The language of Agathias Scholasticus (c. 530–80), like many other aspects of his literary production, has never been the object of extensive research. Averil Cameron laid the groundwork for future inquiry, devoting an article and a substantial portion of her 1970 Agathias – to date the only monograph dedicated to this authorFootnote 1 – to matters of language and style.Footnote 2 Before Cameron, scholarship on the topic was scarce, limited essentially to the studies of Teuffel,Footnote 3 Van Herwerden,Footnote 4 Reffel,Footnote 5 and Franke,Footnote 6 in addition to Keydell’s and Costanza’s Index graecitatis in their respective editions, as well as a few pages of ‘Sprachliche Bemerkungen’ published by Keydell.Footnote 7
The late 1960s were a relatively fertile period for the study of Agathias: in addition to Cameron’s work, Keydell’s new critical edition of the Histories appeared,Footnote 8 followed by that of Costanza only two years later.Footnote 9 Around the same time, Agathias’ poetic production also began to attract attention: McCail devoted several valuable studies to the epigrams and had planned an edition that ultimately was never published.Footnote 10 In a brief but useful article that appeared in 1968, he catalogued the poetic reminiscences in the Histories, demonstrating the marked predominance of epigrammatic poetry in the first two books, which were written soon after the publication of the Cycle.Footnote 11 More recently, two articles have revisited the possible influence of Agathias’ poetic expertise on the Histories.Footnote 12 A short study by Adshead has explored Agathias’ knowledge and use of Thucydides, though its arguments, despite the promising topic, remain underdeveloped.Footnote 13 Two recent translations of the Histories into modern Greek and Spanish also include useful linguistic notes and remarks.Footnote 14
Cameron’s findings remain the most valuable foundation for anyone wishing to explore Agathias’ language in greater depth. With her book, she contributed to a preliminary general assessment of Agathias’ language by identifying three main features: a) an archaizing patina, achieved through typical Attic usages (such as duals, datives, and pluperfect tenses); b) the presence of poetic (Homeric, Nonnian, epigrammatic) vocabulary; c) what she termed ‘Byzantinism’: a consistent inclination, beneath the classicizing veneer, towards stylistic elements that distinguish late Greek from its ancestor (lexical abundance, a tendency to excess and periphrasis).Footnote 15
While Cameron’s insightful conclusions on Agathias’ language – particularly his debt to classical models, his position between archaism and Byzantinism, and the tension between classicism and affectation – remain significant, some of her judgments, especially those based solely on classical Greek standards, no longer hold up to scrutiny today. One of the central arguments in Cameron’s monograph is that Agathias’ performance as a historian, his language and style included, would have been better had he been an abler writer.Footnote 16 Yet this assessment overlooks the possibility that his linguistic choices were the result not of inability but of intent. As Martin Hinterberger has observed, ‘Byzantine classicising language diverges from Ancient Greek norms not primarily because Byzantine authors were incapable of following those norms, but because of their conscious decision not to do so.’Footnote 17
This perspective aligns with recent developments in the study of medieval Greek – particularly its learned (or so-called classicizing) register – and provides a crucial framework for my analysis. It allows for a reassessment of Agathias’ language beyond the lens of classical deficiency, interpreting it instead as a deliberate stylistic negotiation. Geoffrey Horrocks’ work similarly emphasizes that learned Byzantine Greek, rather than being a static dead variety, continued to ‘develop over time, not only according to [its] own internal dynamics, e.g. through analogical extensions of inherited rules and principles, but also because the speakers who use [it] tend, however thorough their training, to reconceptualize traditional elements of grammar in contemporary terms’.Footnote 18 According to Horrocks – and as the following discussion will attempt to show through selected examples – such developments occurred particularly in the domains of syntax and semantics, where prescriptive norms were often lacking, especially by comparison with morphology.Footnote 19
Recognizing high-register Byzantine Greek not as a flawed approximation of Ancient Greek but as a self-standing linguistic variety with its own grammatical rules and innovations is essential to advancing the study of this language. The rigid distinctions that continue to be made between spoken language, lowbrow literature, and highbrow literature are thus usefully challenged by Horrocks, as seen also in his analysis of the language of Georgios Akropolites.Footnote 20 Through a case study of expressions of futurity and modality in both contemporary vernacular texts and Akropolites’ Histories, Horrocks convincingly demonstrates that thirteenth-century high-register prose was not merely ‘a “bodged” and “artificial” version’ of Ancient Greek, but a dynamic entity influenced by contemporary linguistic structures.Footnote 21
Building on this framework, I adopt a similar perspective on Agathias. Rather than viewing his prose – and that of other Byzantine authors writing in a high register – as varying degrees of failure to replicate the norms of fifth- or fourth-century BCE Attic Greek, I examine it as an instance of sixth-century classicizing prose that nonetheless integrates elements of linguistic evolution. This classicism, however, did not merely reflect an orientation toward classical Athens but was shaped by a broader and historically layered tradition, including the Atticizing models of the Second Sophistic and late antique rhetorical education. Through two specific case studies of so-called non-classical usage, I aim to offer new insights into the relationship between literary tradition and language change by exploring how high-register literary Greek in this period actively engaged with ongoing linguistic developments. What has in the past been labelled as ‘non-classical’ or ‘incorrect’ may in fact reflect regular features of sixth-century learned Greek. Rather than judging Agathias’ prose against an idealized version of ‘Classical’ Greek, I consider how his style negotiates between literary models and contemporary linguistic reality in order to reveal the evolving nature of high-register prose.
First case study: πλὴν ἀλλά
This section examines a syntagm frequently encountered in Agathias’ Histories: the combination of the particles πλήν and ἀλλά.Footnote 22 This syntagm is not discussed in The Cambridge Grammar of Medieval and Modern Greek,Footnote 23 and a preliminary TLG search confirms that it is rarely found in the fifth and fourth centuries BCE.Footnote 24 It begins to surface during the imperial period and becomes increasingly common throughout the Byzantine centuries.Footnote 25 By tracing its usage diachronically and examining its function in the Histories, I argue that πλὴν ἀλλά is employed by Agathias to signal a shift in topic while simultaneously introducing a contrast or opposition. For this reason, I will refer to its function as ‘transitional-adversative’. While this dual meaning was already conveyed by the individual particles πλήν and ἀλλά, their combination – possibly originating in the spoken language – becomes characteristic of this function.
In Classical Greek, in addition to its main limitative value (‘except’, ‘apart from’), πλήν can also function as a transition term, indicating a change of subject after a strong pause, and often has adversative value.Footnote 26 Although limited examples of this usage can be found already in Herodotus (Hist. 7.32), Plato (Prot. 328e), and Polybius (1.69.13–14), a large and definitive spread does not occur until the imperial period. In the New Testament πλήν not only acts ‘as a balancing adversative particle, the equivalent of δέ or μέντοι’,Footnote 27 for example in Mt. 28.7, but it can also have a function that is ‘progressive rather than exceptive or adversative […] meaning “moreover,” or “and indeed”’,Footnote 28 as attested in Lk. 10.14 and 19.27.Footnote 29 It thus seems that by the first century CE the particle πλήν had taken on this transitional and adversative value.
In Hellenistic Greek, πλήν had started, albeit rarely, to adopt a transitional function, occasionally reinforced by ἀλλά. This combination, which would later become typical of Byzantine Greek, is already present in Agathias’ prose. Prior to the imperial period, πλὴν ἀλλά is attested only sporadically: four times in the Hippocratic Corpus,Footnote 30 once in Theophrastus,Footnote 31 once in the fragmentary playwright Machon,Footnote 32 and twice in Aristophanes of Byzantium.Footnote 33 Isolated occurrences are also found in Dionysius of Halicarnassus, Strabo, and Plutarch. However, it is only in the second century CE that the syntagm becomes well-attested. Most notably, πλὴν ἀλλά appears 82 times in the works of Lucian; 79 of these instances occur at the beginning of a sentence. This striking concentration in Lucian’s prose, and especially in his dialogues, suggests that the expression was associated with the rhythms and register of spoken language and was employed to mimic conversational shifts.Footnote 34
From an analysis of its occurrences, we see that during roughly the same period πλὴν ἀλλά entered (perhaps still as an element of the spoken language) genres such as philosophical treatises and biographies, as witnessed by five occurrences in Sextus Empiricus,Footnote 35 five in Clement of Alexandria,Footnote 36 and four in Flavius Philostratus.Footnote 37 Agathias’ use of the syntagm may reflect this literary inheritance. Rather than drawing exclusively on classical Attic norms, Agathias – like many classicizing authors of the Byzantine period – probably also took stylistic cues from imperial prose, and particularly from the idiom of the Second Sophistic. A phrase like πλὴν ἀλλά may thus have been recognized not as a deviation from classical usage, but as a legitimate – and even high-register – element of classicizing prose.
Πλὴν ἀλλά at the beginning of a sentence is also widely attested in the Church Fathers from the third to early fifth century, with 84 occurrences in Eusebius, 22 in Gregory of Nyssa, 42 in Basil the Great, and more than 200 in John Chrysostom, to name just a few. While the widespread use of the phrase in homiletic works could suggest a link with spoken-style communication, its distribution is by no means limited to sermons: it also appears in letters and treatises. Although more rarely, it is also attested in high-register writers such as Libanius,Footnote 38 Synesius,Footnote 39 and Isidore of Pelusium; the latter employs the phrase fourteen times in his correspondence,Footnote 40 thirteen of which occur after a strong pause. Over the fifth and sixth centuries, the syntagm becomes increasingly common. We find seven occurrences in the rhetor Procopius of Gaza, seven in the ecclesiastical encomia of Chrysippus, and 21 in Justinian’s Novellae, reinforcing the idea that Late Antiquity was a period of consolidation for this usage in Greek prose. While it would be superfluous to list all occurrences from the sixth century onward – the expression became typical of Byzantine Greek and is well attested across genres, in prose as well as poetryFootnote 41 – it is worth pausing to review the results so far before turning to Agathias and other sixth-century historians.
Πλὴν ἀλλά originated as an expanded form of one of the rarer classical uses of πλήν – the transitional use. It can be rendered in English as ‘and yet’, ‘however’, or ‘but’. Likely originating in the spoken language, it entered Greek prose primarily through the genres of dialogue and philosophical treatises and is extensively present in patristic literature. The phrase continued in use throughout the fourth and fifth centuries and eventually became a regular feature of prose in later centuries. In the overview above, I have deliberately left aside Agathias and other classicizing historians, for they deserve to be looked at separately. Although the syntagm was well attested by Agathias’ time, it remains uncertain whether it had become fully naturalized in learned prose or was still perceived as a spoken or informal feature. Analysing its usage in Agathias and his contemporaries can help clarify this question and shed light on their stylistic choices.
To begin with, the syntagm πλὴν ἀλλά never occurs in any of Procopius’ three works. Only a single instance is attested in the fragments of Menander the Protector (or Guardsman),Footnote 42 Agathias’ historiographical continuator, suggesting that more examples may have existed in the now-lost portions of his text. No instances are found in John Malalas’ Chronographia. In this light, the 32 occurrences in Agathias’ Histories – 28 of which appear at the beginning of a sentence – stand out as particularly significant. A closer analysis reveals that Agathias uses the combination of the two particles in the contemporary and already typical function I have identified: to start a new sentence while maintaining an oppositional or contrastive link with the preceding one – although some examples are more sharply adversative than others. Out of the 32 total occurrences of πλὴν ἀλλά in the Histories, I will focus on three particularly illustrative passages to explore this usage in greater depth. These examples show that the expression was not restricted to a single stylistic mode or textual function but could operate flexibly – linking sentences, marking transitions in topic, or emphasizing contrast within an elevated rhetorical texture.
The first example comes from a narrative passage in Book 2, where Agathias describes the treacherous terrain through which the Roman army must march. The landscape is presented as almost impassable:
τὰ γὰρ ἐχόμενα πεδία ἰλυώδη γέ εἰσι δεινῶς καὶ τελματώδη, λόχμαι τε δασεῖαι καὶ δρυμοὶ ἀνέχουσιν, ὡς καὶ ἑνὶ ἀνδρί, καὶ τούτῳ εὐζώνῳ, ἄπορον εἶναι τὸ χρῆμα, μή τί γε πλήθεσιν ὡπλισμένοις. πλὴν ἀλλὰ καὶ ὣς οἱ Ῥωμαῖοι οὐδὲν ὁτιοῦν πόνου ἀνιέντες… (Agathias, Histories 2.19.4 [K. 65.26–29])Footnote 43
The surrounding plains are tremendously muddy and swampy, with dense thickets and groves rising up, so that they are impassable even to a single man – even someone light-armed – let alone to an army of heavily equipped soldiers. But even so the Romans by no means gave in to difficulties…
While καὶ ὣς conveys the concessive nuance (‘even so’), πλὴν ἀλλά introduces the sentence and ensures continuity while also marking the shift to a contrasting idea. Frendo’s English translation seems to overlook the phrase, reducing its transitional weight.Footnote 44 Its presence, however, reflects Agathias’ care in maintaining rhetorical fluidity while allowing for adversative contrast, reinforcing the heroic perseverance of the Romans despite the natural obstacles. In translating πλὴν ἀλλά, I have opted for the English ‘but’, which captures the basic adversative force of the expression. However, this choice inevitably simplifies the Greek. The syntagm functions in a richer way: not merely marking opposition but also mediating transitions, highlighting topical shifts, and slowing the rhythm to allow emphasis or reflection. In English, such combined effects are difficult to replicate with a single conjunction. Even a more elaborate rendering, such as ‘and yet’ or ‘still, however’, cannot fully convey the layered rhetorical role πλὴν ἀλλά plays in Agathias’ prose.
The second example offers a more clearly adversative use, framed in a passage that reports the direct speech of the Lazic Aeetes during an assembly after the murder of the Lazic king Gubazes by Roman generals:
πρόσεστι γὰρ αὐτοῖς ὡς ἀληθῶς καθάπερ οἰκεῖα καὶ συγγενῆ καὶ τὰ τοιαῦτα ὀνείδη· πλὴν ἀλλὰ τοῖς ἐκ φύσεως μοχθηροῖς τὸ αὐθαίρετον ἀδίκημα προστεθὲν ἐδιπλασίασεν αὐτοῖς τὸ ἀτύχημα τῆς τοῦ κρείττονος προμηθείας διὰ τὸ ἄγος ἐκπεπτωκόσιν. (Agathias, Histories 3.10.1 [K. 95.15–19])
And indeed, outrages of this kind belong to them as natural and innate characteristics; but it was the addition of voluntary fault to the wickedness of their nature that doubled the failure, seeing as they had fallen out of divine protection owing to the sacrilege.
Here, the syntagm sets up an emphatic contrast not just in terms of logic, but in moral responsibility: the rhetorical structure reinforces the idea that natural vice is bad but deliberate vice is worse. While the adversative function is dominant, the transitional role remains: it connects two related observations while intensifying the judgement. A simple ἀλλά could have sufficed, but the expanded πλὴν ἀλλά both elevates the register and slows the shift, lending gravity to the transition.
The third and perhaps most sophisticated example comes from an excursus in Book 4, where Agathias compares Chosroes favourably to the ancient Persian kings:
…καὶ γέγονεν ὁποῖος οὔπω πρότερον ἄλλος τῶν παρὰ Πέρσαις βεβασιλευκότων ἀναδέδεικται, εἴ γε τῷ παντὶ ἐκάστῳ συγκρίνοιτο, οὐδὲ εἰ Κῦρον εἴποι τις ἂν τὸν Καμβύσου οὐδὲ Δαρεῖον τὸν Ὑστάσπεω, οὐδὲ μὴν Ξέρξην ἐκεῖνον, τὸν ἱππήλατον μὲν δεικνύντα τὴν θάλατταν, ἐν δὲ τοῖς ὄρεσι ναυτιλλόμενον. πλὴν ἀλλὰ τοιοῦτός γε ὤν, ἀκλεής γε αὐτῷ ἡ τοῦ βίου καταστροφὴ γέγονε καὶ οἰκτρὰ καὶ τῶν φθασάντων ἀλλοτριωτάτη. (Agathias, Histories 4.29.6–7 [K. 161.1–7])
… and he was such a man as no other of those who had reigned among the Persians had ever proved to be, if indeed he were compared with every one of them – not even if someone were to name Cyrus son of Cambyses, nor Darius son of Hystaspes, not even that famous Xerxes, who rendered the sea passable on horseback and sailed in the mountains. And yet, despite being such a man, the end of his life was inglorious and miserable and enormously different from what had happened before.
Here, πλὴν ἀλλά perfectly displays its twofold transitional-adversative role: it closes the comparison with his distinguished predecessor, while marking the beginning of a new theme – Chosroes’ end – which is related to the previous one – his greatness – in an oppositional way. He was better than Xerxes and yet died miserably.
This overview, however brief, shows that Agathias considered πλὴν ἀλλά suitable for diverse sections of his work: the three passages include speeches, digressions, and narrative. He thus did not regard the syntagm as a feature confined to spoken or informal language but as fully compatible with the texture of high-register literary prose. Its absence in Malalas, whose chronicle altogether lacks such elevated stylistic features, supports this reading. Rather than indicating conscious innovation, Agathias’ use of πλὴν ἀλλά likely reflects broader developments in cultivated Greek, whereby structures of earlier spoken origin had, by the sixth century, become normalized in written prose – particularly through influential models such as those found in Second Sophistic and patristic literature. At the same time, its complete absence in Procopius suggests that authors committed to a stricter, more archaizing form of Atticism continued to avoid such expressions. By contrast, Agathias’ style appears more open and flexible. The second case study, to which we now turn, will explore another instance of this stylistic openness.
Second case study: the future optative
In the Index graecitatis of his edition of the Histories, Keydell included the heading optativus futuri, noting a total of fourteen occurrences. Reffel likewise observed Agathias’ penchant for this form.Footnote 45 The decision to highlight the future optative as a notable aspect of Agathias’ language was likely motivated by two factors. First, the optative mood, already in decline during the Hellenistic period, had largely disappeared from spoken Greek by the sixth century, to be replaced by the subjunctive.Footnote 46 Second, and perhaps more importantly, the optative of the future tense was never extensively used in Classical Greek.Footnote 47 Its function was limited to expressing an original indicative future in subordinate clauses (typically after verbs of saying, thinking, fearing, and in conditionals).Footnote 48 This was decisively demonstrated as early as 1884 by Fassbänder, in what remains the most comprehensive study of the tense in Classical Greek.Footnote 49 From its first appearance in Pindar (Pythian 9.115–16) to Demosthenes, Fassbänder counted a total of 254 occurrences, of which 141 alone appear in Xenophon, by far the most prolific user of the tense.Footnote 50 While these figures might be adjusted slightly with the help of modern databases such as the TLG, they remain substantially valid.
In Hellenistic Greek, the future optative occurs only rarely and always in very limited numbers – typically fewer than ten instances per author, as in Polybius, Posidonius, and Diodorus Siculus; these low numbers reflect the concurrent decline of both the optative mood and the synthetic future.Footnote 51 Of the few instances of the optative mood in the New Testament, none is in the future tense.Footnote 52 This situation does not change significantly in the imperial period. The future optative remains marginal but is attested in Dionysius of Halicarnassus and Strabo, and with somewhat greater frequency in Flavius Josephus. However, it must be remembered that the morphological overlap between certain aorist and future optative forms (e.g. πέσοι) can make the form difficult to identify with certainty.Footnote 53
Given the results of the first case study, it is particularly relevant to consider the use of this tense among the authors of the Second Sophistic. Here, too, it appears sparingly – typically no more than ten to twenty times per author – in writers such as Dio Chrysostom,Footnote 54 Aelius Aristides,Footnote 55 and, with more occurrences, in Philostratus.Footnote 56 I could find no certain occurrences in Lucian. Its presence in these authors is likely due to Atticizing affectation. The presence of this tense in patristic literature, also with non-classical, expanded uses, is well attested and has been extensively studied.Footnote 57 Neil O’Sullivan was the first to place this development in a broader linguistic context. In his 2013 study, he demonstrated that grammatical papyri and wooden tablets dated from the third to the fifth century CE show that late antique students were still actively learning and conjugating the future optative.Footnote 58 These texts confirm the morphological survival of the tense but offer little insight into its perceived syntactic function. The tense is always introduced by the particle εἰ, possibly to emphasize its primary conditional nuance, but no further explanation is provided.
O’Sullivan extended his analysis beyond grammatical texts to include documentary papyri of legal content – such as property sales and transactions, wills, and contracts – from the sixth and seventh centuries. In these documents, he identified at least fifteen occurrences of the future optative used with new, non-classical values, thus complementing the theory reflected in the grammatical papyri with practical examples.Footnote 59 In these documentary texts, the tense appears consistently in conditional sentences but is employed in a manner that diverges sharply from Classical Greek: it occurs in main clauses rather than being restricted to subordinate ones. The context is that of set phrases, such as clauses regulating penalties in the event of non-compliance. The construction is as follows:
Protasis: εἰ + future optative – Apodosis: indicative (more rarely: infinitive)
The novelty of the structure is revealed not only by the presence of the optative in the protasis of direct speech but also by the apodosis with the indicative, where another optative would have been required. To elucidate the case, I give below two of the examples of this non-classical use studied by O’Sullivan. The first comes from one of the Grenfell papyri and is datable to c. 610; the second is from Dioscorus’ archive and is dated to 570.Footnote 60
1. P.Bodl.1.45, 32–3 (TM 22584)
εἰ δὲ ἀσθενήσοιμεν περὶ τὴ[ν ἐ]κδίκησιν καὶ καθαροποίησιν ταύτης τῆ[ς] πράσεως, ἑτοίμως ἔχομεν παρασχεῖν ὑμῖν τὴν ἐγκειμένην τιμὴν ἐν διπλῷ.
If we were to be incapable of the legal remedy and the completion of this contract, we are ready to pay you twice the relevant penalty. (Tr. O’ Sullivan, ‘The future optative’, 95)
2. P.Cair.Masp. 2.67151, 195 (TM 18905)
καὶ εἰ ἀμελήσοι ὄψεται πρὸς τὸν Θεόν.
And if he were to neglect this, he will be answerable to God. (Tr. O’ Sullivan, ‘The future optative’, 97)
O’Sullivan’s numerous examples and detailed discussion demonstrate that the future optative – a tense used rarely and with specific constraints in Classical Greek – remained in limited but persistent use into Late Antiquity. While not part of the spoken language, it continued to be taught and learned, as shown by its presence in educational materials and legal papyri from early Byzantine Egypt. Moreover, its appearance in formulaic and pragmatic contexts, with values distinct from those of classical syntax, highlights a broader syntactic evolution in late Greek that has received insufficient scholarly attention. While, as mentioned above, a few studies have examined the use of the future optative in patristic literature, its presence and application in early Byzantine literature remains largely unexplored. A deeper examination of this feature could shed valuable light on the syntactic developments of Greek during this period.
The future optative in Agathias deserves close attention, especially in light of the patterns identified in the first case study. While a comprehensive diachronic study across Byzantine Greek remains a desideratum,Footnote 61 even a focused analysis of the Histories – supplemented by references to Procopius – yields findings consistent with O’Sullivan’s conclusions from the papyri. In Agathias’ Histories the optative future appears 55 times, a striking number when compared to earlier authors.Footnote 62 Even when compared to his chronologically and thematically closest counterpart, the figure remains impressive: in the entirety of Procopius’ literary output (including the Wars, Buildings, and Secret History), which is much more extensive than Agathias’ five books, Ι was able to identify only eight definite occurrences of the future optative.Footnote 63 Menander the Protector’s count is even more notable given that his history survives only in fragments: his work shows eighty occurrences, suggesting that the total would have been even higher had the complete text survived.Footnote 64 All three authors – Procopius, Agathias, and Menander – had a legal training. Given the tense’s use in contemporary legal papyri, it is reasonable to assume that their education may have reinforced or even revived familiarity with it. At the same time, stylistic preferences likely played a role: Procopius’ stricter adherence to an archaizing idiom, possibly more actively looking to the language of strictly classical models than to that of the Second Sophistic, may explain his more restrained use of the form.
It is notable that John Malalas does not use the tense at all. Although broader contextualization would be needed, this absence suggests that the future optative, whether in its classical form or the later usages studied by O’Sullivan, might have been perceived as a hypercorrect form – a marker of elevated style and refined language. In the following pages, I will argue that Agathias’ extensive use of the tense should be understood in the context of its resemanticization. He used it not only in the classical sense but also with later meanings: of the 55 occurrences of the future optative, I could identify, with some margin of error, nineteen as used in Classical Greek, while almost twice as many, 36, appear within new structures.
As O’Sullivan’s study on documentary legal papyri shows, in sixth-century Egypt the future optative was used in primary sentences to express a remote possibility, associated with an apodosis in the indicative (or infinitive). Agathias’ Histories exhibits a comparable use: in five instances, we find conditional sentences structured as protasis: εἰ + future optative – apodosis: indicative.Footnote 65 For example, at Histories 1.13.2 (K. 26.1–2) we find, in the speech addressed by Narses to the inhabitants of Lucca, the following conditional construction:
ἀλλ’ εἴ γε καὶ νῦν ἐθελήσοιτε μεταμαθεῖν τὸ συνοῖσον καὶ ἔργῳ τὰ ξυγκείμενα διανύσασθαι, οὐδέν τι ἔλαττον ἕξετε.
But if even now you were willing to reconsider the agreement and actually carry out what has been settled, you would not lose anything.
The conditional structure is clear: the protasis consists of εἴ γε καὶ νῦν ἐθελήσοιτε (future optative of θέλω), while the apodosis uses the future indicative ἕξετε. The optative here conveys not simply a future possibility but a certain deliberative distance – perhaps a courteous or strategic tone appropriate to persuasion. In this case, the construction seems to underscore Narses’ diplomatic posture: offering a favourable outcome that remains conditional on the willingness of the audience – the people of Lucca – to comply.
In addition to the structure attested by documentary papyri, Agathias’ prose bears witness to further new uses of the future optative. In twenty of the non-classical instances, the tense appears with the particle ἄν, an unusual combination for earlier Greek. Fassbänder already addressed a number of doubtful occurrences involving this construction, concluding that ‘optativum futuri cum ἄν particula coniunctum a Graecorum dicendi consuetudine alienum fuisse’.Footnote 66 Out of the twenty occurrences with ἄν, four examples occur in conditional constructions within direct speech.Footnote 67 Here, the future optative is found in the apodosis, accompanied by ἄν, and follows a protasis introduced by εἰ with an indicative, subjunctive, or optative verb form. A representative example occurs at Histories 2.12.9 (K. 56.33–57.2), in the context of another of Narses’ speeches, this time to his soldiers:
Εἰ δὲ καὶ παντάπασιν ἐνδοῖεν, (εἰρήσθω γὰρ ἄμφω) ἀλλὰ τό γε ὑμέτερον ἐν ἀσφαλεῖ κείσεται καὶ τὰ κράτιστα δόξοιτε ἂν βεβουλεῦσθαι.
And even if they were to give in completely – both options must be mentioned – still your position will be secure, and it would seem that you had taken the best course of action.
The conditional structure is again clear: εἰ … ἐνδοῖεν (aorist optative) forms the protasis, while the apodosis is composed of two coordinated verbs: κείσεται (future indicative) and δόξοιτε ἄν (future optative + ἄν). While both verbs refer to the same favourable outcome, only the second is modalized by ἄν, marking it as a potential or evaluative judgment. Their coordination suggests a nuanced distinction: the security of the soldiers is asserted with certainty, while the assessment of their decision is presented more as a possibility than as a factual claim.
The combination of the future optative with ἄν had possibly become one of its most common new uses, for the Histories features six occurrences of future optative + ἄν in main clauses (all in direct speech) to express potentiality, where Classical Greek would have employed optative present or aorist.Footnote 68 This is the case, for instance, in Histories 5.18.8 (K. 186.33–187.3):
τὸ μὲν οὖν φρονοῦν ἀμιγὲς ἀγαθὸν καὶ ἀνόθευτον ἅπαντες εἶναι φήσοιτε ἄν, τῆς δὲ ὀργῆς τὸ μὲν δραστήριον εὐκλεές, τὸ δὲ θράσος φευκτὸν καὶ ἀξύμφορον.
You all would admit that, while prudence is a pure and genuine blessing, the active side of anger is glorious, but its rashness should be avoided and is harmful.
This passage demonstrates the future optative + ἄν construction in action, where the future optative expresses potentiality in a context that contrasts virtue with the dangerous effects of anger.
In eight other passages, the future optative appears in purpose clauses, an extremely rare use in Classical Greek, if it occurs at all.Footnote 69 More interestingly, while the clause is introduced by ὅπως only once, the syntagm ὡς ἄν appears in all the remaining seven cases to indicate a purpose likely to be fulfilled.Footnote 70 This nuance was already present in classical purpose clauses, but there the mood used, along with the particle ἄν, was always the subjunctive, not the optative.Footnote 71
Further cases of non-classical uses of the future optative in the Histories are those in relative and conditional clauses, as in Histories 3.19.6 (K. 109.10–13):
ἐγὼ δὲ καὶ τοῦτό φημι τῇ τοῦ κρείττονος ταλαντεύεσθαι γνώμῃ καὶ οὐ παρὰ τοὺς μέγα βρενθυομένους ἰέναι, ἀλλ᾽ ἐφ᾽οὓς ἂν ὁ πάντων προαγωγὸς ἐπινεύσοι.
As for me, I say that this depends on the will of the Almighty and is not granted to those who take excessive pride, but to those to whom the One-who-rules-all would consent.
and in Histories 4.17.5 (K. 144.16–19):
τό γε μὴν ἐπιχώριον ἐμπειρίᾳ τῶν τόπων ἔκ τινος στενοτάτης ὁδοῦ καὶ ὑπολανθανούσης μόλις μὲν καὶ αὐτοὶ καὶ πεπονημένως, καταφέρονται δὲ ὅμως, εἴ που δεήσοι, καὶ αὖθις ἀνέρπουσιν.
The locals, albeit familiar with the place, hardly and with great effort go down, in case they need it, through a most narrow and concealed path, and then climb back up.
In both passages, the future optative is introduced by ἄν to express potentiality – a function that in Classical Greek was predominantly conveyed by the aorist optative and, less frequently, by the present optative. From this initial examination of Agathias’ usage, we observe a syntactic development comparable to the trajectory outlined by O’Sullivan in documentary papyri. The relatively high number of occurrences within less than 200 pages of Keydell’s edition – most of them reflecting non-classical patterns – suggests ‘a demonstrated knowledge and expanded use’ of the tense in the early Byzantine period.Footnote 72 When these data are interpreted within a broader linguistic and sociocultural frame, however, the picture becomes more nuanced. Was the use of this tense firmly perceived as imparting a classical connotation and archaizing tone to sixth-century prose? Or did the newer syntactic environments in which it appeared reshape its perception, aligning Agathias’ usage with more contemporary norms? A comprehensive study of the future optative from its earliest attestations through its later development (including its trajectory after the sixth century) would be needed to answer these questions. However, a brief comparison with Procopius and Menander may provide some preliminary results.
As previously mentioned, Procopius employs the future optative only eight times: five times in the Wars and three times in the Buildings. Footnote 73 Of these occurrences, four preserve classical usage, representing an original future indicative in subordinate clauses. The remaining four instances diverge from this pattern: one appears in a final clause introduced by ὅπως (Wars 8.32.24); two (Buildings 1.2.19, 1.8.13) occur in main clauses introduced by ἄν, substituting a classical aorist optative; and one (Buildings 4.2.28) features in a concessive clause introduced by εἰ. In Menander’s fragments, among eighty instances, at least 25 can be classified as classical, representing an original future indicative in subordinate clauses. The remaining examples align closely with the patterns observed in Agathias: purpose clauses (often) introduced by ὡς ἄν;Footnote 74 the protasis (introduced by εἰ) of conditional clauses, with an apodosis in the indicative or infinitive;Footnote 75 relative clauses;Footnote 76 and main clauses expressing potentiality – functions that in classical usage would have called for the present or aorist optative.Footnote 77
In sum, the future optative – never common even in Classical Greek – underwent a notable reconfiguration in Late Antiquity. Its frequency and increased functional flexibility, as already identified by O’ Sullivan and seen in Agathias (and Menander), suggest a shift in the grammatical and stylistic repertoire available to learned writers of the sixth century. Procopius’ restrained and predominantly classical use of the tense may reflect a more conservative stylistic stance. By contrast, Agathias and Menander make fuller use of the tense’s expanded potential, incorporating it into syntactic environments that would have been rare or impossible in classical usage. Rather than signalling a clear-cut opposition between classical and post-classical practice, these patterns indicate a transitional moment in the history of the learned register, in which inherited forms such as the future optative were not only preserved but actively resemanticized. In this light, Agathias’ usage should be read neither as an archaism nor as a deviation from classical norms, but as part of a broader process of stylistic and syntactic adaptation in post-classical Greek.
Conclusions
The findings from these two case studies – the transitional-adversative syntagm πλὴν ἀλλά and the future optative – support a more nuanced understanding of Agathias’ prose, one that resists the tendency to evaluate it solely against classical norms. Rather than indicating ignorance or carelessness, these features reflect a deliberate stylistic negotiation aligned with broader developments in late antique and early Byzantine Greek. The use of πλὴν ἀλλά highlights Agathias’ sensitivity to transitional formulas likely of spoken origin that had already entered high-register prose through Second Sophistic and patristic literature. Meanwhile, his frequent and semantically extended use of the future optative mirrors patterns attested in documentary papyri and in other contemporary authors, particularly Menander Protector.
These observations point to the existence of a learned, high-register Greek that was not a static or purely imitative code but one capable of internal developments and responsive to broader linguistic shifts. While this study has focused on just two features, it suggests that a broader analysis – one that includes further syntactic structures and lexical phenomena – has the potential to revise our understanding of sixth-century Greek prose and reassess the linguistic creativity operating below its classicizing surface. A comparison with Procopius further underscores this point. Based on this limited study, Procopius’ language appears to adhere more narrowly to established classical paradigms – perhaps classical even in the sense of fifth- and fourth-century BCE models – with relatively little room for influence from Hellenistic or imperial authors. Agathias, by contrast, although equally invested in the prestige of Attic Greek, likely drew from a wider range of authors and registers. His prose integrates syntactic and lexical developments that suggest an openness to the evolving high-register usage of his time – a usage that, while learned and marked, remained dynamic. As Horrocks has argued in the case of Akropolites, ‘high-register medieval Greek was indeed a “real”, if also highly marked, variety of medieval Greek.’Footnote 78 It is time to extend this perspective to earlier authors such as Agathias and approach his language with the same openness to the internal linguistic complexity and creative potential of Byzantine learned prose.
Larisa Ficulle Santini is a postdoctoral researcher at the Austrian Academy of Sciences and Principal Investigator of the FWF ESPRIT Project ‘Fertility control in Byzantium: women’s reproductive agency in the Eastern Roman Empire (330–1204 CE)’. She completed her PhD in Classics at the Universities of St Andrews and Rome Sapienza (2023) with a dissertation on the Histories of Agathias Scholasticus, which she is currently preparing for publication as a monograph. While her research to date has centred on early Byzantine historiography, her current project turns to social, women’s, and medical history, with a focus on practices of contraception and abortion in the early and middle Byzantine era.
