«Vicious» or «selected»? The linguosemantics of dispondee ending in Cicero's In Catilinam

Koshevskaya Anna Yur'evna ORCID: 0000-0002-8903-0032 Lecturer, Department of Classical Philology of the Maurice Thorez Institute of Foreign Languages, Moscow State Linguistic University 119034, Russia, g. Moscow, ul. Ostozhenka, 38 castrensiana@mail.ru

I

There is still no single dominant method of study for the rhythmics of ancient Greek and Latin prose. In modern works devoted to this problem, the methodology goes back either to the teachings of F.F. Zelinsky, or to the methodology of A.V. de Groot, or (for French-speaking scientists) to the views of A. Bornek.

The ideas of de are now getting the most widespread Groot and his follower G. Aily, who also became the founder of the method of internal (i.e., not going beyond the limits of a specific text) comparison ^{[1, pp. 32-39]}, which is followed by many researchers of the rhythm of prose, for example, J. Dangel ^{[2, p. 45]}: the method of statistical study of clauses turned out to be the most convenient for machine analysis text. On the other hand, another area of research, the founders of which can be considered F.F. Zelinsky and A. Bornek, namely the study of the rhythm of prose in the aspect of its internal (metric or prosodic) structure, does not lose relevance, which requires a slightly different methodology: so, the main difference in these two approaches, laid down in the works of the beginning of the past It consists in distinguishing either a segment of fixed length as the main unit (which is convenient for statistical research), or a special rhythmic structure – a clause with specific features (length, metric composition, etc.). For example, V. Schmid insisted on the artificiality of this unit ^{[3, p. 2]} and, later, G. Aily ^{[1, pp. 32-33]}; currently, most researchers who do not follow the statistical method or the method of internal comparison are supporters of colometry (dividing the text on a semantic and rhythmic basis) and consider columns to be the minimum rhythmic unit (e.g., B. Streterhoff ^{[4, p. 6]}, Ya. Pall ^{[5, p. 37]}, S. Koster ^{[6, p. 7-12]}, etc.).

The grounds on which this or that rhythmic structure relates or does not relate to clauses are not generally accepted and differ from one researcher to another. For example, a feature of the French-speaking science of the rhythm of ancient prose is the juxtaposition of "metric" and "rhythmic" clauses. This contrast is due to the fact that the French tradition of studying prose rhythm is a direct heir to the doctrine of the cursus, a system of rhythmization of the end of a phrase that replaced clauses in medieval Latin after the loss of vowel lengths and relied on the tonic structure of the last two words of the phrase ^{[7, pp. 163-176]}. For quite a long time, the ancient clauses characteristic of texts in classical Latin were also called cursus by French scientists; in order to distinguish the quantitative rhythm from the tonic rhythm characteristic of medieval Latin, the cursus was called "metric" or "rhythmic", respectively ^{[8, pp. 8-9]}. Such an approach is presented, for example, in the works of A. Bornek ^{[9, p. 2]} and J. Omon ^{[10, pp. 164-183]}. The juxtaposition of these terms was already outlined in the Middle Ages, when there was a need to distinguish the ancient quantitative versification and the tonic that replaced it. For this purpose, the opposition "metrum" and "rhythmus" is formed, in which the first term denoted the quantitative method of versification ("nach antiker Weise gebauter Vers" ^[11]), and the second – tonic ("Gedicht aus akzentuierend gebauten Versen" ^[11]). In this respect, medieval (and then French) terminology differs from classical – ancient Greek and Latin, because the main position of ancient theorists (e.g., Dionysius of Halicarnassus) is that the prosaic rhythm consists of the same meters as the poetic one – that is, these terms originally described phenomena of the same order.

That is why for the absolute majority of works in which the concept of "clause" is used, it is possible to formulate such a definition: any clause necessarily contains a cretic foot (– ? –), which is no further than two syllables from the end of the phrase. In other words, any "standard" clauses can be reduced to one of the following types:

1. ... x x – ? –.

2. ... x – ? – x.

3. ... – ? – x x.

Taking into account all possible combinations of long and short syllables, as well as the dissolutions of long ones, the above schemes really cover the vast majority of clauses of classical Greek and Latin prose.

Of course, all existing theories of prosaic rhythm cannot be reduced to a single thesis; nevertheless, the mandatory presence of a cretic in a clause, regardless of the theory and notation system, was noted by A. Bornek ^{[12, p. 373]}; see also his fundamental work Les clausules m?triques Latines ^[13].

Nevertheless, there are also cases that contradict the above rule and do not correspond to the above schemes. These are clauses that do not contain a cretic, even with long syllables; in other words, these are those cases when the period ends with a sequence of four or five long syllables. In addition, cases with a pronounced alternation of long and short [– ? – ? –] are often referred to as "non-standard" clauses; in addition to the named rhythmic structures, it does not contain a cretic and so on. The "heroic clause" [– ? ? – –], which was considered unsuccessful and fell out of use in antiquity (no later than the I century AD; see Quint. i nst. IX 63–64)..

Although clauses represented by a sequence of long syllables were quite common in ancient Greek and Latin prose and were actively used by authors since the 5th century BC – at least, they are already used regularly by Thucydides ^{[14, p. 181]} – such cases cause certain methodological difficulties for researchers; as a rule, these clauses are left without attention, being considered an unavoidable anomaly (which is typical for English-speaking science, for example, for the works of G. Gatchinson ^{[15, pp. 485-486]}), or are evaluated negatively at all (up to the elimination of the "anomalous" clause by means of conjuncture). Such views were held by A. Bornek (1871-1935), calling such clauses "vicious" (clausulae uitiosae). His opinion is not without reason: indeed, in a homogeneous flow, which is characterized by continuity (continuatio), there can be no clear rhythm – this is evidenced, for example, by Cicero:

Numerus autem in continuatione nullus est; distinctio et aequalium aut saepe uariorum interuallorum percussio numerum conficit, quem in cadentibus guttis, quod interuallis distinguuntur, notare possumus, in amni praecipitante non possumus. (Cic. de or. III. 186)

"In continuity, there is no rhythm; rhythm is created only by clear intervals between accents, intervals equal or even unequal – we notice rhythm in the falling of drops separated by intervals, and we do not notice it in the rapid movement of the stream." (per. M.L. Gasparov)

Bornek argued that in artistic rhythmized prose, dyspondea (four long syllables) as clauses are uncommon, and if they occur in texts, they are necessarily preceded by trochaeus, creticus or tribrachium ^{[12, p. 374]}. On the other hand, Bornek had his own, special view on the specifics of the use of rhythm in prose: the researcher allowed the coexistence of both rhythmized and non-rhythmized, as well as partially rhythmized texts in the corpus of one author ^{[13, 566-567]}; this provision became fundamental in his dissertation ^{[9, pp. 2-3]}.

F.F. Zelinsky (1859-1944) describes these phenomena differently, raising all possible rhythmic sequences to a single model [– ? – | – ?], in which the cretic [– ? –] is the basis of the clause, to which the trochaic cadence is attached ^{[16, pp. 12-17]}. Due to the colossal complexity of F.F. Zelinsky's theory and notation, the description of rhythmic structures in this article has been changed in favor of clarity and ease of understanding; all changes, however, do not contradict Zelinsky's theses in essence.

Zelinsky also divides clauses into "good" and "bad" (more precisely, into "true" (class V, uerae), "permissible" (class L, licitae), "bad" (class M, malae), "worst" (class P, pessimae) and "refined" (class S, selectae) ^{[16, pp. 15-17]}), however, uses completely different criteria and evaluates certain cases differently. Thus, he refers rhythms containing long chains of short ones (class M, malae) or a dactylic element (class P, pessimae) to "bad" clauses, which fully corresponds to the provisions of Cicero, Dionysius of Halicarnassus, Quintilian and other ancient theorists of eloquence. In turn, those clauses that Bornek considers "vicious", Zelinsky refers, depending on the structure, to "permissible" (class L, licitae, rhythm [– ? – ? –] is perceived by him as a cadence (an integral part) of the clause) or to "refined" (class S, selectae), which, according to Zelinsky, should contain a holose (from the Greek. "lame"), i.e. moloss instead of the last cretic clause ^{[16, pp. 16-17]}.

Zelinsky also determines the percentage ratio of each type of clauses in the text, deducing the following. "the Ciceronian formula":

V: 60,3 – L: 26,5 – M: 6,1 – S: 5,2 – P: 1,4 ^{[16, p. 247]}

Although the system proposed by Zelinsky is far from perfect, and mistakes are not uncommon in his calculations, one cannot but admit that his work was in many ways ahead of its time and became a good basis for other concepts; it is also important that, with all the shortcomings of his theory, Zelinsky was able to deduce fair and factual laws of the functioning of clauses, and also get quite reliable statistical data.

II

In his commentary on Cicero's "Catilinaries" ^[17] Bornek notes six "vicious" clauses (although in fact there are quite a lot of such clauses – according to the calculations of the author of this article, there are 45 clauses in total in the "Catilinaries" corpus (taking into account the factur(um) ess? (Cic. Catil. III. 10)), ending with four long ones, which is approximately 6.5% of the total number of clauses in speeches "Against Catilina" (cf. with the data of Zelinsky's "Ciceronian formula" for class S (selectae), given above)), to one of which gives the conjuncture:

d?tr?ment? c?p?ret (Cic. Catil. I. 4) [– – – – // ? ? –] (NB! The above clause is part of a stable legal formula (Senatus consultum ultimum), quoted by Cicero);

praet?rem u?nist? (Cic. Catil. I. 19) [– – – // – – –];

fact?r(um) ess? (Cic. Catil. III. 10) [– – // – ?] >>> esse facturum [– ? // – – –] (NB! Such reading is not attested in the manuscripts (the "ego" conjecture));

iustres numquam (Cic. Catil. III. 23) [– ? – – // – –];

l?t?r ?t (Cic. Catil. IV. 17) [? ? ? // – ? –];

gr?t?l?tnem d?cr?uistis (Cic. Catil. IV. 20) [– ? – ? – – // – – – –].

Of these clauses is only ?l?t?r ?t [? ? ? // – ? –], which Bornek obviously considered to be "vicious" because of the abundance of short.

On the other hand, it is not entirely clear why Bornek left the remaining 39 clauses ending in four or more long syllables without comment; it is also unclear why he proposed a conjuncture only for factur(um) ess? (Cic. Catil. III. 10) [– – // – ?], but left "vicious" clauses iustres numquam (Cic. Catil. III. 23) [– ? – – // – –] (the options offered for the sole purpose of the thought experiment: *numquam iustres [– – // – ? – –], darocha, one of the most common clauses in Czernowski prose – this and further hypothetical (*) conjectures – ego (A. K.)), ?l?t?r ?t (Cic. Catil. IV. 17) [? ? ? // – ? –] (*?t ?l?t?r [– ? – // ? ? x], is quite a common clause very characteristic, e.g., for Pliny the Younger ^{[16, p. 223-224]}) and gr?t?l?tnem d?cr?uistis (Cic. Catil. IV. 20)[– ? – ?– – // – – – –] (*d?cr?uistis gr?t?l?tnem [– – – – // – ? – ? – –], ditroia).

Meanwhile, if you analyse the text "Catilinaria" methodology Zelinsky, you'll find that almost all of the clauses specified by Bennekom as "vicious", belong to the class of "exquisite" (the only exceptions are клаузулыd?tr?ment? c?p?ret (Cic. Catil. I. 4) [– – – – // ? ? –], which falls into the class of "admissible", and ?l?t?r ?t (Cic. Catil. IV. 17) [? ? ? // – ? –], which falls into the class of "bad"). It is noteworthy that there are no discrepancies in the manuscripts for any of the clauses noted by Bornek – i.e. these clauses are quite reliable.

In other words, it is obvious that Cicero quite deliberately interspersed "normal" clauses with dyspondea. But why ?

It seems necessary to consider clauses somewhat differently. For this purpose, the author of the article has developed a new methodology, which is based on the views of Quintilian, according to which the clause consists of two metric stops (Quint. inst. IX. 94-95), as well as Dionysius of Halicarnassus, who described the linguosemantics of each foot (D. H. comp. 17). Based on the nature of the stops included in the clauses, the new methodology distinguishes between primary clauses (with an obvious alternation of long and short ones, composed on the basis of trochaeus and/or creticus, e.g. arb?tr?ris (Cic. Catil. I. 1) [– ? – –]), heavy ones (ending in four or more long ones, for example, n?s ?l?det (Cic. Catil. I. 1) [– – – –]) and derivatives (regularly reproduced rhythmic structures formed from one of the primary clauses, usually through dissolution, e.g. ess? u?dtur (Cic. Catil. I. 14) [– ? ? ? - -] = [– ? – – –]) ^{[18, pp. 146-151]}.

According to our assumptions, "heavy" rhythms had a vivid connotation with the development, continuation of thought, and not with its completion. Thus, of the 45 "heavy" clauses we have identified in the corpus "Catilinary", only one is used at the end of the paragraph (pericopes) (the already mentioned gr?t?l?tnem d?cr?uistis (Cic. Catil. IV. 20)); in the position of the "penultimate" clause of the paragraph, we find only two cases. At the same time, "heavy" clauses end the first period of a new paragraph 12 times (and this is more than a quarter of all cases!). The usual position for clauses of this type is the position of the middle of the paragraph (30 cases or 67%) – that is, the neutral, least marked part of the super-phrasal unity. This indicates that clauses also played a meta-textual role, helping to logically divide the text: where it was necessary to complete a period for reasons of grammar or articulation (after all, the period should be pronounced in one breath), but the thought remained incomplete, the dyspondean clause allowed the speaker to slow down. It is possible that such clauses played the role of modern enumerative intonations in oratorical prose.

In addition, since the rhythm of a sequence of long ones correlates not with the completion of a thought, but with its middle, it is important to note that spondei and molossians (and indeed sequences of long syllables) are frequent harbingers for "standard" clauses: according to our calculations, the entire body of the "Catiline" in the role of the third foot from the end spondei or molossus occur in 203 cases out of 707 (29%); for comparison, creticus in the same position occurs only in 107 cases (15%), trochei – in 71 (10%).

It should also be noted that within this group of clauses, it is necessary to distinguish separate subtypes ^[18]: for example, in "Catilinaria" dyspondea [– – – x] (22 cases or 3%) and combinations of creticus and molossus [– ? – – – x] (14 cases or 2%) are most common; less often – sequences of five long [– – – – x] (9 cases or slightly more than 1%).

Summing up what has been said, we can say that the empirical material testifies to the correctness of F.F. Zelinsky rather than A. Bornek: dyspondei, as well as longer sequences of long syllables, are undoubtedly specific variants of the norm for Latin rhythmized prose, and not subject to correction anomalies. This is also indicated by the statistics of the use of such clauses (in the corpus "Catilinary" the proportion of such clauses is 6.5%: for comparison, Bornek admits the proportion of dispondees in the text of no more than 2.1% ^{[12, p. 374]}, and the above "Ciceronian formula" Zelinsky – 5.2%), and a rather pronounced limitation their use is undoubtedly related to the logical division of the text.