. 2), 250 states that ‘[i]n the end, few usages can be justly called colloquial or substandard’, but nevertheless has to list nine phenomena as clear colloquialisms.

11 Adams, J.N., ‘The Bellum Africum ’, in Reinhardt, T., Lapidge, M. and Adams, J.N. (edd.), Aspects of the Language of Latin Prose (Oxford, 2005), 73–96 CrossRefGoogle Scholar, at 86.

12 Particularly reliable evidence for the non-literary spoken character of linguistic phenomena are occurrences in Plautus (and, with some caution, Terence), in letters and in many passages in Petronius: cf. Hofmann, J.B., Lateinische Umgangssprache (Heidelberg, 1978), 2–5 Google Scholar, with A. Chahoud, ‘Idiom(s) and literariness in classical literary criticism’, in Dickey and Chahoud (n. 2), 42–64, at 47, and the remarks at Reinhardt, T., Lapidge, M. and Adams, J.N. (edd.), Aspects of the Language of Latin Prose (Oxford, 2005), 7CrossRefGoogle Scholar, esp. n. 8. On the difference between vulgar and spoken Latin in particular, Löfstedt (n. 3), 8–13 is still worth consulting.

13 Klotz adopts John Davies's conjecture, which still makes for the best text. I have printed Davies's original conjecture without Klotz's prodelision (which seems to be unparalleled elsewhere in the BHisp.). Böhm's eccentric argument for praemissus nuntius est quod ( Böhm, G.R., ‘Aus einem Aenigma werden durch Dummheit nur Fehler. Textkritisches zum Bellum Hispaniense 10,1.2’, RCCM 30 [1988], 81–95 Google Scholar, at 85–6) does not avoid the perceived awkwardness of quod, and Davies's praeterire quod at least has the authority of (the slightly different) Cic. Clu. 188 praetereo quod … sibi domum sedemque delegit and the similar constructions at BHisp. 36.1 (and perhaps 18.5: see n. 21).

14 See the passages discussed in G. Mayen, De particulis quod, quia, quoniam, quomodo, ut pro acc. cum infinitivo post verba sentiendi et declarandi positis (Diss., Kiel, 1899), 17–24, and the overview of the distribution of the infinitive construction vs quod (quia/quoniam) at Perrochat, P., Recherches sur la valeur et l'emploi de l'infinitif subordonné en latin (Paris, 1932), 146–7Google Scholar. Cf. also the further evidence discussed by Adams, J.N., ‘The accusative + infinitive and dependent quod/quia-clauses. The evidence of non-literary Latin and Petronius’, in Kiss, S., Mondin, L. and Salvi, G. (edd.), Latin et langues romanes: études de linguistique offertes à József Herman à l'occasion de son 80ème anniversaire (Tübingen, 2005), 195–206 CrossRefGoogle Scholar.

15 See Mayen (n. 14), esp. 15–16. Mayen also provides the passages from Petronius discussed above, but does not comment on their colloquial character.

16 On the colloquial character of coepi in Petronius in particular, see the discussion by Petersmann, H., Petrons urbane Prosa. Untersuchungen zu Sprache und Text (Syntax) (Wien, 1977), 189–93Google Scholar.

17 This is implicitly endorsed by Adams (n. 14), 197, who suggests that the vulgarisms in Petronius mainly consist in the ‘perversion of the more “educated” construction by the analogical replacement of the correct subordinator quod with quia’. It is also one of the conclusions of Cuzzolin, P., Sull'origine della costruzione dicere quod. Aspetti sintattici e semantici (Florence, 1994)Google Scholar.

18 Note that there is no such usage in Caesar. The only possible exception may be Caes. BCiv. 1.23.3 pauca apud eos loquitur, quod sibi … gratia relata non sit, where the neuter adjective prepares for the quod. See Pascucci, G., Bellum Hispaniense. Introduzione, Testo Critico e Commento (Florence, 1965), 207Google Scholar, ad BHisp. 10.2.

19 Gaertner (n. 2), 250 n. 68.

20 On which, see Herman, J., La formation du système roman des conjonctions de subordination (Berlin, 1963)Google Scholar.

21 The passage is thus identified at Löfstedt (n. 3), 117. If we accept Koehler's conjecture, an easy and attractive change that unfortunately has not made it into Klotz's apparatus, we have another instance at BHisp. 18.5: duo … fratres transfugae nuntiarunt quod Pompeius contionem habuisset; see Koehler, A., De auctorum Belli Africani et Belli Hispaniensis Latinitate (Erlangen, 1877), 77Google Scholar.

22 See Adams, J.N., Social Variation and the Latin Language (Cambridge, 2013), 825–7CrossRefGoogle Scholar.

23 See n. 16.

24 See the discussion at Löfstedt, E., Syntactica II. Syntaktisch-stilistische Gesichtspunkte und Probleme (Malmö, 1956 2), 450–2Google Scholar.

25 On Greek ἄρχομαι + infinitive, see Hesseling, D.C., ‘Zur Syntax von ἄρχομαι und Verw.’, ByzZ 20 (1911), 147–64Google Scholar.

26 See also BHisp. 22.4 manus intentare coeperunt and 39.2 in speluncam Pompeius se occultare coepit.

27 See e.g. Caes. BGall. 5.7.5 at impeditis omnium animis Dumnorix cum equitibus Haeduorum a castris insciente Caesare domum discedere coepit and BCiv. 3.97.2 Pompeiani, quod is mons erat sine aqua, diffisi ei loco relicto monte uniuersi iugis eius Larisam uersus se recipere coeperunt. The only passages that I would regard as borderline cases which hint at a phraseological character of the construction are BGall. 3.23.2 and 4.27.7. Cf. Gaertner, J.F. and Hausburg, B., Caesar and the Bellum Alexandrinum. An Analysis of Style, Narrative Technique, and the Reception of Greek Historiography (Göttingen, 2013), 56CrossRefGoogle Scholar n. 112.

28 Gaertner (n. 2), 252, esp. n. 79. On the ubiquitousness of the phenomenon in the BHisp. and its colloquial character, see also Petersmann (n. 16), 191.

29 See Adams (n. 22), 826.

30 See e.g. Plaut. Epid. 248–9 hoc quod actum est. egomet postquam id illas audiui loqui, | coepi rursum uorsum ad illas pauxillatim accedere; Ter. Phorm. 75 coepi aduorsari primo: quid uerbis opust?; Phorm. 78–9 … coepi is omnia | facere, obsequi quae uellent …; Phorm. 82 hanc ardere coepit perdite; and Cic. Fam. 7.5.1 coepi uelle ea Trebatium exspectare a te, quae sperasset a me, with Kühner, R. and Stegmann, C., Ausführliche Grammatik der lateinischen Sprache. Zweiter Teil: Satzlehre (München, 1962 4), 570Google Scholar.

31 For the clusters in the BHisp. see the occurrences cited above and, for another example, 31.4 premere coeperunt and traduci coepta sit, 31.5 premere coepit, 31.7 agere coeperunt. The occurrence in clusters is identified by Adams (n. 22), 827. For an example of the distribution in Petronius, Adams points to 74.8 osculari diutius coepit, 74.9 male dicere Trimalchioni coepit, 74.12 gemere ac flere coepit.

32 Klotz calls the phenomenon ‘phraseologisch in vulgärer Sprache’: Klotz, A., Kommentar zum Bellum Hispaniense (Leipzig and Berlin, 1927), 38Google Scholar. See also Hofmann, J.B. and Szantyr, A., Lateinische Syntax und Stilistik (München 1972 2), 329Google Scholar (‘vor allem umgangssprachlich’).

33 The use of magis/maxime as intensifiers is classical with most adjectives ending in -uus and -ius. See further Hofmann and Szantyr (n. 32), 165.

34 For the colloquial character of bene + adjective see, by way of example, Plaut. Capt. 966 bene morigerus fuit puer; Cas. 596 ut bene uociuas aedis fecisti mihi; Lucil. 268 Marx bene plena; 1073 bene longincum … morbum; Cic. Fam. 12.13.4 classis bene magna, and cf. the representative distribution at TLL II.2125.53–2126.80, s.v. bonus.

35 bene magnus: 1.2, 4.2, 12.4, 13.4, 15.6, 16.2, 22.6, 35.2, 35.3, 40.2, 41.1; bene multus: 12.6, 34.6, 36.4; cf. also bene longe, 25.2.

36 See Wölfflin, E., ‘Lateinische und romanische Komparation’, in Meyer, G. (ed.), Ausgewählte Schriften (Leipzig, 1933), 126–92Google Scholar, at 140.

37 See e.g. firmissimus, 6.1; crassissimus, 6.3; uehementissime, 13.7.

38 For this view, see Gaertner (n. 2), 251.

39 The Ennian hexameter could thus be reconstructed as <et> premitur pede pes <atque> armis arma teruntur: see Norden, E., Ennius und Vergil (Leipzig, 1915), 158Google Scholar. Vahlen's Teubneriana has <et iam> pes premitur pede <et> armis arma teruntur: Vahlen, J., Ennianae poesis reliquiae (Leipzig, 1928 2)Google Scholar.

40 Pascucci (n. 18), ad loc.

41 See again Pascucci (n. 18), ad loc. For a sophisticated Ennian allusion, using metre in his favour, cf. Tacitus’ archaising Ennian hexameter with which he starts the Annals.

42 As he himself notes, Gaertner essentially builds here on Eden's argument about the stylistic relationship between Caesar's commentarii and the Roman late annalists: see Eden, P.T., ‘Caesar's style: inheritance versus intelligence’, Glotta 40 (1962), 74–117 Google Scholar.

43 Gaertner (n. 2), 251.

44 The fragments and testimonia are now well presented and commented upon in Cornell, T.J. et al. (edd.), The Fragments of the Roman Historians (Oxford, 2013)Google Scholar.

45 See e.g. Gries, S.T., Statistics for Linguistics with R: A Practical Introduction (Berlin, 2009), 24 CrossRefGoogle Scholar, who instructs: ‘it is important that you choose your sample such that it is representative and balanced with respect to the population to which you wish to generalize. Here, I call a sample representative when the different parts of the population are reflected in the sample, and I call a sample balanced when the sizes of the parts in the population are reflected in the sample’ (italics in the original).

46 Until the opposite has been shown, I remain unconvinced that they do. Rather, I would subscribe to the view that most linguistic phenomena of the BHisp. discussed by Gaertner and in the present paper, just as the archaisms in the Bellum Africum which Adams (n. 11), 79 discusses, ‘represent rather the last throes of an older way of writing. The writer need not have been consciously using archaism as stylistic adornment’.

47 See n. 3.

* This article developed out of a paper written for a seminar on the BHisp., run by Professor Gerrit Kloss. I am grateful to him and to Professor Tobias Reinhardt for their comments on an early version. My thanks also to my wife Kalina, Dr Barnaby Taylor and CQ’s anonymous reader and editorial team, Professors Bruce Gibson and Costas Panayotakis, for their help and suggestions at a later stage.