On 4 June 968, Liudprand of Cremona arrived in Constantinople on an embassy to the Byzantine Emperor Nikephorus Phokas. His mission, undertaken on behalf of the Holy Roman Emperor Otto the Great, was to ask Nikephorus for the hand of Anna Porphyrogenita, daughter of the former Eastern Roman Emperor Romanus II, for his son, the younger Otto (afterwards Otto II). As Liudprand narrates in the account of his embassy, the Relatio de legatione Constantinopolitana, from the beginning the mission was a disaster. The house he was confined to neither kept the cool inside nor shielded its inhabitants from the heat. It was remotely placed, and guards were posted outside of it, who stopped Liudprand and his retinue from going out, and others from coming in. This calamitous confinement was aggravated by the fact that the wine of the Greeks was undrinkable, mixed as it was with resin, tar, and plaster. Furthermore, there was practically no water to drink.Footnote ¹ To these travails, another was soon added: ‘Huic magno vae, vae aliud appositum est: homo †sciliorum† custos, qui cotidianos sumptus praeberet, cui similem si requiras, non terra, sed infernus forsan dabit.’ (‘To this great woe another woe was added: a guardian †sciliorum†, who oversaw the daily consumption, whose likeness, if you look for it, will be found not on earth but rather in hell.’).Footnote ² Thus prints Paolo Chiesa, the most recent in a long line of scholars charged with editing the Latin text: the obeli placed around the word ‘sciliorum’ signal a locus desperatus. Indeed, it appears that no scholar or critic who has edited or studied the text has been able satisfactorily to make sense of the word: there is no Latin word ‘scilius’ or ‘scilium’ of which ‘sciliorum’ could be the genitive plural, and the passage seems undoubtedly to be corrupt.

This note will first review the conjectures proposed by scholars to emend the passage, and then propose a solution to the problem. Before doing this, however, it will be necessary to review briefly the textual tradition of Liudprand's Relatio. There is none. No manuscript of the Relatio has survived.Footnote ³ Our only source for the text is the printed edition by Henricus Canisius of Ingolstadt (Hendrik de Hondt, ‘The Dog’, a nephew of St Peter Canisius) published in 1600.Footnote ⁴ According to Canisius’ preface, the edition is based on the transcription of a manuscript kept in the library of the cathedral of Trier. This transcription was made by Christopher Brouwer, a priest of the Society of Jesus and an ardent antiquarian. Brouwer sent his transcription to a friend, Marcus Welser, who brought it to Canisius’ attention.Footnote ⁵ The Trier manuscript is now lost, and Canisius’ print of Brouwer's transcription remains the oldest extant version of the text and forms the basis of all subsequent editions.

Although the manuscript upon which Canisius’ edition is based is lost, judging by other texts that he printed and whose base manuscript survives, the editio princeps is likely to be a faithful reproduction of the Trier manuscript.Footnote ⁶ For some passages, however, which may have been corrupt already in the Trier manuscript or in Brouwer's transcription, Canisius proposed alternative readings which he printed in the margins. One such passage is the one above, which in Canisius’ text reads ‘homo, Sciliorum custos’, and receives the marginal note ‘al. Siciliorum’,Footnote ⁷ implying that there is something wrong with the text as it stands, and that an alternative reading, alias or aliter, suggests that the guard may have been of Sicilian origin.

Previous editors of the text have proposed various emendations to explain the contested passage. Pertz, in his 1839 MGH edition, cautiously prints ‘homo scili … orum custos’ in the text, but in the apparatus presents his full conjecture, ‘homo scilicet nostrorum’, and suggests domorum as a potential misreading of nostrorum.Footnote ⁸ In the text of his 1915 MGH edition, Becker repeats the ‘homo scili … orum custos’ of his predecessor, but adds, in the apparatus, the abbreviations that could possibly have been misread by the scribes (for scilicet and nostrorum).Footnote ⁹ Bauer and Rau also print ‘scili … orum’ but capitalize on the marginal suggestion of the editio princeps, and propose in a note that ‘Siciliorum’ should be read, indicating perhaps a ‘Befehlshaber einer ursprünglich aus Sizilien rekrutierten Truppe’.Footnote ¹⁰ Why Liudprand would have regarded it as an insult to be guarded by someone from this island is not explained. Koder, most recently, suggests that the corruption could be cured relatively easily by reading scolarum, and that the guard thus would have been a member of the σχολαí stationed in the palace, which would yield the text, ‘Huic magno vae, vae aliud appositum est, homo scolarum, custos …’Footnote ¹¹ Again, I do not know why being guarded by a man defined merely as being of or from the palace cohort would have been such a humiliation. At any rate, Chiesa accepted none of these conjectures and, as we have seen, despaired of the reading. Also Bougard, whose recently-published French translation contains a facing page Latin text, repeats Chiesa's obeli.Footnote ¹² This is surprising because Bougard, in a footnote, mentions the suggestion made by Immanuel Bekker buried in a note to his edition of Codinus Curopalates, De officialibus palatii Constantinopolitani Footnote ¹³ that it could be the word σκύλλος (‘dog’) that stands at the root of the reading. Although he did not explore the full consequences of Liudprand's use of the word, to my mind Bekker was on the right track.

Contrary to previous critics, therefore, I do not think that there is anything wrong with the transmitted text and I would like to suggest that the problem may be solved if we only pay attention to Liudprand's use of Greek. In Liudprand's writings, alongside Greek words more or less correctly written in Greek characters, Latinized Greek words are a commonplace. The Relatio alone abounds with both proper transliterations such as kosmocrator for κοσμοκράτωρ (12.210 and 17.271), parakinumenos (15.247), which is slightly off for παρακοιμώμɛνος (‘night-nurse’), and distorted ones such as porniogenitum (12.205), a hybrid form construed from Greek πορνɛία ‘fornication’, and Latin genitus ‘born’, and porphyrogeniti and porphyrogenita (15.253), ‘born of the purple’, modelled presumably on such a proper Greek word as πορφυροφόρος (‘purple-wearing’).

In light of this, sciliorum need not be a misreading for something else but appears rather to be a Latinization of the Byzantine Greek word for dog, σκύλος, or, even more likely, its neuter form, σκυλίον, both derived from Ancient Greek σκύλαξ. This was transliterated, by Liudprand or his scribes, into Latin as scilium and used in the text in its genitive plural form, making the man in the passage a ‘guardian of dogs’, or ‘dog-watcher’. Perhaps even more likely, sciliorum could be a hybrid Greco-Latin form, skili-orum, ‘dog-faced’.Footnote ¹⁴ If the former is right, assigning a watchman of canines to provide food supplies to an imperial ambassador's house would obviously add to the insults already listed by Liudprand and make a good fit at this point in the narrative. Even more so, perhaps, if the latter is correct and the guard had the face of a dog. This is also suggested by the immediately following statement: ‘cui similem si requiras, non terra, sed infernus forsan dabit’.Footnote ¹⁵ Liudprand's observation that the guard, with the face of a dog, seemed to issue from hell rather than from earth, carries with it the allusion to Kerberos, the watch-dog of the underworld. Such a coinage and word-play would not be unfamiliar to Liudprand's usual practice. The Greek word for dog, furthermore, σκύλος or σκυλίον, is not incidentally connected with Σκύλλα, the notorious sea-monster. Moreover, just like Skylla, barking like a dog from her cavern in the Straits of Sicily, vexed (σκύλλω) passing sailors, so Liudprand's dog-faced watchman, in the immeditely-following section, makes life a living hell for his guests: ‘is enim, quidquid calamitatis, quicquid rapinae, quicquid dispendii, quicquid luctus, quicquid miseriae excogitare potuit, quasi torrens inundans in nos effudit. Nec in centum viginti diebus una saltem praeteriit, quae non gemitus nobis praeberet et luctus.’ (‘for he poured out like a deluge over us every kind of mischief, plunder, affliction, and misery that he could think of. Not one single day in one hundred and twenty passed in which he did not serve us affliction and woe.’).Footnote ¹⁶