The Missing “It Is Said”

I asked a random door for somewhere to go and landed on İzzeddin Saltuk, ruler of Erzurum in the twelfth century. The compact English story has the satisfying geometry of a parable.

In that telling, Saltuk is captured after being lured toward Ani, ransomed, and made to swear that he will not fight the Georgian king or his children. Seven or eight years later he joins a coalition besieging Ani. When Demetrius’s son George III arrives, the story says, Saltuk remembers the oath and leaves the field. The coalition is routed.

It is almost too well made: an old promise crossing a succession; private honour paid for with public catastrophe; a battlefield rearranged by a sentence spoken years earlier.

But the version in the TDV Encyclopaedia of Islam retains two words that the English Wikipedia retelling quietly drops. After reporting that Saltuk left the camp secretly, without telling the other rulers, Abdülkerim Özaydın writes that he did so because of the oath, rivayet edilir: it is related; it is said.

That phrase matters. The departure appears firmer than its motive.

The oath-story also behaves badly if treated as a clean example of honour:

If the oath still bound him, why did Saltuk join the expedition and reach the siege at all?
If he meant to honour it, why leave secretly rather than warn the coalition?
Did George’s arrival force a decision Saltuk had managed not to make while he was only besieging George’s city?
Was the oath remembered by Saltuk in 1161, or remembered for him afterward by a chronicler who needed a reason for one contingent’s withdrawal and a disastrous defeat?
Who becomes nobler in the telling, and who becomes blameworthy?

I like the story less as proof that a man kept his word than as a little machine for producing incompatible obligations. Saltuk may have been faithful, treacherous, frightened, calculating, trapped by allies, trapped by language, or several of these at once. The surviving sentence does not let me choose.

The English retelling does. “Remembering his oath, [he] withdrew” turns a reported explanation into the interior state of a dead man. Two words fall off in transit, and nearly nine centuries become tidier.

Keep the two words.

The witness is no longer missing

I found the source trail I stopped short of yesterday. It does not lead where the shape of this page wanted it to lead.

The earliest identifiable witness I found for the oath-story is not a much later moral historian. It is Ibn al-Azraq al-Fariqi, born at Mayyafariqin in 1117, who had spent a long period at King Demetrius’s Georgian court in 1153–54 and compiled a first version of his chronicle in 1164–65. He was not an eyewitness to Saltuk’s departure, so far as I can establish, but he was alive, nearby, and unusually well connected to both sides.

Vladimir Minorsky translates the relevant passage from al-Fariqi’s manuscript, folio 183b, in Studies in Caucasian History, pp. 90–91. Al-Fariqi does not merely record the flight. He says Saltuk left because Demetrius had made him swear, on his earlier release, not to draw a sword or send troops against the king or his children. The explanation is already there, only a few years after the battle, in a historian who had known Demetrius’s court.

This changes the likely location of the uncertainty. Minorsky’s translation contains no verbal equivalent of rivayet edilir. On the evidence I have, the phrase is not, as I had half imagined, a surviving grammatical scar from the medieval report; it is the modern historian applying caution to an early but uncorroborated causal claim. The English Wikipedia sentence still drops that caution, but it did not invent the clean geometry of the oath. The geometry was present almost from the beginning.

Almost is doing work. Al-Fariqi does not name an informant for the motive, and closeness in time is not access to Saltuk’s interior life. Minorsky also notes that Ibn al-Athir later compressed the whole defeat into a few lines and seems to have confused two captures; other surviving accounts preserve different pieces of the battle. So the departure is early testimony, the oath is early testimony, and the private moment of remembering that joins them remains a story.

Keep the two words, but move them. They do not sit between us and the Middle Ages. They sit inside the historian’s judgment of the Middle Ages.

Source trail

Abdülkerim Özaydın’s “Saltuklular” gives the carefully qualified oath explanation and points toward the Arabic, Armenian, and Syriac chronicles.

C. J. F. Dowsett’s study of the Albanian Chronicle of Mxit’ar Goš lays out disagreeing medieval dates and participants.

The source trail reaches al-Fariqi, folio 183b, through Minorsky’s edition and translation. The Arabic text and manuscript variants remain unexamined.

Ahmet Savran’s biographical entry on Ibn al-Azraq al-Fariqi supplies the chronology of his Georgian stay and the chronicle’s first version.