Future paradigms in Latin : Pesky anomaly or sophisticated technique ?

The article deals with paradigms of the future simple (3rd and 4th conjugations only) and the future perfect active that can be treated as anomalous since they form the first person singular and other forms by adding different suffixes to the verbal stem. This entails, first, a certain heterogeneity within the paradigms and, second, a partial overlapping of these paradigms with two other verbal paradigms. Although attempts to unify the future simple and future perfect paradigms were made by archaic authors, Classical Latin has preserved this “inconvenient” distinction, presumably, to highlight the first person singular. The question arises as to why Latin sought to single out the first person singular in this particular way. I will explain this phenomenon as a manifestation of language egocentrism. I will argue that the forms under consideration may function as egocentric devices. Since Latin is a pro-drop language, it requires special means to highlight the speaker as the most significant speech act participant and to give him/ her a privileged status with respect to the other speech act participants. Thus, by using an -am form, the speaker received an additional opportunity to express some modal values better than the other participants did, while with the -ero form, the speaker, conversely, could express his/her thoughts more definitely or unambiguously. In both cases, the singling out of the first person locutor seems to be much more significant for the language as a communicative system than the unified character of the paradigms. The argument is based on an analysis of examples from the works of Latin authors as well as comparative material from other languages.


Introduction 1
In this study, I will focus on two paradigms which can be treated as anomalous since they form the first person singular and the other forms by adding different suffixes to the verbal stem.
The first paradigm is Futurum indicativi (henceforth 'future simple'), the 3rd and the 4th conjugations only. It is formed by adding -a-to the infectum stem in the first person singular and -e-in the other persons. At the same time, the suffix -a-functions as a marker of Praesens coniunctivi (henceforth 'present subjunctive') the first person singular of which is, therefore, homophonous with that of the future simple and, thus, creates ambiguity in some contexts, see Table 1. The second paradigm to be considered is Futurum exactum indicativi activi (henceforth 'future perfect') which is formed by adding -er-to the perfectum stem in the first person singular and -eri-in the other persons. It is well known that the suffix -eri-in Classical Latin functions also as a marker of Perfectum coniunctivi activi (henceforth 'perfect subjunctive') which leads to it being confused with similar forms of the future perfect in all persons but one: the first person singular, see Table 2. From this short observation, it is clear that the Latin language seeks to somehow single out the first person singular, even if it brings about the anomaly of paradigms.
The phenomenon under consideration seems even more intriguing if we take into account the fact that attempts to unify the paradigms of the future simple by using either the -b-suffix or the -e-suffix were made by the early Latin authors and gave rise to the following forms: dicebo, vivebo (Nov. 8;10), exsugebo (Plaut. Epid. 188), sinem (Plaut. Truc. 963, in some manuscripts) (Sihler 1995: p. 558;Tronsky 2001: p. 255;Ernout 2004: p. 192). Such unified forms, however, were rejected in Classical Latin. As regards the unification of the future perfect / perfect subjunctive paradigms, the -ero form was used instead of the -erim form only in Late Latin while Classical Latin kept preserving the distinction between these two paradigms (Pinkster 2015: p. 471). 2 The question arises why the Latin language has preserved the anomaly in both paradigms of the future tenses.
It is worth stressing that in the case of the future simple, the Latin language admits of only one syncretic form in the paradigm while in the case of the future perfect, it allows, on the contrary, all syncretic forms except one, but in both cases, the exceptional form is permitted only for the first person singular.
Interestingly, for each of these "anomalous" strategies, one can find an analogy in other languages. Thus, on the one hand, in Ancient Greek one can find the syncretic form of the first person singular in the future simple and the aorist subjunctive (e.g., τιμήσω). On the other hand, the English language marks differently the first person and the other persons in the future simple (cf. shall and will, respectively). All these phenomena that regularly occur in various languages are unlikely to be random, hence, there is something behind them to be discovered.

Latin future tenses in historical perspective
It seems reasonable to look first at the historical evolution of the forms under consideration and to make clear how the historical grammars explained such anomaly.
The common opinion is that in historical perspective, the future simple was closely related to the present subjunctive whose suffixes -ā-/-ē-it has borrowed (Handford 1946: p. 39;Hofmann & Szantyr 1972: p. 309;Sihler 1995: p. 557;Baldi 1999: p. 398;Tronsky 2001: pp. 250, 255;Ernout 2004: p. 191). These two suffixes go back to different archaic paradigms of the subjunctive: the first one was constructed with the suffix -ā-, and the second one -with the inherited PIE long thematic vowel -ē-"leveled analogically from an original PIE -ē-/-ō-alternation which is evident in Greek" (Baldi 1999: p. 398). In Classical Latin, the -ā-paradigm was used for the present subjunctive (e.g., dicam, dicas, etc.), while the -ē-paradigm -for the future simple (e.g., dico, dices, etc.). However, since the first person singular form of the future simple therefore turned out to be homophonous with that of the present indicative (dico), it has been eventually replaced by the form of the present subjunctive (dicam). This is how some authors of the historical grammars explain the heterogeneous forms of the future simple paradigm (Palmer 1988: p. 231;Sihler 1995: pp. 558, 595;Ernout 2004: p. 192). Other scholars, however, underline the obscure origin of the first person singular -am form (Baldi 1999: p. 398). 3 In my opinion, the explanation put forth by the authors of the historical grammars is not convincing. First, a similar homophony of the first person singular in the future simple and aorist subjunctive is found in Ancient Greek (cf. τιμήσω) but did not cause the replacement of one form by another one borrowed from a different paradigm, as it happened in Latin. Second, the Latin language had indeed the possibility to unify the paradigm of the future simple, and such attempts, as I have shown, were made by the archaic authors but could take root neither at the archaic nor at the later stages of the Latin language.
As regards the suffixes of the future perfect and the perfect subjunctive, they proved to have no direct relation to each other. In the suffix -eri-, which in Classical Latin was common for both tenses, 4 the final vowel -i-was historically of different origin. Thus, in the future perfect paradigm, it is classified as thematic vowel -i-which goes back to the PIE vowel -e-/-o-, while in the perfect subjunctive paradigm, -i-(< -ī-) goes back to the suffix of the ancient optative mood (Baldi 1999: p. 403;Ernout 2004: p. 255;Pinkster 2015: p. 462). As for the element -er-, historically it is the preterite morpheme *-is-, also present in other perfective paradigms, e.g., in the perfect indicative (cf. amav-is-ti, etc.) (Baldi 1999: p. 403). 5 In Classical Latin, these two suffixes of different origin merged into one common suffix -eri-which therefore was used as the marker of both tenses in all persons but one: the first person singular (Tronsky 2001: p. 291). The explanation of why this form stands out from the paradigm is not found in the historical grammars of Latin.

Morphological affinity and semantic correlation between future and subjunctive
The morphological affinity of the future simple and the present subjunctive manifests itself at the semantic level, too. Baldi (1999: p. 400) points out, that functionally, the subjunctive was used to express volition and reservation (doubt) about some future events and, therefore, "had a future orientation". Pinkster (1990: p. 226) suggests that "the future is often used with a so-called 'modal' nuance. In other words, the predication is formulated as referring to the future and as a rule does, in fact, have future reference, but the attitude of the speaker with regard to the predication is such, that the hearer does not interpret it in an exclusively temporal way. Statements in the future concerning a first 3 Cf. Sihler (1995: p. 558): "The exact source of -am, earlier -ām, is however a mystery". 4 See, e.g., Pinkster (2015: p. 462): "from roughly Cicero's time onwards, the future perfect indicative forms and the perfect subjunctive forms were no longer morphologically distinct, except in the passive and in the first person singular (tulero versus tulerim, respectively)".
(6) Sit nox cum somno; sit sine lite dies! (Mart. 2, 90, 10) "Let it be a night with sleep, let it be a day without a quarrel!" (desire). Significantly, such a formal and semantic correlation between future and subjunctive does exist in many languages: the future tenses can express will, an impulse to action, obligation, possibility and other modal nuances, since they are often grammaticalized expressions of intention (Plungian 2011: p. 434). Therefore, it does not come as a surprise that markers of the future tenses can display modal values of intention, exhortation or will in modern English (Greenbaum 1996: pp. 259, 262;de Haan 2012: p. 126), French (Mosegaard Hansen 2016: pp. 105-106), Russian (Stojnova 2016), and other languages.
To sum up, the semantic zones of future and subjunctive are overlapping significantly, but not to the point that one category is completely replaced by another, since each of them, in addition to a common semantic zone, has its own specific function that can be expressed by no other means. For this reason, they do in fact coexist in languages and preserve reasonable distribution of the markers in all persons but one -the first person singular. 6 The examples with translations are taken from Pinkster (1990: p. 226).
7 Pinkster (2015: p. 323) underlines the importance of grammatical person in assessing the communicative goal of the information presented in the simple future tense: "When the speaker is talking about his own future actions or states, his statement is most likely taken as a declaration of his intention. When, in turn, the speaker is talking about the addressee, the statement is most likely to be taken as a prediction or an instruction. Finally, when the speaker is talking about a third person, the statement is most likely to be taken as a prediction. In general, given the nature of 'futurity' itself, the simple future indicative is less assertive than the other tenses and has some uses which resemble those of the present subjunctive". In my opinion, the range of values conveyed by the future simple and present subjunctive is even broader and includes evidential and mirative overtones as well (Zheltova 2018: pp. 232-233).

Language egocentrism: a theoretical framework
In what follows, I will try to explain this phenomenon as a manifestation of language egocentrism. Generally speaking, egocentricity is an immanent property of human language associated with its subjective character. According to E. Benveniste, language as such is marked with manifestation of subjectivity so deeply that the question arises whether it could function and be called so if it were structured in another way. 8 The fact that the language allows each speaker to appropriate an entire language when he/she refers to himself/herself as a speaker 9 is precisely due to the egocentric inventory that exists in any language. The theory of egocentrism was developed in detail in the works of K. Bühler (1965), E. Benveniste (1976), R. Jakobson (1984), Yu. Apresjan (1995), B. Uspensky (2007) The egocentric inventory usually includes deictic or modal elements of a language such as personal pronouns, tenses, moods etc. (Benveniste 1976: p. 262;Jakobson 1984), because it is through the deictic elements that the speaker relates the statement to the moment of speech and expresses his/her ownsubjectiveattitude to the content of the statement. Thus, personal pronouns have been considered as prototypical deictic words since ancient grammarians onwards, 10 verbal tenses have also deictic reference since the action expressed by a verbal form always correlates with the moment of speech (Uspensky 2007: p. 13). As for the category of mood, this is, in fact, a "grammaticalized modality" (Plungian 2011: p. 423), that is, one of the main "egocentric" mechanisms of natural languages which allows not only to describe the world as it is, but to represent a subjective image of the world -the world perceived through the prism of the speaker's consciousness (Plungian 2011: p. 424).
Among the egocentric elements of any language, the priority without a doubt is given to the category of person. The first person singular, by definition, is the most subjective and egocentric, since it is from the angle of the first speech act participant that utterance is generated.
It is noteworthy in what terms Apollonios Dyskolos, a grammarian of the second century A.D., justified the priority of the first person over the other persons: "...because what other persons say comes from him (sc. from the first person singular)" (Polikarpov 2007: p. 98).
The special status of the first person locutor (speech act participant) may have various manifestations in languages. Thus, E. Benveniste (1976: pp. 264-265) drew attention to the fact that some groups of verbs have different semantics in the first person singular than in the others: these are the verbs of mental operations (suppose, conclude, assume, 8 "II est marqué si profondément par l'expression de la subjectivité qu'on se demande si, autrement construit, il pourrait encore fonctionner et s'appeler langage" (Benveniste 1976: p. 261).
etc.) and some verbs of speaking that are now classified as performatives (swear, promise, guarantee, certify). The verbs of such type take completely different meaning in the first person than in the others: while I swear is an obligation, he swearsjust a description of an action similar to that of He runs or He smokes (Benveniste 1976: p. 265).
The study of the Russian verbs in the way suggested by Benveniste was conducted by E. Paducheva (2008: pp. 136-142) who singled out several groups of verbs that are not compatible with the first person singular pronoun and, therefore, highlight its special status. This approach may be applied to any language and presumably will give similar results.
In modern linguistics, egocentrism as part of a broader domain of subjectivity has proved to be a topical issue. The very notion of egocentrism seems to partly overlap with the notion of egophoricity, as it is treated in San Roque et al. (2018: p. 2): "At its very broadest, egophoricity is a general phenomenon of linguistically flagging the personal knowledge, experience, or involvement of a conscious self; it can furthermore be understood as differential linguistic marking of 'privileged access' to a real or mentally projected activity or state".

Privileged status of the first speech act participant
In view of these observations, the special verbal forms of the first person singular in Latin may be considered as egocentric/egophoric devices. Since Latin is a pro-drop language, it requires special means to highlight the speaker as the most significant speech act participant and to give him/her a privileged status or privileged access with respect to the other speech act participants. This observation raises a new question: what does the privileged status mean or, in other words, what are the advantages of the particularly marked first person singular forms as compared to the others? In my opinion, the advantages each of the particular markings gives to the first speech act participant depends on the communicative purpose of the speaker.

Neutralization of the tense/mood opposition and irrealis
Let us first consider what communicative challenge the Latin language meets by using the syncretic -am-forms. The syncretism of the -am-forms entails, as we remember, the neutralization of the future simple and the present subjunctive just in these forms. Strange as it may seem, a neutralization that reduces the opposition in a given categorial feature may create another categorial feature which can be very important for a language. 11 Based on this assumption, I will argue that in the forms under consideration, the neutralization of the tense/mood opposition creates a new categorial feature of irrealis in the sense suggested by T. Givón (1994) 12 and maintained by many other scholars. Thus, according to J. Elliott (2000: pp. 66-67), a proposition is realis if it asserts that a state of affairs is an "actualized and certain fact of reality", whereas it is classified as irrealis if "it implies that a state of affairs belongs to the realm of the imagined or hypothetical, and as such it constitutes a potential or possible event but it is not an observable fact of reality". Although the existence of irrealis as a grammatical category is a highly debatable issue, 13 in a number of works the notion of irrealis is applied to the forms "that encode some type of unrealized states of affairs, that is, states of affairs that are not presented as positively occurring or having occurred. 'Irrealis' in this sense is a descriptive label for particular forms, roughly equivalent to more traditional terms such as, e.g., future, subjunctive, or conditional" (Cristofaro 2012: p. 131). In this vein, the notion of irrealis allows for a broader range of modal values: from a higher degree of certainty through probability/presupposition/possibility to a lower degree of certainty.
Interestingly, Givón (1994: p. 270) has presented the correlation between tense/aspect and modality as follows: Past/perfective => realis (or presupposition) Perfect => realis (or presupposition) Present-progressive => realis Future => irrealis Habitual => irrealis or realis From this observation, it is obvious that future unambiguously correlates with irrealis, and this allows us to assume that neutralization of the tense/mood opposition in the -amforms brings about a new categorial feature of irrealis and thus provides the speaker with an additional opportunity to express some modal values better than the other speech act participants do.

Egocentric functions of the -am forms
I will try to corroborate my hypothesis by referring to the texts of the Latin authors, in particular, to the contexts in which Cicero used the form dicam. I will restrict my analysis to the occurrences of dicam in independent clauses, to cut off the uses of dicam in the subordinate clauses where they could be determined by other reasons. Look at ex. (7): (7) Aspendum vetus oppidum et nobile in Pamphylia scitis esse, plenissimum signorum optimorum. Non dicam illinc hoc signum ablatum esse et illud: hoc dico, nullum te Aspendi signum, Verres, reliquisse (Cic. Verr. 2, 1, 53) "You know that Aspendus is an ancient and noble town in Pamphylia, full of very fine statues. I do not say (C. D. Yonge) / I shall not allege (L. H. G. Greenwood) / I would not say (mine -E. Zh.) 14 that one or another statue was taken away from thence: this I say, that you, Verres, left not one statue at Aspendus." 15 The three translations above show that dicam is not unambiguous. Although this -amform can be treated as a statement in the simple future, I prefer to see here the present subjunctive that conveys Cicero's unpreparedness or unwillingness to say (a sort of a hedging strategy). Significantly, the form dico allows us to better understand the meaning of dicam: these two forms seem to be opposed to each other on the modal axis rather than on the temporal one, as the opposition of uncertain vs. categorical statements. My interpretation doesn't rule out the possibility, however, of seeing a rhetorical omission here, too. 16 Quite a different pragmatic context is given in ex. (8) where dicam is rhetorically repeated twice, and the adverbial modifier alio loco indicates a certain point in the future when Cicero is going to say about Lucullus: 17 (8) Sed de Lucullo dicam alio loco, et ita dicam, Quirites, ut neque vera laus ei detracta oratione mea neque falsa adficta esse videatur (Cic. Manil. 10, 8) "However, of Lucullus I will speak hereafter, and I will speak, O Romans, in such a manner, that his true glory shall not appear to be at all disparaged by my pleading, nor, on the other hand, shall any undeserved credit seem to be given to him." In ex. (9), dicam looks as ambiguous as in ex. (7) but the presence of the hedging adverbial modifier paene argues in favor of the subjunctive rather than of the future: (9) Siqua enim sunt privata iudicia summae existimationis et paene dicam capitis, tria haec sunt, fiduciae tutelae societatis. (Cic. Q. Rosc. 16) 14 The translations provided clearly demonstrate that translators are unlikely to pay much attention to the semantic nuances of grammatical forms that, unfortunately, sometimes leads to inaccuracies or even a misinterpretation of meaning. Concerning the importance of careful approach to the translation of each word in the Classical text, cf. D'Angour (2019).
"For if there are any private actions of the greatest, I may almost say, of capital importance, they are these threethe actions about trust, about guardianship, and about partnership." In ex. (10), again, Cicero used the form (non) dicam as a sort of polite, mitigating assertion which is characteristic of the potential subjunctive rather than of the future simple: (10) At in ipsum Habitum animadverterunt. Nullam quidem ob turpitudinem, nullum ob totius vitae non dicam vitium sed erratum. (Cic. Cluent. 133) "Oh, but they visited Habitus himself with their censure. Not for any baseness, nor for any, I will (C. D. Yonge) / would (mine -E. Zh.) not say vice, but not even for any fault of his own in his whole life." On the contrary, in ex. (11), non dicam goes along with the future perfect protulero which argues in favor of its interpretation as the future simple: (11) Qua de re tota si unum factum ex omni antiquitate protulero, plura non dicam (Cic. Cluent. 134) "And with respect to this whole business, if I produce one precedent from the whole of our ancient history, I will say no more." I would say that the -am-forms can be definitely treated as the future simple only when used along with the future perfect forms or in other contexts where the tense/aspect opposition is concerned, as in ex. (12). The temporal meaning of dicam is backed up by the adverbial modifier semper: (12) Maxima voce ut omnes exaudire possint dico semperque dicam (Cic. Sull. 33) "I speak, and I always will speak, with my loudest voice, in order that all men may be able to hear me." As is clear from the observation above, the -am-form demonstrates quite a few meanings at once and thus makes the addressee experience mixed feelings about the proposition (distance, doubt, lack of confidence, etc.). Such polysemy may be part of the speaker's hedging strategy to make a statement "weakly asserted as either possible, likely or uncertain" (Givón 1994: p. 268) or provide other modal meanings which belong to the domain of irrealis. One cannot rule out that the -am-form repeated twice can exemplify a particular rhetorical technique as well.
To sum up, the syncretism of the -am-forms entails a neutralization of the opposition between the future simple and the present subjunctive. It may be considered as an egocentric marker which highlights a special status of the first speech act participant by adding a whole array of nuances to his/her utterance.

Egocentric potential of the -ero/-erim forms
Now I will turn to the analysis of the -ero/-erim-forms. Given all the forms of the future perfect and the perfect subjunctive are syncretic with the only exception of the first person singular, one could suppose the distinction between -ero/-erim forms to be preserved by the same reason: the language was keen to single out the first locutor but in this case by preserving the opposition, i.e. through completely opposite technique. Whereas neutralization can be called a technique of gluing, the preservation of the opposition -a technique of scissors. 18 Now we will consider what communicative task could be accomplished by means of this technique.
I have analyzed all the occurrences of dixero and dixerim in the Cicero's corpus. 19 It is worth stressing that dixero occurs 28 times while dixerim -59. The comparison of these results does not seem senseless and clearly shows that the perfect subjunctive is used two times as often as the future perfect. The reason for such a distinction is probably that the future perfect -as opposed to the perfect subjunctive -performs mainly the taxis functions in subordinate clauses and does not have modal nuances, ex. (13): (13) Ego autem si omnia quae dicenda sunt libere dixero, nequaquam tamen similiter oratio mea exire atque in volgus emanare poterit (Cic. Rosc. 3).
"But if I freely say all the things which must be said, yet my speech will never go forth or be diffused among the people in the same manner." Conversely, the perfect subjunctive is used predominantly with modal meaning both in independent and in subordinate clauses including those where the subjunctive mood is not determined by grammar rules, and, consequently, signals modal overtones. Out of 59 occurrences of dixerim in Cicero's works, 34 (more than a half) were used in the function of the potential subjunctive, as in exx. (14-16): (14) Citius dixerim iactasse se aliquos ut fuisse in ea societate viderentur… (Cic. Phil. 1, 25) "I should sooner say that some men had boasted in order to appear to have been concerned in that conspiracy..." (15) Ibi est ex aere simulacrum ipsius Herculis, quo non facile dixerim quicquam me vidisse pulchrius (Cic. Verr. 2,4,94) "In it there is a bronze image of Hercules himself, than which I cannot easily tell where I have seen anything finer." (16) Omnibus fere in rebus sed maxime in physicis quid non sit citius quam quid sit dixerim (Cic. nat. deor. 1,60) "In almost all things, but most of all in natural ones, I should sooner say what does not exist than what does exist." 20 These examples suggest that the speaker could express his/her thoughts more definitely and unambiguously through the -ero-form which is restricted to only temporal contexts without modal connotations, rather than through the -erim-form which almost always has modal overtones. Seemingly, the need to express such a distinction overtly in the first person singular prevented the paradigms of the future perfect and the perfect subjunctive from the complete unification.

Concluding remarks
I tried to show that special forms of the first person singular in the paradigms of the future simple and the future perfect may function as egocentric devices in Latin. Using such devices, Latin highlighted the speaker as the most significant speech act participant and gave him/her a privileged status with respect to the other speech act participants. In the case of the -am-form, which may feature the overtones of uncertainty or subjectivity, the speaker received an additional opportunity to express some modal values better than the other participants do. In the case of the -ero-form, the use of which is restricted to only temporal contexts without modal connotations, the speaker, conversely, could express his/her thoughts more definitely or unambiguously. In both cases, the singling out of the first person locutor seems to be much more significant for the Latin language as a communicative system, than the unified character of the paradigms.