Stories of emperors, sultans, and cities: comparing protagonists in the histories of Doukas and Leonardo Bruni

One of the defining features of the Byzantine historiographical tradition is the dominant narrative roles played by emperors and, in the later period, by Ottoman sultans. This article explores this characteristic feature of the tradition through comparative analysis of the structuring roles occupied by such characters in the fifteenth-century History of Doukas and the protagonistic role of the Florentine people in the contemporary History of the Florentine People by Leonardo Bruni. Transhistorical comparison, organized around two case studies, serves to denaturalize the roles played by emperors and sultans in both Byzantine and modern historiography.

monopoly over protagonistic roles, albeit without ever entirely fading into the supporting cast.
In Byzantine history writingemperors and later sultansstand at the top of a hierarchy of predominantly male characters. The majority of the narrative space not occupied by the emperor is taken up by a selection of generals, courtiers, state-elites, churchmen, foreign rulers, and members of the imperial household, who occupy intermediate roles between these imperial protagonists and the host of minor characters who occupy the rest of each narrative's space and about whom the audience learns relatively little. The majority of minor characters in Byzantine history writing are elite and male, reflecting the profile of the more prominent members of the cast; almost all of Byzantine historiography's non-male and non-elite characters occupy relatively minor roles, despite some notable exceptions, especially in the form of empresses.
The fact that the Byzantine state archives have not survived in any significant quantity means that historiographical narratives constitute the most substantial source material for modern reconstructions of the Byzantine past. 1 This is not to claim that no alternative studies are possible, simply that the organisation and dynamics of Byzantine historiographical narratives have exerted a huge influence over modern reconstructions of that past. The leading roles played by Byzantine emperors and Ottoman sultans are no exception. Modern histories of the Byzantine world consistently reproduce the character hierarchies of their principal sources. It is not just that biographically orientated studies of particular emperors and their reigns constitute one of the most common types of monograph publication concerning the late Byzantine world, but that modern histories consistently deploy imperial protagonists as the principal vehicles for their narratives and arguments. 2 Even beyond explicit biographies, modern historians consistently adopt the emperors and sultans that dominate their source material as the principal protagonists of their own histories of the late Byzantine world. A handful of passages illustrate how emperors and sultans serve to describe and explain the fourteenth-century past in Donald Nicol's classic narrative history The Last Centuries of Byzantium, 1261-1453.
Bayezid had played his first trick skilfully.
This was not what Bayezid had wanted. John V and Manuel had to pay dearly for upsetting his plans. At the end of 1354 John V inherited a situation that called for the combined qualities of a Justinian and a Belisarius. 3 Byzantine emperors and Ottoman sultans provide Nicol's principal means of describing action, while their personalities supply his primary mechanism for explaining it. Today, his style reads as somewhat old-fashioned, but the same types of narrative logic, shorthand, and personality-centred explanation are regularly deployed in modern scholarship, in explicitly narrative histories, in commentaries, and in almost any other format. 4 The dominance of imperial protagonists in Byzantine historiographyclosely associated with the disproportionate weight given to war, politics, diplomacy, Constantinople, the imperial court, and its (principally male) literate elitehas been widely acknowledged. 5 Indeed, much energy has been invested into preventing the positive or negative portrayals of specific emperors and their reigns in particular historiographical narratives from 'misleading' historians in their quest to reconstruct the Byzantine past. Less often interrogated, however, is how the dominance of imperial protagonists affects the narrative structures of Byzantine historiography and how this in turn limits and predetermines not only the manner in which action is presented, but how it is produced as meaningful. While the elite and imperial predisposition of Byzantine historiography has been acknowledged as inhibiting accurate reconstruction, the fact that imperial characters occupy such dominant roles in these narratives has been accepted as a somehow 'natural' structure of historiography itself, rather than as a contingent feature of the Byzantine tradition. Transhistorical comparison with contemporary historiographical traditions can serve to denaturalize the character systems and hierarchies of Byzantine historiography.
Comparison between the character systems of the fifteenth-century Byzantine history of Doukas and the roughly contemporary Florentine history of Leonardo Bruni facilitates the identification and interrogation of the idiosyncrasies of the Byzantine tradition. My basic argument is simple. Leonardo Bruni's Historiarum florentini populi (History of the Florentine People), is, as the title suggests, organized around a protagonistic populus. 6 That is to say, it is centred on a different type of protagonist and thus produces a different character system to that found in Byzantine historiography. Consequently, the character system and narrative structure of Bruni's history serve to denaturalize one of the foundational organising structures of Byzantine historiography, namely the characters that dominate the stories and monopolize the meaning-making functions of the narrative, something that has conditioned and framed the type of Byzantine pasts that modern historians have set out to imagine.
Chris Wickham captures the fundamental potential of comparative analysis that this article seeks to exploit, when he writes that '…if you don't compare, you end up believing that one type of historical development is normal, normative, and that every other is a deviation.' 7 Wickham frames comparative analysis in more materialist terms than the present article sets out to do, but the potential to disrupt what he calls the 'cultural solipsism' engendered by lack of comparison is just as instructive when it comes to historiographical traditions as the historical developments Wickham has in mind. Stories about emperors, sultans, and kings have become normalized in Byzantine studies as the natural mode of narration for the Byzantine past, but extended comparison with another tradition demonstrates that it is possible to tell stories about very similar subject matter with different casts of characters.
This article was originally envisioned as a comparison of urban populations in these two traditions, but the divergent structural positions occupied by such characters in the histories of Bruni and Doukas and the manner in which those narrative positions are reflected in modern scholarship proved too jarring to ignore. Before the minor characters of Byzantine historiographical narrative, such as urban populations, can be centred by scholars in the stories we tell, their relatively marginal position and the discursive violence of their subordination and marginalization must be identified and acknowledged. 8 Consequently, this article seeks to render one of the foundational narrative structures of late Byzantine historiographical narrative, namely the protagonistic monopoly of emperors and later sultans. It begins by introducing the histories of Doukas and Bruni and their fundamental similarities and differences, before exploring two illustrative case studies in detail. The first relates to a typical formulaic description of a military campaign from each history, while the second draws on passages relating to the failed sieges of Florence in 1312 and Constantinople at the end of the fourteenth century.

The histories of Doukas and Bruni
The text that will stand for the wider Byzantine historiographical tradition in this article is the History of Doukas. 9 Although it begins with Adam and a list of biblical personages and earlier Byzantine emperors, the narrative rapidly reaches the fourteenth century in a few short jumps and coverage begins in detail in the year 1341. It breaks off mid-sentence during the Ottoman siege of Mytilene in 1462. As with many other Byzantine historians, there is no possibility of distinguishing the author from the narrator of the History, since the narrative is the only source of information about Doukas, whose first name is unknown. Nevertheless much modern analysis has been organised around using the author, his politics, and motivationsartificially extracted from the narrativeto interpret the text and its reliability. 10 The Oxford Dictionary of Byzantiumwhich notes that 'Doukas was an eyewitness to several of the events he describes, and his narrative is generally considered biased but reliable'accurately captures both scholars' traditional treatment of and reconstructionist interest in the text. 11 Doukas' bias, which has been understood primarily as anti-Ottoman or anti-Turkish and pro-Latin and pro-church union, is not absolute, but relative, and principally determined in intertextual dialogue with the authorial personages extracted from the three other major Byzantine histories of the last century of the Byzantine state: Laonikos Chalkokondyles (1380s-1460s), Michael Kritovoulos (1451-1467), and George Sphrantzes (1413-1477). 12 All four have principally been analysed for what they can tell us about the Byzantine past, within the reconstructionist/empiricist framework of analysis that provides the discipline of Byzantine history with its dominant philosophy of history and methodological apparatus.
Bruni's History of the Florentine People is a classicizing history written in Latin in twelve books, published in instalments over the course of the first half of the fifteenth century. It begins with the ancient origins of Florence, but only offers detailed coverage from the 1250s to 1402. Although much analysis of the text has likewise revolved around its reliability for the reconstruction of the past and its relationship to other narratives of the period, the History of the Florentine People has played an important role in the debates surrounding the Baron thesis, civic humanism, the reception of classical literature and rhetoric, and the emergence of so-called scientific modern history. 13 Like Doukas, Bruni has basically been approached as a relatively reliable source, albeit one increasingly understood as having political and ideological commitmentsmost notable for his alignment with the fourteenth-century oligarchs, in the first half of the history, and with the Medici, in the second. 14 The huge number of other texts written by and about Leonardo Bruni mean that, as with much else, students of Italian history are in a privileged positions in terms of source material, when compared with Byzantinists. Riccardo Fubini, Gary Ianziti, Anna Maria Cabrini, and James Hankins in particular have been able to identify the development and transformation of Bruni's historical and political thought through his voluminous works in a manner completely impossible for Doukas, whose total textual survival is as the narrator of and a character in his own history. 15

Statist narratives with alternative protagonists
Comparison between the histories of Doukas and Bruni is particularly useful because of their considerable overlap in subject matter. Both histories centre the stories of specific state projects. Doukas' story is principally that of the contraction of the Byzantine and the expansion of the Ottoman state (although not necessarily in that order). Likewise, Bruni's historywhich had been adopted as the state-sanctioned version of the Florentine past by the end of his lifeis the history of the Florentine territorial state. 16 These statist narratives both centre on military and political action. This is neatly summed up in the proem of the History of the Florentine People, which identifies the three activities of the Florentine people that were to form its core areas of narration: internal struggles, external struggles against its immediate neighbours, and external struggles against the Duke of Milan and the King of Hungary. 17 While such prefatory remarks should not necessarily be taken at face value, this passage does in fact reflect the preoccupations of the text fairly accurately. 18 Doukas' History offers no comparable prefatory remarks, since the text begins directly with a list of Biblical and then imperial personages. However, once the narrative begins in earnest in the mid-fourteenth century Doukas' History is replete with the intermingled stories of the military expansion and contraction of the Byzantine and Ottoman states, the diplomatic intrigues between their rulers, and the internal dynastic struggles of their ruling houses. The state provides a common-sense framework for the evaluation of action in both narratives, although it takes different forms in each. Eventsnatural disasters, battles, personal decisionsare constructed as good or bad depending on the extent to which they enable the state to reproduce itself and/or extend its territories. 19 This common sense, of course, looks different in each history, not least because Doukas' narrative intermingles two state projects and because Bruni's text is more explicitly didactic.
However, despite their similar subject matter, these narratives of the struggles of, between, and within the Byzantine, Ottoman, and Florentine states are animated by decidedly different character systems. This difference is signalled at the start of both narratives. Doukas' History, from its opening sentence, is the story of individuals and rulers. Beginning with a list of biblical personages, it switches into a slightly eccentric list of Byzantine emperors beginning with Constantine the Great, until it reaches the late thirteenth century, at which point the territorial extent of Turkish conquests in Anatolia and their expansion across the Hellespont come to dominate the narrative. Just as Doukas' History begins with imperial figures so too does it continue. The clash of the Ottoman and Byzantine states that the History narrates is organized around a succession of emperors and sultans, who periodically confront and negotiate with each other throughout the narrative.
In contrast, the History of the Florentine People, after a proem which declares that it will narrate the res gestas florentini populi (deeds of the Florentine people), begins its narrative with the foundation of the Roman colony of Florentia in 80 B.C. 20 It then jumps yet further into the past to narrate the early history of Tuscany and with it the earliest territorial expansion of Rome. This narrative is not devoid of kingsjust as the proem explicitly featured the dukes of Milan and the kings of Hungarybut these characters neither structurally organize nor dominate action in the narrative. The object of the narrative is explicitly identified as civitates Etruriae, whose factionalism and inability to present a unified front against Roman expansion is repeatedly highlighted in prefiguration of the medieval Tuscan history that was to form Bruni's principal subject matter. 21 The wars between these cities, most notably Veii, and Rome feature well-known characters, such as Tarquinius, Brutus, and Horatius Cocles, but it is the inhabitants of these various Etruscan cities and of Rome who dominate the narrative, as can be seen in a passage summarizing the Roman conquest of Veii.
To summarize a great many battles, this one Etruscan city (Veii) carried on the war against Rome down to 403 B.C. and beyond. Sometimes she had the help of other cities, sometimes she fought alone. Sometimes she was beaten in battle, sometimes she won. Her final overthrow, however, came when she decided on her own, not by agreement among the Etruscan people, to resume the war. 22 Later, in a summary passage, the romani duces made famous by these Etruscan wars are listed, but in the narration of action it is principally the Etruscans, Romans, and inhabitants of specified cities who dominate the narrative. 23 This story of Etruscan cities is interrupted by an extended excursus on the history of the Roman empire, or rather its collapse due to the institution of imperial rule (internal struggles) and repeated barbarian invasions (external struggles). During this interlude the cities of Tuscany are presented as stifled by and submerged under Roman imperial rule. The urban characters of the earlier conflicts are abandoned in favour of a multitude of wicked emperors and barbarian hordes and their rulers. These characters dominate the narrative of the Roman republic's collapse into empirewhich Bruni lamentsand multiple waves of barbarian invasions of the Italian peninsula. 24 However, at the end of the first book of the History of the Florentine People, the narrative returns to its protagonist, with the medieval re-emergence of Florence and various other Tuscan cities, of which Pisa and Siena are singled out for particular attention, in a passage that explicitly reaches back to the earlier Etruscan passages. 25 At this point Bruni returns to Tuscan discord, introducing the conflict between the Papacy and the Holy Roman Empire that led to the Guelf-Ghibelline factionalism that animates much of the early parts of the History of the Florentine People. 26 Thereafter Florence and the Florentines sit at the heart of the narrative, confronting its various enemies, whether in the form of communes, tyrants, or internal factions.
Both histories imply a similar range of past actors, but from the beginning of each narrative the balance of space, importance, and meaning-making functions occupied by different types of characters are dissimilar. The string of Byzantine rulers and later Ottoman Sultans who begin and dominate Doukas's History stand in stark contrast to the people of Florence, whose foundation story and regional prehistory offer the starting point for Bruni's narrative. While the populations of Constantinople, Thessaloniki, and other cities appear in Doukas' story and emperors, kings, tyrants, and dukes litter the pages of Bruni's text, the positions that such characters occupy in these narratives and the explanatory functions they perform differ dramatically. The following case studies examine two examples of relatively typical presentations of military and political action in order to excavate the common-sense logic and mechanics that animate these narratives. 27

Narrating Ottoman and Florentine expansion (Case study 1)
The presentation of the first years of the reign of Bayezid I (r. 1389-1402), in book IV of Doukas' history, offers an illustrative example of how military action is typically produced in the narrative. 28 Bayezid first appears in the narrative in its presentation of the so-called battle of Kosovo (1389), an event enshrined in Serbian folklore and nationalist mythology, which saw the death of both Bayezid's father, Murad I (r. 1362-1389) and the Serbian King Lazar (r. 1373-1389). 29 At the battle Bayezid is introduced for the first time as the leader of the army's left wing, at which point he is described as 'terrifying and mighty above all others', and the narrative relates how he had his brother blinded immediately after learning of his father's death. 30 This opening characterization is shortly afterwards crystalized by an assessment of the character of Bayezid, in which he is presented as a barbaric heathen terror: Bayezid was acclaimed ruler of the Turks. He was a feared man, precipitate in deeds of war, a persecutor of Christians as no other around him, and in the religion of the Arabs a most ardent disciple of Muhammad, whose unlawful commandments were observed to the utmost, never sleeping, spending his nights contriving intrigues and machinations against the rational flock of Christ. 31 This introduction to the character who will dominate the narrative until his death fourteen books later directly precedes the passage under discussion here. 32 The next chapter (IV) describes, in a short passage of just 451 words, the stabilization and expansion of the Ottoman state under Bayezid, by describing his subjugation, by turns, of various cities and regions, mostly in western Anatolia. The narrative centres the person of Bayezid, who is repeatedly named and renamed throughout the passage. When explicitly named, Bayezid only appears as the grammatical subject of the sentence in question, although his presence extends across the whole passage, with his actions repeatedly implied. The narrative moves each time with Bayezid to the next territory he attacks.
Over and above Serbia, where this passage begins, Bayezid travels to and conquers six defined areas: Kutahiya and Phrygia; Ionia; Caria and Lycia; Magnesia, Lydia, and the Aeolian cities; and Philadelphia. Each episode largely follows a simple three-part formula. First, Bayezid travels to a new location (a city or region), which is signposted by the description of his route and travel (such as crossing the Hellespont; or descending Mount Tmolos). Second, he captures the place either by force or surrender. Third, the fate of the enemy ruler (consistently identified as ἀρχηγός) is described (they are generally killed, imprisoned, or flee to the Persians). After this, the formula repeats as Bayezid moves on to the next location. 33 The presentation of the subjugation of all the Turkish emirates described here is really quite formulaic, as the examples of Ionia, Caria, and Lycia illustrate.
Traversing Phrygia and marching from Laodicea to Ephesus, Bayezid became master of Ionia. Seizing Isa, Aydin's grandson, the ruler (ἀρχηγόν) of Ionia, he exiled him to Nicaea in Bithynia, and there he spent the remainder of his life. Bayezid transported his entire force over the Maeander, and after he had boldly taken all of Caria and Lycia, Ilyas, the ruler (ἀρχηγός) of these provinces, fled to the Persians. 34 The cases of the Christian territories of Serbia and Philadelphia, which book-end this list of conquests, offer slight variations on this formula. Since Bayezid is already in Serbia and the battle of Kosovo already won, only the third part of the formula is relevant and this section focuses on the consecration of the new Serbian ruler, his future military obligations to the sultan, and the marriage alliance that joins their two houses.
In the case of Philadelphia there is a shadow of communal resistance to Bayezid's inevitable conquest, although the city's inhabitants soon surrendered due to lack of provisions. Bayazid next took the road to Philadelphia, because that city, both extensive in size and densely populated, had remained free for nearly 100 years. The whole earth was subjugated by the Turks, but this city shone like a star in the clouded meridian. The Turks laid siege to the city … but, unable to continue because of a lack of provisions, they surrendered. After first installing dependable commanders and governors (ἀρχηγοὺς καὶἡγεμόνας) in the provinces, Bayazid took all his forces from the East and marched to the western regions. 35 Here, no local ruler (ἀρχηγός) is identified and although the passage lacks clarity due to a lacuna in the manuscript, the Philadelphians are presented as surrendering their own city. This is the only moment when Bayezid's conquests are not explained exclusively through the transfer of power from local rulers to Bayezid, but this missing element of the formula is immediately supplanted by mention of the 'rulers and leaders (ἀρχηγοὺς καὶἡγεμόνας)' appointed by Bayezid after the conquest. 36 33 Whilst ἀρχηγός is a common lemma in a number of late Byzantine historiographical narratives, it has the highest number of occurrences in the History of Doukas of any single text in the TLG corpus. In large part this is because it is one of Doukas' preferred terms for referring to Ottoman sultans. Note also that it is also used to identify a range of other characters, such as Niccolò II Gattilusio, the ruler of Lemnos. Doukas, Istoria This fourth book illustrates the basic mechanics for the presentation of military action in Doukas' history and Byzantine historiographical narrative in general. The subjugation of huge swathes of western Anatolia is described through the actions of a protagonistic ruler and explained principally by reference to the rulers he captures, kills, deposes, and appoints. While the fleeting presence and resistance of the Philadelphians reminds the reader that these conflicts were not literally between a handful of named individuals, the narrative of these Ottoman victories flows around and over such transient minor characters. Furthermore, this campaign is made meaningful within the wider narrative by illustrating and confirming the initial characterization of Bayezid and the image of the implacable barbarian aggressor presented on his first appearance in the narrative.
Although they do not always take place on anything like the same geographical scale as some of those described by Doukas, Bruni's narrative is replete with military campaigns. One comes at the beginning of the second book of the history to the period following the death of Frederick II. 37 At this point in the narrative the first book has already covered the earliest history of Tuscany/Etruria, the decline and division of the Roman empire, and the emergence of Guelf/Ghibelline factionalismthe last of which is blamed on Frederick. The narrative that follows Frederick's death, however, unlike Doukas' passage following the death of Murad I, does not relate to the imperial succession, but rather to the insurgency of the florentinus populus, who are described as being 'consumed with hatred for the arrogance and ferocity of those who had seized the commonwealth [i.e., the Ghibellines]'. 38 Bruni then describes in some detail how this new government of the people organized itself, summing up the change as follows: The People was now itself a lord (dominus) and a font of honour, and men who only a short while before had been frankly servile towards princes and their supporters, now, having tasted the sweetness of liberty, bent all their strength on raising themselves up and acquiring an honourable standing in their own community. 39 Bruni's presentation of the military campaigns that immediately followed the establishment of this popular government are illustrative of a different mode of narrating military action. Like the passage just examined in Doukas, it comes early in the text, before the narrative is at its most detailed. It is similarly formulaic in the way it presents successive campaigns and like Bayezid's conquests only describes the victories of one side.
The presentation of these campaigns in the History of the Florentine Peoplewhich occupies seven pages in the critical edition (II.3-15)is considerably longer than the 37  passage examined in Doukas, and there is more detail and variation in the text. However, there is still a relatively coherent formula. First the florentinus populus makes a decision or experiences an emotion (it aroused, for example, by the report of imperial interference in Tuscany, elated by previous victories, or feels grief because of the defeat of Lucca). Second, they move to a location or take up a position. Third, there is some kind of military action or at least the threat of military action (for example, a battle or a siege), which generally precipitates some decision or emotion from either the enemies or the Florentines. Fourth and finally there is some kind of settlement in which the method of peace-making is described (generally the exchange of ambassadors) and finally the conditions of the settlement (such as the surrender of towns or the return of exiles to a city).
The presentation of the campaign against Pistoia is illustrative of this formula, although the absence of an extended description of military action means that it is relatively short: The People (populus) were elated at this series of triumphs and as soon as spring came they went out in force once more and encamped around Pistoia. The Pistoiese had no hope left either in themselves or in their friends and decided at length to give way to the will of the Florentine people (Florentinorum voluntati) rather than fight it out to the end. So the Florentines sent out to draw up an agreement an ambassador named Ildebrando di Ottobono, who at that time was a man of great authority in the commonwealth, together with two judges. They allowed peace to the Pistoiese on the following conditions: that the league and friendship between the Florentine and Pistoiese people (Florentino Pistoriensique) should henceforth be perpetual; that the Pistoiese should take all exiles back into their city; that they should restore the possessions of the latter… 40 Ildebrando di Ottobono, the ambassador sent to arrange a peace with Pistoia, is one of only two named characters to appear in the seven pages analysedwhich describe confrontations in more than fifteen different locations. The characters that power this narrative are collectives. The Florentine people are the principal protagonist in the narrative. Their psychological disposition is presented repeatedly and they are the principal actors. They are aided and abetted by a similar cast of characters, made up of other urban populations.
To cut an elaborate story short to make a simple point: these two passages illustrate the fundamental difference in the cast of characters in each of these histories. Florence's armies, presumably, had generals, just as Bayezid presumably had an army, but it is Bayezid and the florentinus populus who respectively drive these narratives. 40 Bruni, History, I, 116-18 ( §II.10); Hankins, History, 117-19.

Narrating the sieges of Florence and Constantinople (Case study 2)
Later in Bruni's narrative, when Florentine interests and ambitions reach further afield, the composition and accordingly the presentation of Florence's enemies change. Whereas in Book II Florence's enemies are presented in similar terms to the protagonistic Florentine people, in Book V (my second case study), these enemies look different. Between 1312 and 1342, Florence was at war almost constantly. 41 During this period, the Florentines are presented as engaging in military conflict with emperors, kings, and Ghibelline tyrants in Tuscany, as well as well as with urban communes, from which those individuals are sometimes indistinct. In 1312, Henry VII, Holy Roman Emperor, laid siege to Florence. This event has been described as prefiguring the increasing foreign intervention that characterised the fourteenth-century history of the Italian peninsula. 42 It offers a good example of how Bruni's narrative produces imperial characters within an alternative system and hierarchy of characters.
In the run-up to the siege, Bruni describes various manoeuvres between the imperial and Florentine forces in the Arno valley. Although the Florentine force was able to avoid a pitched battle the imperial forces managed to outmanoeuvre them, by crossing a mountain pass, which allowed them surprise the Florentine army and reach the city before its defenders, leading its citizens to assume that the Florentine force had been defeated. In Bruni's presentation the emperor features prominently in the narrative, but the principal character involved in the passage remains the Florentine people, with the enemy forces and their leader intermingled to form a kind of secondary composite character.
Thus the emperor (imperator) could neither fight nor pass, since the castle overlooking the road on the high cliffs forbade easy passage. So, acting on the information of the exiles, he began to make his way through the extremely difficult passes in the nearby mountains to the north. / When the Florentines at Incisa saw this, they grew afraid that they would be cut off by the enemy. They at once advanced standards, fell into marching order and headed off towards Florence. / Seeing them from the hilltop (for a certain part of the imperial forces had already made their way past the town) the enemy (hostes) attacked in force. / The pass was somewhat obstructed, and the enemy (hostis) were raising a clamor on all sides as they rushed down the hill to fight. / The Florentines (Florentini), however, had determined not to fight but to march rapidly to the city. / Thus when they saw and heard the enemy rushing towards them they at once turned around and returned [to Incisa]. / The retreat (fuga, of the Florentines) was somewhat hasty and soon took on the appearance of a rout, but the town was nearby to help them evade any signal calamity that day. / There was no great number (magnus numerus) of killed and captured (Florentines), but they were mentally shattered as though they had suffered a total defeat. 43 In the eight sentences that make up this passage, the emperor is the subject of just one. In contrast the Florentines (or a noun related to them or their actions) are the subject five times and the imperial forces (termed either enemies or the enemy) are the subject twice. The Florentines are not always explicitly named, as one would expect in a highly inflected language like Latin. Consequently, James Hankinsin his close translation of the textis regularly forced to supply 'the Florentines' for clarity and ease of comprehension, since the narrative takes for granted that any floating third person plurals, as in the second sentence in this passage, refer to the history's popular protagonist, even if that requires the reader to reach back several sentences. Neither this brief example, nor the disproportionate occupation of the subject position is intended to provide empirical proof of the narrative's character hierarchies. However, it does offer an example of how 'the Florentines' typically dominate the narrative, even when they are reacting to some foreign power or named character.
If such differences between Bruni's Henry and Doukas's Bayezid are not already apparent, comparison of some other passages in and around the fourteenth-century sieges of Florence and Constantinople, will demonstrate them further. Sieges and the conquest of cities are two of the few types of events in which non-elite characters consistently appear in Byzantine historiographical narratives. However, as will become clear, they do so within a narrative framework that builds meaning on the back of individual protagonists.
The preparations of the defenders for these sieges are telling. Measures are taken, both in Constantinople and Florence, to prepare the cities' fortifications. In Doukas this is presented as the emperor John himself building towers and fortifications when he beheld the audacity of the tyrant Bayezid.
When the emperor (Ὁ δὲ βασιλεὺς) beheld the blatant ambition and audacity of the tyrant, he began to build (ἤρξατο κτίζειν) in that part of the City called the Golden Gate, two towers on either side of the gate… 44 The emperor is presented as literally building and completing the fortifications himself and later it is John himself who demolished the fortifications, when Bayezid threatened 43 Bruni, History, II, 6-8 ( §V.8); Hankins, History, II, 7-9. 44 Doukas, Istoria, 75.19-23 ( §XIII.3); Magoulias, Decline and Fall, 81-2, slightly adapted. to blind his son and heir Manuel. Both activity and meaning are thus supplied to these works through the persons of John, Manuel, and Bayzeid.
Bruni's description of the Florentines' defensive measures, in contrast, looks very different. On seeing the imperial army, the terrified Florentines are described as taking up arms, while the fortification and defence of the walls is explained through the allocation of sections to various unnamed gonfaloni (standard bearers).
Nevertheless, the people took up arms, and under their standards raced to protect those parts of the city where the enemy was stationed. There each of the gonfaloni was assigned places and sections of the city to defend, and labour on the earthwork was resumed and continued without interruption, day and night. Wooden towers and redoubts were jury-rigged in the more exposed places and filled with armed men. 45 Bruni's presentation is not devoid of individuals, as the emperor Henry looms large as a terrifying bogeyman threatening the city. The failure of the imperial forces to attack immediately during this panic is framed in personal terms, as resulting from Henry's indecision and overconfidence in local support. 46 Nevertheless, as in the passage examined above, the Florentines dominate the narrative not only in action, but also in the allocation of meaning. Henry's indecision is made meaningful, because it allows the citizens to recover their nerve and the Florentine army and allies to arrive.
Another example can be seen in how both narratives relate the call for aid by the besieged parties. In the run up to the siege of Constantinople, Doukas' history lists the various rulers to whom the emperor Manuel appealed for aid, which in his story leads directly to what historians call the crusade of Nikopolis (1396).
Emperor Manuel, in despair because there was no help whatsoever from anyone, wrote to the pope, the king of France, and the kral of Hungary, informing them of the blockade and the City's desperate condition…With the coming of spring the king of Flanders, many Englishmen, the nobles of France, and many Italians came to Hungary…With them was Sigismund, the kral of Hungary, who was also called emperor of the Romans. 47 In short, the logic that appears here is the same as that which made Bayezid's campaigns in the first case study meaningful through the list of rulers he defeated. 48 In Bruni's history, help arrives not in the form of a list of rulers, but as the armies of various cities, whether they are on the Florentine or Imperial side. 45 Bruni, History, II, 8-10 ( §V.10); Hankins, History, II, 9-11. 46 Bruni, History, II, 10 ( §V. [11][12]. 47 Doukas, Istoria, 79.15-25 ( §XIII.8); Magoulias, Decline and Fall, 83-4. 48 The same logic is repeated in the passage detailing the Ottoman conquests through reference to the lieutenants (ἀρχηγοί) that Bayezid sent out before the siege. Doukas, Istoria, 77.29-79.5 ( §XIII.6).
In this situation, help arrived from the allies in a most opportune way. The Lucchesi (Lucensium) sent three thousand foot and six hundred horse; the Sienese (Senesium) sent the same number of horse and two thousand men; and other allies (aliorum sociorum) sent such troops as their resources permitted.
Seeing that the emperor was in this location, the Pisans (Pisanorum) sent him help three hundred foot and fifty knights. The Genoese (Genuensium) also sent about a thousand archers… 49 Thus, just as with the cities' physical defences, these stories produce aid in divergent ways, even though the human actors represented must be assumed to be similar. Henry's Genoese archers, presumably had a leader with a name, just as the king of Flanders is assumed to have brought soldiers with him to Nikopolis.
Even when the Constantinopolitans appear as a character, their appearance is made meaningful through their interactions with named imperial and sultanic protagonists. As Ιoannis Smarnakis has usefully observed, '[t]he political life of Constantinople is often described as a field of interaction between the plans of the emperor and the wills of its people.' 50 However, while the emperor also regularly acts independently, the Constantinopolitans' actions are predominantly mediated by those of imperial and sultanic protagonists. Doukas himself, in the middle of the siege narrative, wonders aloud what these protagonists meant to do, asking: 'What was the aim of Bayazid, and what the aim of Emperor Manuel?' 51 Shortly after this question the common people are described as contemplating disloyalty, because of the terrible conditions of the siege, but only in the context of Manuel himself noticing the suffering of his subjects: The emperor, a devout and prudent Christian, recalled the words of Holy Scripture when he observed all his subjects suffering from want (ὁρῶν τὸ ὑπήκοον ἅπαν ταλαιπωρούμενον ὑπὸ ἐνδείας). A measure of grain sold for more than twenty gold coins, but where could one gold coin be found? Out of necessity the common people were looking to treachery and betrayal of the fatherland. 52 Perhaps, the Constantinopolitans' most active moment in the siege narrative comes in the following book (XV), when they refuse to surrender to Bayezid. Here rather awkwardly the Constantinopolitans end up responding to a pronouncement sent not to them, but to the emperor John by Bayezid: 49 Bruni, History, II, 10, 12 ( §V.12, V.14); Hankins, History, II, 11, 13. On his return…he made the following pronouncement to Emperor John, 'If I expelled Emperor Manuel from the City, I did this not for your sake but for mine. If you wish to be my friend, then leave the City and I will grant you any province you desire. But if you refuse, as God and the Great Prophet are my witnesses, I will spare no one; indeed, I will kill everyone without exception.' When Bayazid had sent this wrathful message which was followed by similar warnings, the Constantinopolitans placed their hopes in God. Long before they had brought into the City a small stock of supplies. Their answer to Bayazid was as follows, 'Go and say to your lord: Since we are powerless and greatly oppressed, there is no place where we can find refuge except in God who helps the weak and mightily oppresses the oppressors. Do as you like.' 53 The Constantinopolitans, however, demonstrate very little actual activity. Caught between the sultan and emperor, they merely put their faith in God. Shortly after this, the Constantinopolitans (again combined awkwardly with the emperor) once more put their faith in god, before the narrative goes on to explain how they were delivered by the defeat of Bayezid by Tamerlane, a conflict once again explained in personal terms. 54 The Constantinopolitans are clearly supposed to be produced as helpless and at the mercy of providence, but what is most relevant is how this character occupies only a marginal and passive position. Their actions are framed by and made meaningful within the stories of emperors and sultans, just as emperors appear on the margins of Bruni's story of the Florentine people.

Conclusions and future directions
Two different protagonists, character hierarchies, and modes of presenting action can be found in the histories of Bruni and Doukas. Whether in the construction of military campaigns, the building of urban fortifications, the request for aid, or the response to a siege, each narrative weighs very differently the amount of narrative space, attention, and meaning-making functions that certain types of characters receive. There may be rulers and urban populations in both histories, but they do not perform the same narrative and explanatory functions, attain the same levels of coherence, availability, and accessibility, or occupy the same amount of narrative space. In Doukas' narrative there is almost always an emperor and a sultan present at the heart of action, just as the Florentine people consistently dominate Bruni's narrative.
This exploratory article is not trying to suggest that male rulers are not integral to Bruni's history. They clearly are. The actions of named elite male charactersnot to mention their extended speechesare essential to understanding not only specific action, but also the usable past that Bruni was creating for his oligarchic and later Medici patrons. The tactical omission of members of the Medici family in the latter parts of the history is just one of many examples of how individual characters and their names produce meaning in Bruni's story. At the other end of the spectrum, one need only read Bruni's unsympathetic presentation of the Ciompi revolt to understand that he was not engaged in an egalitarian historical project. 55 However, there remains a significant disparity between the manner in which ostensibly similar events and characters are presented in the two histories.
This simple comparison denaturalizes a foundational structure of Byzantine historiographical narrative and invites modern historians to rethink the kind of logics upon which their own narratives, explanations, and arguments depend. Looking outside of the narrative and discursive tradition that has produced the kind of modern scholarship cited at the beginning of this article, has the potential to push Byzantinists out of the reconstructionist framework that has confined research into Byzantine historiography. Reconstructionism has channelled the reading of Byzantine historiography towards places where 'verifiable historical knowledge' might be produced. Here positivist methodology has allied itself to the new realities of the neoliberal academy, which demands the manufacture of new stuff (whether pasts or publications), even from Byzantinists.
What would serve Byzantine history better than tweaking the traditional historical narratives of what happened and when is an alternative framework to reconstructionism. As an alternative, I suggest a (non-reconstructive) narrative poetics of Byzantine historiography. A poetics that will provide space for us to fully explore the textual and narrative qualities of these stories and their characters. Such a project will not only challenge traditional analyses of the Byzantine historical tradition and constructions of the Byzantine past on their own terms, but also free us up to develop alternative reading strategies and different kinds of historiographical narratives. Perhaps even more importantly, such an approach will provide a framework through which the study of Byzantine history writing can be meaningfully placed in conversation with wider pre-modern European and global traditions of history writing and narrative.
Matthew Kinloch is a research fellow at the University of Oslo, where he is currently the principal investigator of the Research Council of Norway project 'Narrative Hierarchies: Minor Characters in Byzantine and Medieval History Writing'. He holds a doctorate in History from the University of Oxford and a masters in Byzantine Studies from the University of Birmingham. His research interests include historiography, narratology, gender, and comparative history.