Here is some additional information. During some of the time he ruled, he was a Christian Vassel to a Muslim Sultan. A difficult position to be in.
http://www.roman-emperors.org/manuel2.htm
"The emperor, who on the coins still bore the title King and Autokrator, was as a vassal of course subject to the sultan's orders on campaign -- the sultan who amused himself at banquets, while the emperor discussed Islam with the Kadi. From October to December of 1391 the emperor enjoyed the hospitality of the Muderris (=Kadi) at Ankara. A Muslim born to Christian parents acted as interpreter between the emperor and the Kadi. The result of these conversations was the "Twenty-six Dialogues with a Persian," dedicated to his brother Theodore I. By 1399 the work had received its final editing. Presumably the emperor took notes at the time of the conversations. Apart from the emperor's writings there is no independent proof that the conversations ever took place. They must represent a mixture of fact and fiction. At the end the Kadi declared himelf ready to come to Constantinople and continue the conversation with Manuel. With this work, which must have been composed between the end of the campaign and the break with Bayazit (1392-94), Manuel made an important contribution to the knowledge of Islam on the part of the Christians."
Apart from the emperor's writings there is no independent proof that the conversations ever took place. They must represent a mixture of fact and fiction.I realize you're quoting someone else, but "they must represent a mixture of fact and fiction" is somewhat non-scholarly, to say the least. Why "must" they? Do historians agree that they must, or is that simply this author's opinion?