August 29, 2015

Abner of Burgos: a missing link between Nasir al-Din al-Tusi and Nicolaus
Copernicus?

Michael Nosonovsky, PhD

University of Wisconsin-Milwaukee,
3200 N Cramer St, Milwaukee WI 53211 USA,
nosonovs@uwm.edu

Abstract:

The theorem of geometry known as the “Tusi couple” was first discovered by Persian
astronomer Nasir al-Din al-Tusi (1201-1274) and independently rediscovered by
Nicolaus Copernicus (1473-1543). The Tusi couple played an important role in the
development of the heliocentric planetary system. It has been suggested that
Copernicus could be aware of al-Tusi’s result; however, a particular way of
transmission is not known, and several hypotheses have been proposed, including
Jewish links between the Muslim world and European scholars. The Spanish Jewish
author Abner of Burgos (1270-1340) was familiar with the Tusi couple. This may
provide a missing link in the transmission of the astronomic knowledge to Europe.

1
 The extent of influence of the Muslim medieval astronomy and mathematics
upon the European Renaissance and early modern science has been a topic of
intensive research. Several authors have shown in a series of articles that Persian
astronomers of the 13th century from the Maragha school in southern Azerbaijan
came very close to the discovery of the heliocentric Copernicus planetary system 1.
Although these scholars did not state explicitly that the Sun is located in the center of
the planetary system, their mathematical formalism was identical with that of
Nicolaus Copernicus and different from the earlier formalism employed by the
geocentric Ptolemaic system. Astronomers of the Maragha School abandoned the
Aristotelian axioms of motion and were able to advance further than their European
colleagues. Putting the Sun at the center of the celestial system remained a matter of
formal definition. Furthermore, there is indirect evidence that Copernicus (1473-
1543) was familiar with the work of Nasir al-Din al-Tusi (1201-1274), the founder of
the Maragha observatory, and that the discoveries of al-Tusi were the necessary key
element to finalize the Copernicus system2. Hartner showed that it is likely that
Copernicus was familiar with al-Tusi’s work, since there is a striking similarity
between the drawings of Copernicus in his De Revolutionibus orbitum celestium (“On
revolution of celestial spheres”) and those of al-Tusi in his Tadhkira fi ‘ulm al-ha’ya
(“Memorandum on the science of astronomy”)3. However, the direct evidence of
transition of the astronomical knowledge to Europe is very limited4.

1
E.S.Kennedy, “Late Medieval Planetary Theory,” Isis, 57 (1966):365-378; S. Kren, “The Rolling
Device of Nasir al-Din al-Tusi in the De Spera of Nicole Oresme?” Isis, 62 (1971):490:498;
W.Hartner, “Copernicus, the Man, the Work, and its History,” Proc. Amer. Phil. Soc., 117 (1973):413-
422; N.Swerdlow, “Derivation of the First Draft of Copernicus’s Planetary Theory,” Proc. Amer. Phil.
Soc., 117 (1973):423-512:490-498; N. Swerdlow and O.Neugebauer, Mathematical Astronomy in
Copernicus’s De Revolutionibus (Springer, NY, 1984); G.Saliba, A History of Arabic Astronomy:
Planetary Theories During the Golden Age of Islam, (NYU Press, 1994); J. Ragep, “Copernicus and
His Islamic Predecessors: Some Historical Remarks,” History of Science, 2007, 45:65–81
M. di Bono, “Copernicus, Amico, Fracastoro, and Tusi’s Device: Observations on the Use and
Transmission of a Model,” J. Hist. Astron., 1995, 26:133–154.
2
Hartner 1973:421-422.
3
Hartner 1973:422.
4
For a discussion see G. Saliba, “Rethinking the Roots of Modern Science: Arabic Scientific
Manuscripts in European Libraries”, Georgetown University, Washington, 1999; E. Huff, “The Rise of
Early Modern Science: Islam, China and the West”, Cambridge University Press, 1993.

2
 The al-Tusi work at the focus of the discussion is a theorem or geometrical
device called by Kennedy the “Tusi couple”5. The Tusi couple involves two spheres,
one of which has twice the same diameter as the other. The small sphere is inside the
big one and can roll upon it (Fig. 1a). Al-Tusi, who used this device for his lunar
theory, showed that a point upon the small sphere would move back and forth along a
straight line, and thus the straight periodic motion can be presented as a superposition
of two circular motions. Copernicus used the same mechanism for his (erroneous)
theory of trepidation and in his theory of Mercury. Although it was believed that
Copernicus re-invented the Tusi couple independently, Hartner presented evidence
that it was borrowed6.
While there are many examples of translation of the Arabic and Persian
astronomical works into Latin, particular ways of transmission of Eastern
astronomical material to Europe are not well known. Kennedy notes that the school in
Maragha attracted students from as far as Muslim Spain as well as from the East 7. It
has also been suggested that Byzantium could play a role of a way station in the
transmission of this material to Europe, because Hulagu, the grandson of Genghis
Khan and the patron of al Tusi and his observatory, was involved into diplomatic
relations with the Byzantine Emperor Michael VIII Palaeologus8.
In search of possible pre-Copernicus references to the Tusi couple in Europe,
Kren called attention to the manuscript De Spera by French scholar Nicole Oresme
(c. 1323-1382), written before 13629. In this work, Oresme describes how to produce
a back-and-forth rectilinear motion by combining three uniform circular motions and
suggests a mechanism very similar to the Tusi couple:
“It is possible for some planet to be moved perpetually according to its own
nature in a rectilinear motion composed of several circular motions… Proof: Let us
propose, conceptually, as do the astrologers, the A is the deferent of some planet or its
center, B is the epicycle of the same planet, and C is the body of the planet or its
center, I take these as equivalent. Let us also imagine line BC from the center to the

5
Kennedy 1966:370.
6
Hartner 1973:422.
7
Kennedy 1966:378.
8
Kren 1971:497.
9
Kren 1971:490-498.

3
epicycle to the center of the planet on which BC falls perpendicularly. Let circle A
move on its center toward the east and circle B toward the west. The planet C
revolves on its own center toward the east. Moreover, since line BC is of constant
length, because it is a radius, let us propose that the distance B descends in the motion
of the deferent is the distance at which point C may ascend with the motion of the
epicycle. From this one can obviously observe that point C in some definite time will
be moved in a straight line… Fourth corollary: Conceptually, it is possible for a
rectilinear motion to be eternal with the provision that at the point of return, the
mobile cannot be said to be moved or to be at rest.”10
The situation is shown in Fig. 1(b). My interpretation of the “Oresme device”
is slightly different from that of Kren, who did not provide any illustration; however,
it is consistent with the text. The difference between the “Tusi device” and the
“Oresme device” is that in the first device, the small sphere touches the big one and
rolls upon it, whereas in the second device “the distance B descends in the motion of
the deferent is the distance at which point C may ascend with the motion of the
epicycle”. However, kinematically this is the same motion. Oresme also does not
state that the deferent has the same radius as the epicycle, which is required in order
to have the same mechanism as that of al Tusi. Kren cautiously concludes: “Despite
speculation as to al-Tusi’s availability in the West and the likelihood that Oresme
would have found the eastern astronomer’s device attractive, the crux of the matter
lies in the difficulty of interpretation of the above passage. Confused and incomplete
as it is, one cannot say with surety that it does represent al-Tusi’s mechanism. We
must await the discovery of further material which may at least strengthen our
conjectures.”11
A different path of transmission was suggested by Morrison, who argued that
Moses Galeano could transfer the knowledge of the Tusi Couple to Copernicus.
Galeano was a Jewish scholar with ties to Crete and the Ottoman Sultan’s court, who
travelled to the Veneto around 1500. Morrison concludes that “while it is impossible

10
Kren 1971:490-491.
11
Kren 1971:498.

4
to prove that Galeano had direct contact with Copernicus, he most likely had contact
with some European astronomers.12”
Although the above-mentioned channels of transmission of Muslim
astronomical and mathematical concepts are quite plausible, I would like to call
attention to another possibility. Amazingly, the “Tusi couple” is mentioned in the
Hebrew mathematical treatise Meyashsher ‘Akov (“Straightening the Curved”) by an
author named Alfonso, found in manuscript in the British Museum13. The treatise is
devoted to the prominent problem of the quadrature of the circle, relation of the
curved and the straight, and to the philosophical problems of the motion. The author,
Alfonso, embraces Plato’s viewpoint in the Aristotle-Plato controversy14. He also
believes that besides the potential (koah in Hebrew) and actual (po‘al in Hebrew)
causes, there is also the intermediate (emtza‘i) cause. The intermediate cause
corresponds to the motion, which relates a finite point to the actually infinite line15.
Gluskina undertook the difficult task of identifying the author, and came to
the conclusion that Alfonso is Abner of Burgos (c. 1270-1340), a Jewish scholar from
Spain, who converted into Christianity at the age of 50 and acquired the name
Alfonso de Valladolid16. Although this claim by Gluskina was at first skeptically
accepted by some western scholars, later Gad Freudental and Shlomo Pines found

12
R. Morrison A Scholarly Intermediary between the Ottoman Empire and Renaissance Europe. Isis,
Vol. 105, No. 1 (March 2014), pp. 32-57:32.
13
First, I.Leveen, “Note on Some Names in a MS in the British Museum,” Jewish Quarterly Review,
13 (1922) and S. Luria “Die Infenitisimaltheorie der Antiken Atomisten,” Quellen und Studien zur
Geschichte der Mathematik, Astronomie und Physik, vol 2 (1933) 106-185 called attention to the
manuscripts (MS Add. 26984, vi, ff. 93b-128a, British Museum, London). In 1960, Luria (1891-1964)
obtained a copy of the manuscript and suggested that Dr. Gita Gluskina prepare its scientific edition
and translation. Gluskina’s book, which appeared in 1983 in Moscow after overcoming numerous
difficulties, includes a facsimile of the manuscript, scientific edition of the text, translation into
Russian, many explanatory notes and geometrical diagrams [G.M.Gluskina,,] Alfonso, Meyashsher
Akov (“Streightening the Curved”, in Russian), (Ser. Monuments of Literature of the Orient, Vol., 62,
ed., trans., commentaries by G.M.Gluskina, S.Ya.Luria, B.A.Rozenfeld, Moscow, Nauka, 1983).
Some commentaries were prepared by the Soviet historian of mathematics B.A. Rozenfeld.
14
Gluskina 1983:21.
15
Gluskina 1983:23-24.
16
Gluskina 1983:17-24. G. M. Gluskina, “On the Authorship of the Mathematical Treatise Meyashsher
‘aqob,” Palestinskiy Sbornik, 25 (1974) 152-156 (in Russian).

5
additional arguments in support of this identification and agreed that it is proven
beyond any doubt that Abner of Burgos wrote Meyashsher Aqov17.
Abner of Burgos was a theologian, physician, and astrologist18. In his
philosophical concepts he was an opponent of Aristotle and supporter of Plato. He
was familiar with astronomy, for instance, in 1334 he tried to convince the elders of
Toledo that they had erred in fixing the date of Passover, which would require
significant astronomical knowledge19. Although Abner wrote in Castilian, most of his
works were composed in Hebrew20. While there is no evidence that Abner knew
Arabic, he read Greek and Arabic authors in Hebrew translations21.
The Tusi couple is in Statement 33 of the treatise (see the Appendix). The
passage is not very clear and it includes a number of mistakes, however, it was
identified by Rozenfeld (who wrote his comments to Gluskina’s translations included
in the same book) as Tusi’s theorem 22. Rozenfeld states, however, that “Alfonso’s
proof is incomplete (or has not entirely survived), and the correct solution of this
problem was found only by Copernicus in his De Revolutionibus.” Rozenfeld also
identifies “Ridwan” in the passage as Abu-l-Hasan Ali ibn Ridwan (998-1061), an
astronomer, philosopher, and physician from Cairo. Rozenfeld does not discuss,
however, whether Abner could read the Arabic original or a translation of his work23.
Apparently, the problem that concerns Abner, is not whether the
straightforward motion can be presented as a combination of circular motions (which
is quite close to his main topic, straightening the curved), but whether the oscillating
point stops at the moment when it changes direction of its motion. Note that this is
very similar to the question raised by Nicole Oresme, whether “at the point of return,
the mobile cannot be said to be moved or to be at rest”. Note that the time of
completion of the works by Abner and by Oresme was very close. The treatise of

17
G. Freudental, “Two Notes on Sefer Meyasher ‘aqov by Alfonso, Alias Abner of Burgos,” Qiryat
Sefer, 63 (1990-1991) 984-986 (Hebrew), English translation in G. Freudental, Science in the Medieval
Hebrew and Arabic Traditions (Variorum, Aldershot, 2005 9. 1-4).
18
Z. Avnery, “Abner of Burgos,” Encyclopedia Judaica, (2nd Ed., 2007) Vol. 1.
19
Abnery 2007:97.
20
Avnery 2007:98.
21
Gluskina 1983:27.
22
Gluskina 1983:107.
23
Gluskina 1983:107-108.

6
Abner was written after he was baptized, most likely, in the 1330s, while the work by
Nicole Oresme was completed before 1362.
An independent discovery of the Tusi couple by several scholars in the 14th
century is not likely. The form of the Tusi device considered by Abner involved
touching rolling circles, and it was much closer to that of al-Tusi himself, rather than
the one considered by Nicole Oresme, although there is no data that the work by al-
Tusi was translated into Hebrew. Following Harter, who established that the Latin
notation of Copernicus followed the Arabic notation by al-Tusi24, we compared the
Hebrew notation by Abner with these two notations and found similarities (Fig. 1c).
Al-Tusi used the first five letters of the “abjad” system, A, B, G, D, H (which is the
sequence of the letters of the Hebrew alphabet) and the letter R. Abner utilized the
same first five letters plus the letters Z, X, T, and N. Like both al-Tusi and
Copernicus, Abner used A and B for the large sphere and G, D, H for the small
sphere, however, he only swapped the letters G and D at the small sphere. Oresme
just used first letters of the Latin alphabet.
There are several possible ways of transmission of the Eastern astronomic
knowledge to Europe. The Spanish way is possible keeping in mind that students
from Spain studied in the Maragha school and that Arabic astronomic works (e.g. by
ibn Ridwan) were well known to Jewish scholars in Christian Spain, such as Abner.
Further study is required to identify the exact path of transmission of these
concepts. However, one can conclude that the idea of the Tusi couple was known to
many scholars in Europe in the 14th century, and, in a more general sense, that the
development of astronomy in the late middle ages was a multinational enterprise
involving Muslim, Christian, and Jewish scholars. This advances our understanding
of the very nature of the European renaissance.

24
Hartner 1973:422.

7
‫‪Appendix. Hebrew text and English translation of the “Statement 33” by Abner of‬‬
‫)‪Burgos (Gluskina 1983:196-197‬‬
‫סגלה ל"ג נרצה לצייר תנועה ישרה מדובקת‬
‫ותמידית הלוך ושוב על קו ישר בעל תכלית‬
‫מבלי השקט בין הליכה ושיבה כיצד נשים קו א"ב הישר‬
‫ונרשום עגול אז"ב עליו מרכזו ג' ונרשום על קו א"ג עוד עגול‬
‫דח"ג נוגע עגול אז"ב ויהיו שלש אותיות אה"ד בנקדת‬
‫הנגיעה ונסכים שנסוב מרכז הכוכב שהוא ה' בעגול‬
‫דח"ג מנקדת ד' עד נקדת ח' בתנועה שוח בחלקיה ותסוב‬
‫נקדת ה' שהיא נגיעת שתי העגולות בתמועה שוה בחלקיה‬
‫מנקדת א' לנקדת ז' לחפך תנועת ה' בענין שתהיה‬
‫תנועת ה' כפל תנועת ד' ר"ל שכשתשלים נקדת ד' את כל‬
‫עגל דח"ג שתי פעמים תשלים נקדת ה' את כל עגול אז"ב‬
‫פעם אחת ונדביק שני קוי ג"ז ג"ח הישרים ולפי שקשת‬
‫ד"ח דומה לכפל קשת א"ז וזוית דג"ח במקיף העגול רוכבת‬
‫על קשת ד"ח וזוית אג"ז במרקז העגול רוכבת על קשת‬
‫ד"ח תהיה זוית דג"ח שוה לזוית אג"ז ולפי כשירכב קו ד"ג‬
‫על קו ג"ז ירכב קו א"ג על קו א"ג ותפול נקדת ח' שהוא‬
‫מקום מרכז הכוכב בקו א"ג על נקדת ט' נמצא שחלק‬
‫הכוכב מנקדת א' עד נקדת ט' וכן יתבאר כשישלים הכוכב‬
‫כל עגולתו פעם אחת ויגיע נקדת א' לנקדת ב' ויהיה‬
‫הכוכב בנקדת ב' וכן יחזור מנקדת ב' לנקדת א' מבלי‬
‫שיטה מקו א"ב לשום צד מהצדדים ואומר שלא ישקוט‬
‫הכוכב שום זמן לא בנקדת א' ולא בנקדת ב' מופתו‬
‫שאלו היה שוקט ילך הכוכב בזמן ההוא בעגלתו שום‬
‫קשת לפי שתמיד הוא מתנועע על העגול ונסקים שהלך‬
‫מנקדת ד' עד נקדת ג' ונדביק קו ב"נ ואם היתה נקדת ב'‬
‫נופלת על נקדת א' היה קו ב"נ שוה לקו א"נ והוא שקר‬
‫אם כן לא ישקוט הכוכב בנקדת א' הנה מצאנו התנועה‬
‫הישרה הנזכ מבלי שום השקט בין הליכה ושיבה ונראה‬
‫מזה שהסבה בהשקט הנמצא בין שתי תנועות המתהפכות‬
‫אינה הקצה והתכלית המצויה בפעל במרחק שהתנועה‬
‫עליו כמו שחשב ארסטוט והנמשכים לדעתו לפי שאותה‬
‫הקצה והתכלית לא היתה מפעל המתנועע כגי שישקוט‬
‫ברשומה אבל היא רשומה במרחק ואפשר שיעבור עליה‬
‫הגוף המתנועע מצד שהוא מתנועע מבלי שישקוט בה‬
‫וכש אם היה המתנועע נמשך בתנועתו לתנועה אחרת‬
‫סבובית ומדובקת או לכת שומר את הדבוק וכנו שהוא‬
‫בשני גלגלי נגה וככב חמה וכבר כתב אבו עאלי בן‬
‫רצואן שמה שאמרו שהתנועות המתהפכות יש השקט‬
‫בינותן אינו אמת תמיד וכי אין ההשקט המצוי בין תנועות‬
‫הדבוק כי אם כדי שינוח הכח המניע מיגיעו‬

‫‪8‬‬
 “Statement 3325. We want to present a connected straightforward continuous
motion forth and back upon a finite straight line without stopping between moving
forth and back. Let us take a straight line AB and draw a circle AZB with its center G
upon the line AB. Let us draw another circle DXG upon the line AG, touching the
circle AZB. And let three letters, A, H, D be at the touching point. And let us agree
that the center of the planet H rotates in the circle DXG from the point D to the point
X in a uniform motion26. And the point H [must be D] which touches the two circles
during the uniform motion from the point A to the point Z opposite to the motion of H
in the manner that the motion of H will be twice the motion of D. Thus when the
point D completes the entire circle DXG twice, the point H [may be D] completes the
entire circle AZB once27. Let us draw two straight lines GZ and GX, and, since the arc
DX is similar to twice the arc AZ, and the angle DGX in based on the arc DX and the
angle […]28 in the center of the circle is based on the arc DX. The angle DGX will be
equal to the angle AGZ29 and when the line DG passes to the line G[Z], the line AG
will pass to the line AG [must be GZ]. And the point X [must be H], which is the
center of the planet on the line AG, will pass to the point T30. We will find that the
planet has passed from the point A to the point T, and so it will be clear that when the
planet will complete its whole rotation once, and the point A will reach the point B,
the planet will be at the point B. And so it will return from the point B to the point A
without leaning to any [side] of [the sid]es from the line AB. And I say the planet will
not rest any time neither at the point A, nor at the point B.
The proof is if it rested, the planet would pass during that time a certain arc,
because it is always moving at its circle. Let us assume that it passed from the point D
to the point N and let us draw the line BN. If the point B fell upon the point A, the line

25
My translation using the Hebrew and Russian text in Gluskina 1993:85-86, 196-197, see the
appendix. Note that the letters he, daleth, and heth are similar to each other in the MS. Our
emendations are different from those by Gluskina in several cases, and we were able to reduce the
number of corrections in the text. Our schematic drawing (Fig. 1c) is also different. The notes in square
brackets are by Gluskina. The points naming translation convention is: A for aleph, B for beth, G for
gimel, D for daleth, H for he, Z for zayin, X for kheth, T for teth, N for nun.
26
The three points, A, D, and H are at the same point at the initial moment. The point A is at the large
circle, D is the touching point between the circles, and “the planet” H is attached to the small circle.
27
There is an apparent mistake here because the point H is not at the circle AZB.
28
Damaged in the MS.
29
The angle DGX (45o) is equal to half the angle AGZ (90o) at the initial moment of the motion.
30
At this position of the small circle, points T, G, and H are at the same point.

9
BN would be equal to the line AN [must be HN]31, which is wrong. Therefore, the
planet will not stop at the point A. Thus we found the above-mentioned straight
motion without stopping between going forth and back.
It is observed from this that the reason for stopping between two opposite
motions is not in the end or the actual objective at the space of the motion, as
Aristotle and his disciples thought, because that end and objective did not act upon
the moving [object] for it to stop in form, but was formed at the distance. And it is
possible that the moving body passes along it from the side that moves without
stopping in it. Even more if the moving [body] was passing from one motion to
another circular connected motion, or a motion that keeps a connection, as it is in the
motion of Venus and the Sun. And Abu Ali ben Ridwan has written that those who
say that between opposite motions there is a stop are not always correct. And that
there is no actual stop between connected motions, except for the moving force to rest
in its effort.”

31
The text is corrupted here, however, the idea is obvious: since the small circle rotates uniformly
(without stopping), if the planet H stops for some time at the point B, the small circle will pass in that
time from B to N, and the planet H will move from its original locations.

10
 (a) (b) (c)

Figure 1. (a) The Tusi couple. As the small sphere (DGH) rotates upon the big sphere

(ADB), the point A oscillates along the straight line AB. Original Arabic notation of

al-Tusi is shown in parentheses according to Hartner 1973:421. (b) The device

proposed by Nicole Oresme. The planet C oscillates along the line CB. (c) Our

reconstruction of the Statement 33 by Abner.

11