Did a Medieval Persian Astronomer Have Secret Knowledge of Brazil's Coastline?

Today’s story is a strange one, and I don’t think I have all of the parts to know exactly what happened or why. It begins with a passage I read in fringe writer Andrew Tomas’s mystery-mongering We Are Not the First (1971) concerning the work of the thirteenth century Persian astronomer Nasir al-Din al-Tusi. Here is how he gives the story:

The Azerbaijan Academy of Sciences (U.S.S.R.) made a discovery in 1964 that the 13th-century scholar Nasireddin Tusi was aware of the existence of America two hundred and fifty years before Columbus. The astronomer mentioned the land of Eternal Isles in one of his books, giving its geographical co-ordinates. When these were joined, the contour corresponded to the east coast of South America. Where did Nasireddin Tusi obtain his information about a faraway continent? 

When Tomas recycled the same passage, almost verbatim (and embedded in a longer discussion of maps repeated point for point), for his 1972 book Atlantis: From Legend to Discovery, he added additional detail that fleshes out the claim a bit more. I apologize that I don’t have access to the full text of the 1972 book in English, only in French, so I’ve had to patch together the following quote from an excerpt I had in English (in bold type) and the surrounding text from the French version:

A mathematician and astronomer of the thirteenth century, a native of Azerbaijan, Nasireddin Tusi also knew, 220 years before Columbus, something about the existence of America. G. D. Mamedbeily, of the Azerbaijan Academy of Sciences, has recently found out that seven centuries ago the scholar mentioned in his works the land of “Dzhezair Haldat” (“eternal isles”) the geographical coordinates of which correspond exactly to the eastern coast of South America. Just like Piri Reis, Nasireddin Tusi must have drawn his knowledge from archaic science.

What’s particularly disturbing is that Tomas’s version of the text, at least in its French form, is still being repeated as evidence of lost civilizations and ancient aliens as recently as this year, when it appeared on a number of Francophone and Spanish-language blogs, like this one. I have no idea what caused the claim to have a renaissance this year.
 
The two versions are subtly different, for the first makes the coordinates into a connect-the-dots, while the second simply places them in the same part of the Earth as South America. So, my next step was to try to figure out who the fellow was who proposed this idea in the first place. This led me to discover that G. D. Mamedbeili (in more typical transliteration) was the first Azerbaijani astronomer trained at the Leningrad State University in Soviet astronomical science, having graduated in 1938. He held a doctorate in physics and mathematics, and as an astronomer he became fascinated with Nasir al-Din al-Tusi, who in 1261 had built the greatest medieval observatory in the region. Tusi had been born in Khorasan but he is a national hero in Azerbaijan because the observatory he constructed in Maraqeh was in Greater Azerbaijan at the time, though it is now in Iran. Tusi also trained a generation of Azeri scientist-scholars. Mamedbeili wrote an entire book about the subject, Osnovate’ Maraginskoi observatorii Nasireddin Tusi, published in 1961, one of many works Mamedbeili devoted to Tusi.
 
You can see where this is going.
 
By sheer coincidence, Mamedbeili “discovered” that the Azerbaijani hero was also secretly informed of the existence of the Americas centuries before that Western upstart and knave Columbus, thus showering glory down on Azerbaijan.
 
I am not entirely familiar with all of Mamedbeili’s work, but I know that he laid out his theory in the oddball fringe research journal New World Antiquity in 1959 (vol. 6, no. 4) in an article called “The Problem of the Discovery of America.” (This weird journal is worthy of discussion on its own, running work by archaeologists alongside wacky material about diffusionism, the Kensington Rune Stone, etc.) The New World Antiquity ​article, in turn, was a translation and summary of a portion of an article from the Reports of the Academy of Sciences of the Azerbaijan S.S.R. vol. 12, no. 7 from 1956. The New World​ piece cannot be Tomas’s direct source since a Google Books search finds no mention of the Eternal Isles in the article; however, the eccentric transliteration of jazāʾir khālidāt as “Dzhezair Haldat” certainly speaks to a non-English source. Unfortunately, I don’t have access to either of these articles to know exactly what Mamedbeili said, but the gist seems to be that if one assumes that Tusi correctly estimated the size of the earth, then one must place his prime meridian along the eastern coast of South America, where his Eternal Isles (jazāʾir khālidāt) appear on his geographical chart.
 
To accept this is to do some mental gymnastics that speak to motivated reasoning. Abu’l-Fazl ibn Mubarak, writing around 1590, surveyed the various ancient and medieval geographies and concluded in his Ain-i-Akbari that the Eternal Islands were located beyond the Fortunate Islands in the Atlantic Ocean, approximately ten degrees of longitude west of the Atlantic coast of Africa. (He also seemed to think these were the islands of Atlantis, writing that they were sunken.) One could, on that basis, suggest that the Azores were meant, though this would be a century or more before their official discovery.
 
However, earlier medieval texts suggest that the Eternal Islands were the farthest Western point at which there was inhabitable land, but the overall thrust of the medieval sources, as the Islamic scholar F. Jamil Ragep pointed out in translating Tusi in 1993, is that the Eternal Islands were the Canaries, meaning that they had been accidentally duplicated with the Fortunate Islands, probably when geographers failed to understand that the Fortunate Islands and the Islands of the Blessed were two names for the same thing in Greco-Roman geography. We need not speculate, though; as far back as chapter 8 of al-Masu‘di’s Meadows of Gold, the Eternal Islands and the Fortunate Islands were equated. The Persian scholar al-Biruni, writing two centuries before Tusi, and preserved in Yaqut’s Mu’jam al-buldan in 1228, specified that the Fortunate Islands and the Eternal Islands were one and the same, and were both located off the coast of Africa. He adds that while there is disagreement over which meridian to use to calculate longitude, this disagreement is whether the prime meridian should be the meridian of North Africa’s west coast or that of the Fortunate Islands, as Ptolemy had preferred, located ten degrees to the west. But just to make things more confusing, al-Biruni’s Fortunate Islands appear to be the Cape Verde Islands and not the Canaries. In the fourteenth century, though, Ibn Khaldūn, writing in his Muqaddimah (1.116), specified that the Eternal Islands were the Fortunate Islands of Ptolemy, and thus from Ptolemy and other Greco-Roman sources we can presume the Canaries were meant. Al-Idrisi and Ibn Said concur.
 
But with so many candidates to choose from, why would anyone try to argue that the islands were not actually islands? This doesn’t match medieval Islamic geography, nor does it create a connect-the-dots map of Brazil even if we did accept it since Tusi gave one measurement for the Eternal Isles, not a series, so we can’t play “connect-the-dots” as Tomas asked us to do.