The autograph of Copernicus's De revolutionibus orbium coelestium

We should begin by explaining the meaning of 'autograph' in our title. It is not the usual meaning of the signature of a person in their own handwriting, but rather the more general meaning of a handwritten document written in the hand of the author. We have not used the term 'manuscript' since a manuscript does not need to be in the hand of the author. Today this priceless autograph of De revolutionibus is kept in the library of the Jagiellonian University in Cracow, Poland. It has not been there since it was written by Copernicus but has had a complex, and at certain times unknown, route to reach there.

For a history of how the autograph copy reached the Jagiellonian University in Cracow, Poland, see THIS LINK.

The formal description of the autograph in the library is as follows:-
Jagiellonian University Library, Cracow, Ms BJ 10,000. Latin and Greek. Written about 1520-1541. Paper, 28 x 19 cm, 213 leaves, 2 paste-downs, 4 flyleaves. Sixteenth-century humanistic cursive. Bound at the beginning of the seventeenth century in a parchment document.
The year 1973 marked the 500th anniversary of the birth of Copernicus and to celebrate this event the Polish Academy of Sciences published a 3-volume work of the Complete Works of Nicholas Copernicus. It was published by the Macmillan Press, London, and the Polish Scientific Publishers, Warsaw, and appeared in 1972. The first volume contained a facsimile of Copernicus's autograph of De revolutionibus.

The beginning of the Introduction to this publication of Copernicus's Complete Works is at THIS LINK.

For the Introduction to Volume II, an English translation of De revolutionibus, see THIS LINK.

Volume 1 begins with a section entitled The Analysis and History of the Manuscript written by Jerzy Zathey and much of the information in the rest of this article comes from this section.

Perhaps the most obvious question to start with is, "Are we certain that this manuscript copy of De revolutionibus is in Copernicus's own hand?" First there is a declaration on the manuscript made on 19 December 1603 by Jakob Christman that the manuscript is in Copernicus's own hand. Of course we cannot take this as a proof since Christman, Dean of the Faculty of Arts at Heidelberg University, owned the manuscript at this time and clearly it has a far greater value in Copernicus's hand than if written by a scribe so it would greatly benefit Christman to make this declaration even if it were false.

Before looking at further evidence that the manuscript is in Copernicus's own hand we should note that there was a manuscript copy of the autograph sent to the printer by Georg Joachim Rheticus. Now Rheticus had spent two years with Copernicus at Frauenburg in 1539-41 and it was Rheticus who arranged for the publication of Copernicus's De revolutionibus in Nuremberg in 1543. This publication was from the manuscript copy which, in fact, has many improvements and corrections from the autograph copy. We will say a little more about this towards the end of this article.

Having been less than satisfied that Christman's declaration that the manuscript is in Copernicus's own hand is a proof, we need further proof. This, however, is fairly easy to obtain since quite a number of letters by Copernicus's still exist and handwriting experts can verify with certainty that these are in the same hand as the autograph of De revolutionibus.

There is one further complication which we must mention at this point. The writing in the manuscript is written in two different styles. Are these both due to Copernicus? The answer by handwriting experts is a definite 'yes'. When we write using computers today we often use two different styles, one being the 'normal' style, the other being a slant 'italic' style. This is exactly what Copernicus does, he has two different styles of handwriting one corresponding to 'normal' or 'formal' and the other to 'italic' or 'cursive'. Both styles [4]:-
... are typical of Renaissance humanism and quite different from the models followed by the medieval scribes.
The article [4] describes in detail the way Copernicus wrote. Here, by way of an example, is the description of how he wrote a capital A:-
Its left-hand stroke, made by an upward motion of the quill pen, usually starts without serif. It is thin at the bottom and gradually thickens as it slopes forward and upward. From the top of the left-hand stroke the pen descends vigorously and almost vertically without being lifted to form the right hand stroke. Stopping suddenly, the pen is raised to proceed from right to left in the horizontal stroke, which slopes slightly downward. It may sometimes start to the right of the right-hand stroke, and usually it ends well past the left-hand stroke.
We have given this illustration in part to illustrate how experts can be certain that Copernicus's letters and his autograph of De revolutionibus is by the same hand. There are a few minor marginal notes on the autograph made in another hand which has been verified as belonging to Rheticus.

What about the paper that Copernicus was writing on? All the paper has watermarks which, in theory, should make it possible to determine the particular mill from which the paper was produced and an approximate date of production. Let us note that watermarks are produced by putting an imprint into the paper while it is still wet which explains the 'water' part of the name. Although in theory the watermark identifies where and when the paper was made, in practice this is difficult to do. In the case of the autograph of De revolutionibus four distinct watermarks are found and each of these has a number of variants. The four types have been designated C, D, E, and F. If you are wondering why the naming does not start at A and B see later in this article.

The C type of watermark, a serpent, has not been decisively identified but similar marks were made in southern France and also in Holland. It seems more likely that this was manufactured in Holland around 1520-1525. The watermark D is of a hand with fingers spread out beneath a crown resting on the middle finger and seems likely to have been manufactured in central France in the 1520s. It is the poorest quality paper in the work. The watermark E is of the letter P with a four leaf flower sitting on top. It was manufactured in Maastricht, Holland around 1537 and was used by Copernicus in letters dated August 1537 and March 1539. The final type designated by F is of a hand with fingers together and a three leaf flower on top of the middle finger. The paper with this watermark only occurs on a few sheets and was almost certainly inserted when the manuscript was in the final stages of completion. A sheet with the same watermark was used by Copernicus in a letter dated 15 April 1541. Paper with a similar mark was produced in Osnabrück in 1538 and Lorraine in 1540; both these are possible sources.

What does the distribution of these four papers tell us about how the autograph version of De revolutionibus was produced by Copernicus? Before examining this question we must look at the way the sheets of paper are assembled in the autograph.

The sheets of paper were assembled by Copernicus into quires. The term quire comes from an old word for 'four', sometimes referred to as a quaternion, and consisted of four sheets of paper laid on top of each other, secured in a join down the middle, then folded along the join. Usually the original sheets had approximate size 1:√2 so that when they are folded in this way the ratio of the breadth to length is the same. When folded in this way it formed eight sheets which in a modern book would consist of sixteen pages since both sides of a sheet are used and numbered. The autograph of De revolutionibus consists of twenty-one quires but Copernicus has normally used five sheets, rather than four, so his quires are quinternions with typically ten double pages and twenty pages in modern book notation. This typical size of quire is, however, changed by Copernicus when he sometimes adds an extra sheet to obtain a sexternion. The sixth sheet of paper is one manufactured later (see below) than the other five so almost certainly these sexternion quires were produced by Copernicus when he wanted to add extra material. In a couple of cases he has removed a sheet from a quinternion to form a quaternion when he has decided to change the order of material. We should note that some quinternions have had one sheet removed and replaced but a sheet manufactured later as Copernicus has made alterations to his text.

There is no simple progression so the idea that first he used C, moving on to D, then E and finally F is just not born out by the way the quires are made up of the papers. There is plenty of evidence of quires with both papers D and E which must have been made up by Copernicus in this way before he wrote the text on them. Why he has done this is a bit of a puzzle. Paper F has only been used for a few late additions and this has been used after the work was almost completed. What is suggested by the way these papers are used is that much of the work was originally on quires of papers C and D. Then Copernicus made extensive revisions to the text using papers D and E. Of course the text itself refers to various dated observations and it might be thought that these would be useful in dating parts of the autograph. This, however, is not the case since we only know that he must have written this after the date of the observations but we have no idea how long after. In fact these dated observations do not add to the dating implied by the paper types, simply helps to confirm it in some cases.

Finally in discussing the paper, we should note that even given a precise date on which the paper was manufactured gives only clues to when Copernicus used it. Of course, it would establish a lower bound but after paper is manufactured it may be quite a while before it is purchased by a customer and even after it had been purchased by Copernicus, he may have kept it for quite a while before using it. The final conclusions would be, therefore, that the autograph was composed between 1525 and 1541. In fact he may have begun a few years later than 1525 but certainly worked on it extensively in the late 1530s up until 1541.

Now we have seen that Copernicus had 21 quires forming De revolutionibus. It is reasonable to wonder if he bound them together. In fact we know he did not for it is evident that the autograph has only been bound once and that was in 1603. Had it been bound by Copernicus and rebound in 1603 we would see evidence of two sets of holes but this is not the case. It appears that Copernicus kept everything in a box which has long since vanished. He did letter the quires a, b, c, ... so that their correct order could be maintained. When they were bound in 1603 a parchment, which seems to have been used by Rheticus to keep the quires together, became part of the binding and at this stage further sheets were added which are identified by watermarks and now known as A and B. Both paper A, which is a blank flyleaf, and paper B which makes up two further front flyleaves and two rear flyleaves were manufactured after Copernicus's death. Paper A was manufactured around 1550-60 and B around 1580-1600.

A summary of the binding is given in [4]:-
... our investigation of the binding of Copernicus' autograph would indicate that the author himself kept his quires loose and unbound. When the autograph came into the hands of Rheticus, it was still unbound. Later a discarded legal parchment was used as a protective wrapper. After the autograph had been transported to Heidelberg, it was bound around 1604. It was then that pieces of the page proofs of a book printed at Heidelberg in 1603 were inserted as padding inside the parchment wrapper.
We mentioned above the differences between the autograph version of De revolutionibus and the version that Rheticus gave to the printer and which became the printed text published in Nuremberg in 1543. In [3] Swerdlow denotes the autograph by M and the 1543 printed text N. He writes:-
M is a composite document, written over a period of years and containing hundreds of internal corrections and even substitutions of entire sheets. M is, of course, immensely valuable for studying the evolution of Copernicus's technical work, and sometimes clarifies derivations that are inconsistent or incomprehensible in N since cancelled numbers that actually stand behind some of the derivations in the revised version may be retrieved in M. However, since M contains a stage of the text prior to the revision printed in N, it must be used with caution in correcting N, from which it differs in hundreds of particulars, both significant and trivial. While M was substantially complete when Rheticus visited Copernicus in 1539, it may have received corrections and additions until about 1541.

As it stands, M contains so many serious inconsistencies that Copernicus would never have consented to its being printed without revision. One needs only consider the part on the mean motions of the planets, in which the text and the tables contain distinctly different values, an obvious absurd situation. It is the greatest irony, indeed the greatest injustice, that modern editors have chosen to base their texts on M. Pity the author whose drafts are published as his finished work.

I believe it is necessary to postulate at least two further copies of the text between M and N. The first would be the working copy in which Copernicus, possibly with the aid of Rheticus, made the revisions that appear in N. The revisions are so extensive and, in some cases, so complicated, that it is almost inconceivable that Copernicus would have directly written a fair copy on the basis of M without passing through additional work sheets and a complete working copy, a copy which nevertheless carried over some purely scribal errors from M. The second would be a fair copy sent to Nuremberg, and this could have been executed by a scribe from the working copy, preserving its errors and perhaps adding some more.
Finally we note that a few copies of the printed version contain an errata sheet. Certainly this could not have been anything to do with Copernicus if it was produced after De revolutionibus was printed because he had died by this time. Of course he may have seen errors in his working copy, if such existed, and the errata list, which only corrects errors in the first half of the text, may have been his work. More likely it was the work of Rheticus, but one then has to ask why he did not produce an errata list for the whole work.

References (show)

  1. E Rosen, When did Copernicus write the 'Revolutions'?, Sudhoffs Archiv 61 (2) (1977), 144-155.
  2. N M Swerdlow, The Holograph of De Revolutionibus and the Chronology of its Composition, Journal of the History of Astronomy 5 (1974), 186-198.
  3. N M Swerdlow, On Establishing the Text of De Revolutionibus, Journal of the History of Astronomy 12 (1981), 35-46.
  4. J Zathey, The Analysis and History of the Manuscript, in Complete Works of Nicholas Copernicus (Macmillan Press, London; Polish Scientific Publishers, Warsaw, 1972), 1-25.

Additional Resources (show)

Written by J J O'Connor and E F Robertson
Last Update January 2019