It is generally agreed that one can reconstruct a large part of the Syriac chronicle of Theophilus of Edessa (d. 785) on the basis of the similarities displayed between the four main witnesses to the same in their descriptions of events of the period c. 590–755, where these witnesses are the chronicles written by Theophanes the Confessor (d. 818) in Greek, Agapius of Manbij (wr. 940s) in Arabic, and Michael the Syrian (d. 1199) and the anonymous author of the Chronicle of 1234 in Syriac.Footnote ¹ However, since the last two authors only knew Theophilus’ work through its use by their common source, Dionysius of Tellmahre (d. 845), one is effectively reconstructing it upon the basis of three sources, the chronicles by Theophanes, Agapius, and Dionysius. If all four witnesses include the same detail, then one can be reasonably certain that Theophilus of Edessa included this detail in his original chronicle, but the attempted reconstruction of his text becomes much more difficult if only one of the four witnesses to it seems to preserve a particular detail. Such is the difficulty in which one finds oneself when considering the statement by Theophanes (AM 6160) that Constans II died in a bath called Daphne ‘Laurel’: ‘In this year, the emperor Constans was murdered at Syracuse in Sicily in a bath called Daphne.’ (Τούτῳ τῷ ἔτɛι ἐδολοφονήθη ὁ βασιλɛὺς Κώνστας ἐν Σικɛλίᾳ ἐν Συρακούσῃ ἐν τῷ βαλανɛίῳ, ᾧ ὄνομα Δάφνη.).Footnote ²

Since Theophanes is the only one of the four witnesses to Theophilus’ text to include this detail, Mango did not ascribe this detail to what he referred to as his ‘eastern source’, that is, to Theophilus, in his analysis of his text, even though he did ascribe all the rest of the sentence before it to this source, and Hoyland followed him in this.Footnote ³ Given that Theophanes proceeds next to engage in a detailed explanation of why Constans was assassinated that blames it on his treatment of Pope Martin I of Rome and Maximus the Confessor, the details of which explanation have no parallel within the other witnesses to Theophilus’ chronicle, the obvious temptation is to ascribe the name of Daphne to the same, presumably Greek source from which this next material also derives. Nevertheless, no surviving Greek source preserves this particular detail independently of Theophanes, although an anonymous Greek life of Pope Martin I, composed probably c.740, does record that Constans was killed in the bath of Phatne ‘Manger’: ‘This Constantine was killed by a sword in the bath of Phatne in Sicily.’ (ὅστις Κωνσταντῖνος ἐν Σικɛλίᾳ ἐν τῷ τῆς Φάτνης λουτρῷ μαχαίρᾳ ἀνῃρέθῃ).Footnote ⁴ It is clear from the similarity of the names, Daphne and Phatne, that they represent variants of the same name, but it is not clear which is correct, or better preserves the original name. It seems equally plausible that Constans could have visited baths at a place of either name, although the popularity of the former as a place-name during the Roman period does incline one to favour it, even if it does not now seem possible to locate such a place in Sicily itself.Footnote ⁵

As to whether Theophilus originally named the baths, or the location where they were situated, or how he did so, it is arguable that Agapius provides the answer to this question in his description of how Constans travelled from Constantinople to Syracuse in Sicily:

This description of Constans’ journey from Rome to Sicily via Antioch is clearly problematic. On the one hand, there was no known place called Antioch on the route southwards from Rome to Syracuse. On the other hand, if Agapius is referring to the most famous town of that name instead, as seems probable, he seems to be saying that Constans journeyed from Rome to Antioch in Syria before returning to Sicily. Nevertheless, this would be a ridiculous route to take on any journey from Rome to Sicily. Furthermore, it would also have entailed a major conflict with the Arabs who had occupied northern Syria since c. 638. It is therefore obvious that Agapius errs in his claim that Constans visited Antioch on the way to Sicily. One also notes that neither Michael the Syrian nor the chronicler of 1234 support him in this, which suggests that the error is peculiar to his text, that is, that he is responsible for this error rather than his source. So why has he erred in this way? One may assume that he did not deliberately invent the claim that Constans visited Antioch while on his way to Sicily, since there is no obvious reason why he should have done so. That leaves only one other possibility, that he misunderstood something in his source as he translated and abbreviated it. In order to determine what that something was, one must first compare and contrast his account of Constans’ journey to and death in Sicily to the parallel accounts of the same in the other witnesses to Theophilus’ chronicle in order to determine what, if anything, is missing from his text, and how this missing element might have contributed to the strange claim that Constans visited Antioch.

When one does this, one is immediately struck by the coincidence between the alleged name of the baths where Constans was killed according to Theophanes, Daphne, a detail omitted by Agapius in his description of how Constans was killed at the baths, and the name of the most famous suburb of Antioch in Syria, Daphne also.Footnote ⁷ This suburb lay about 9 kilometres south of the city, and was famous for its natural beauty, its gardens, springs, baths, and a large temple of Apollo.Footnote ⁸ Consequently, it is my suggestion that Agapius read some reference to a visit by Constans to Daphne while on his way to Syracuse in Sicily, mistakenly identified that Daphne with its far more famous namesake in the suburbs of Antioch, and concluded that he must have visited Antioch while on his way to Sicily. It is arguable, therefore, that both Theophanes and Agapius had access to some edition of Theophilus’ chronicle reporting that Constans had visited baths called Daphne. This conclusion is reinforced by the fact that they share many other pieces of information not preserved by either Michael the Syrian or the chronicler of 1234.Footnote ⁹ Their common identification of the name of the baths as Daphne seems to be yet another example of the same phenomenon, even if it has been thoroughly obscured by Agapius’ misidentification of the Daphne in question with the more famous suburb of Antioch.

If one turns next to the question of how Constans was assassinated, rather than where he was assassinated, it is noteworthy that the four witnesses to Theophilus all preserve the same basic account claiming that he was killed by being hit on the head with a bucket while bathing. As all the witnesses to Theophilus’ chronicle agree, he had closed his eyes in order to prevent soap getting into them, and this provided his assassin with the opportunity to strike him while he was off his guard. It will suffice to quote Theophanes alone here as he resumes his description of the assassination proper following his digression on the reasons for it:

There are some disagreements between the various witnesses, but they are relatively slight and need not concern the reader much here. For example, Agapius and the chronicler of 1234 both agree that, after Andrew had escaped, other servants entered the baths and removed Constans, who was merely unconscious rather than dead, although he did finally die a day or two later.

As one commentator has recently noted, ‘this is a rather peculiar account’, raising more questions than it answers.Footnote ¹² Its originator seems to have wished to emphasize that Andrew had acted alone so that there was no connection between his actions and the subsequent decision to raise Mezizios to the throne. The clear implication is that Andrew acted as he did as a result of some personal or private grievance against Constans, but no attempt is made to explain the same. Of course, it is not impossible that Andrew did act alone and as a result of a purely personal grievance. After all, a disgruntled slave, angry because the caliph had not accepted his complaint against the financial demands of his master, acted alone in fatally stabbing the caliph ‘Umar while he was at prayer in a mosque in 644, or so it seems.Footnote ¹³ However, the fact that Andrew seems to have managed to escape the scene of the crime unharmed, never to be heard of again, suggests the existence of a network of accomplices or collaborators. Indeed, since his father Troilus was very likely the patrician of this name who played a large part in the trial and imprisonment of Maximus the Confessor in 655–656, and was clearly a very important person within court-circles, Andrew ought to have been very well connected within the same court-circles as well.Footnote ¹⁴ Hence, there is a real suspicion that the death of Constans was rather more complex that Theophilus of Edessa seems to have preserved. Indeed, the anonymous author of the Spanish Chronica Byzantia Arabica (Chronica Anni 641) and Nicephorus both agree that several servants were involved in the death of Constans.Footnote ¹⁵ Of course, this belief in a wider conspiracy against Constans may simply reflect imperial propaganda after the crushing of the attempted revolt by Mezizios, but it does have the advantage of plausibility.Footnote ¹⁶

There is no reason to doubt that Constans died while at the baths, since, as already noted above, the anonymous author of the Greek life of Pope Martin I supports this, as does Nicephorus. Furthermore, it would have made good sense to have attempted to kill him at a place where he would have been naked for most of the time and thoroughly exposed to any form of attack. Unsurprisingly, mythology and history provide numerous examples of rulers killed in the bath.Footnote ¹⁷ However, the claim that Andrew killed Constans by hitting him over the head with a bucket is problematic. First, it seems inherently implausible that any assassin should have tried to kill his victim by hitting him on the head with a bucket. No one using this method could possibly have hoped to achieve his aim with one swift blow, but would surely have needed several repeated blows in order to ensure that he had not merely stunned his victim, yet the longer the process took, the greater the possibility of failure, of being caught in the act. Numerous other emperors had been, or would be, assassinated over the centuries, but no one else was ever killed by being battered to death with a bucket, and that ought to raise doubts about the plausibility of this alleged method of assassination. Next, and of far greater importance, two earlier, relatively reliable sources specifically record that Constans was killed by a sword, not a bucket. Writing during the late seventh century, Anastasius of Sinai describes the death of Constans as follows: ὁ Μαρτίνου διώκτης ἐν Σικɛλίᾳ ὑπὸ μαχαίρας ἀπέθανɛν. (‘The persecutor of Martin died by the sword in Sicily.’).Footnote ¹⁸

He is supported in this by the author of the anonymous Greek life of Martin I, as already noted above. Since these authors wrote earlier than Theophilus of Edessa, and their version of events seems so much more plausible than that which he apparently preserved, one is tempted to accept that they are probably correct and that Theophilus is wrong.Footnote ¹⁹ Nevertheless, neither Anastasius nor the anonymous author of the Greek life of Martin I reveal any serious knowledge of the circumstances surrounding the death of Constans, and their claims that he was killed by a sword in particular may simply represent their own assumptions based upon no more than a knowledge that he was assassinated and the fact that a sword would have been an obvious weapon in such circumstances.

It is precisely the implausibility of the claim that Andrew killed Constans with a bucket that should cause one to pause before dismissing it, because it is difficult to understand why anyone responsible for the transmission of the larger story should have deliberately invented this detail. If no one deliberately invented this detail, that leaves two other possibilities instead, either that someone in the chain of transmission invented it as a result of some misunderstanding of their source or that this detail had formed part of the story of the assassination of Constans since its first promulgation probably by the imperial administration itself.Footnote ²⁰

As regards the first possibility, it is difficult to imagine how someone might have misunderstood his source in such a way as to believe that Andrew had killed Constans with a bucket, but one possibility lies in the misunderstanding of his office. By the late ninth century, there was a senior palace official entitled the ‘protospatharios of the basin’ (ὁ πρωτοσπαθάριος τῆς φιάλης).Footnote ²¹ He was the judge of the imperial oarsmen, but the significance of the term ‘basin’ (φιάλη) in this context is not really understood. It may refer to the basin of a fountain that stood on the quay of the palace-harbour, the so-called Boukoleon, or it may refer to the round ‘basin’ of the harbour itself.Footnote ²² It is unlikely that an official of this exact title existed under Constans, since the first reliable evidence for the title protospatharios is when Justinian II sent the protospatharios Zacharias to arrest Pope Sergius in the 690s.Footnote ²³ However, someone must have performed the same, or similar, role earlier, although probably not before Constans, who seems to have been responsible for the creation of the navy as a permanent branch of service in response to the sudden development of an Arab navy.Footnote ²⁴ Furthermore, given the essential conservatism of the Byzantine administration in the naming of such posts, it would not be surprising should this post have been named after the ‘basin’ then also, whatever exactly was meant by this term. It is possible, therefore, that Andrew was serving as the ‘controller of the basin’, or in a post of similar title, when he attacked Constans, and that the ultimate source for Theophilus had said no more than that he had killed Constans, failing to identify how exactly he did this. So when a subsequent source in the chain of transmission read this source stating that the ‘controller of the basin’ had killed Constans while he was taking a bath, it would have been all too easy for him to assume, first, that this official bore this title because he was charged with using an actual basin during the bathing process and, second, that he had probably used this basin to kill the emperor as the nearest potential weapon by which to do so.

However, this is a highly speculative suggestion, possible rather than plausible, so perhaps one should prefer the alternative possibility instead, that the claim that Andrew had used a bucket to kill Constans had formed part of the story of his assassination ever since the imperial administration had first released details of what had happened or, to be more precise, of what it thought had happened. There seems to be no reason why those officials who were eventually tasked with investigating the death of Constans should have deliberately fabricated the claim that he was killed with a bucket, so one has to wonder why they might have thought this. The obvious suggestion is that it was clear, when they had recovered the body and had spoken to early witnesses of its condition after Constans’ death, that it had suffered a blunt force trauma to the head such as might have been inflicted by a bucket, and that these witnesses had also confirmed that a bucket had been found near at hand when the body was first discovered. However, it would hardly have been surprising if a bucket was found near at hand, because Constans was actually in the middle of a bath at the time of his death. Furthermore, it was not necessarily the only cause of death. He may well have slipped and have accidentally hit his head against the side of the bath or on the floor, all the more so if he was covered with soap at the time, as the transmitted story suggests. Here, one notes that an emperor, or a usurper, was as liable to accident as anyone else. So, for example, the emperor Theodosius II (402–50) fell from a horse, damaged his spine, and died a few days afterwards as a result, while the rebellious Byzantine general Saborius died in 668 when his horse bolted and dashed his head against a low archway.Footnote ²⁵ However, the difference between these accidents and that which Constans may have suffered, is that their accidents occurred in public, with multiple witnesses presumably, while his occurred, if it occurred at all, in private with minimal witnesses, perhaps just his servant Andrew. As for the latter, the mere fact that he fled the scene does not confirm that he played an active part in the killing of the emperor. It merely confirms that he was afraid of what would happen to him as the sole witness to, and chief suspect in, the death of an emperor.

It is my suggestion, therefore, that Constans was not really assassinated at all, but died rather of an unfortunate accident when he slipped and hit his head while in the bath. However, the lack of witnesses prevented the easy recognition of this fact, a situation that was exacerbated when the main, or even sole, witness fled the scene and used his family connections to ensure that he stayed hidden. The advantages of this explanation are that it explains both why there was no apparent connection between the death of Constans and the raising of Mezizios to the throne soon thereafter, because those who raised Mezizios to the throne really had not played any part at all in the death of Constans, and why the tradition preserved by Theophilus of Edessa should have claimed that Constans was hit on the head by a bucket, that is, had suffered blunt force trauma to the head rather than being stabbed as one would more normally expect in the case of an alleged assassination attempt, because he really had suffered a blunt force trauma to the head.