Over the past several posts, I’ve used the excellent history of the scholarly discussion surrounding the star catalog in the Almagest as given in Gerd Grasshoff’s History of Ptolemy’s Star Catalogue. At this point, we’ve caught up to the time the book itself was published. However, we’re not even half way through the book. The remainder of the text is Grasshoff reviewing the previous authors and offering a more comprehensive and in-depth analysis than done previously.
But before getting into that, Grasshoff attempts to ensure that we really understand the Ptolemaic catalog in its own right. For Grasshoff, this requires “a meticulous edition of the text and a correct identification of the ancient positional description with the stars known by their modern names,” as well as “the accurate position of the star at the historical epoch.”
Authoritative Editions
Putting together an authoritative text is no simple feat, as generations of scribal errors have corrupted the text itself, which makes identification of the stars challenging and, in turn, means we cannot be entirely certain of the historical positions.
The scribal errors are quite easy to achieve because of the similar shapes of many Greek characters – especially as the handwriting changed throughout the centuries. For example, A, Λ, and Δ1 all have a similar shape. Adding an easily lost tick mark to characters could also change their meaning. For example, Γ is $6º$ but Γ’ is $10$ arcminutes. Lastly, the shape of characters changed over time and sometimes characters like ε ($5º$) could be written more like a half circle with a horizontal line which, if the half circle was more complete, could easily be mistaken for θ ($9º$).
Various attempts have been made to compile the most reliable version of the text possible by comparing as many extant manuscripts as possible. The first of these efforts was C.H.F. Peters beginning in $1876$ when he reviewed the versions of the star catalogs compiled by Flamsteed, Lalande, Bode, and Baily and determined they were full of errors. He compared various manuscripts attempting to correct as many as possible until his death in $1890$, at which point his work was given to E.B. Knobel who continued the work and published their combined conclusions in $1915$. This work was based on $21$ Greek, $8$ Latin, and $4$ Arabic manuscripts.
Meanwhile, work was also being done on critical editions of the larger text. The first of these was done by J.L. Heiberg as a Greek text in $1898$. This was then translated into German by Karl Manitus in $1913$, both of which have been largely superseded by the Toomer translation I use for this blog.
However, Grasshoff also notes that a $1974$ translation of an Arabic manuscript by P. Kunitzsch is also of interest. In short, these are the texts to which Grasshoff will be referring as it is beyond the scope of his work to compile a new text.
Stellar Identification
With regard to identifying stars, Grasshoff notes that “[i]n the case of brighter stars, one is hardly ever plagued by difficulties.” However, for fainter stars there is less certainty. Grasshoff is critical of the Peters/Knobel star catalog as they “assume scribal errors in the manuscript merely on the basis of large deviations from the accurate positions and change them accordingly” whereas “Toomer and Kunitzsch preserve the more authentic version”.
Overall, Knobel considers $252$ of the stars in the catalog as having uncertain identifications. But with the later Toomer and Kunitzsch translations, Grasshoff considers there to be “only a few cases with still possible variations.”
Still, Grasshoff does his own stellar identifications using the entirety of the Bright Star Catalog of Dorrit Hoffleit2 with the positions adjusted to the epoch of $128$ BCE (since this is the time that best fits the data due to the longitudinal error) and notes that the quality of the data is sufficient that it should predict historical positions within $1$ minute of arc on average.
These positions are then compared to the positions from the Almagest taking into consideration several factors. In particular, various groups of stars “show varying systematic errors in longitude and latitude.”
In other words, in addition to the longitudinal error previously discussed, some constellations as a whole also show a second error with all the stars being shifted in another direction. Thus, Grasshoff takes this into consideration as well. Furthermore, Grasshoff takes into account potential transcription errors.
After taking these into account, Grasshoff reports that “about $5\%$ of the stars still correlated to another identification than Toomer’s.” Aside from disagreements with Toomer, there were still some other stars that were not well established, but Grasshoff states that “after Ptolemy’s description of the position of the stars within the constellation had been reconsidered, almost all traditional identifications could be confirmed. Only some few cases, about $15$, were either undecided or suggested a different identification.”
For example, he refers to star $836$ in the Almagest. This star is listed in the constellation of Canis Major (although it falls in the modern borders of Monoceros). This star has two potential identifications: either $19$ Mon or δ Mon. What is interesting about this star is that the other, well-identified stars in this constellation all only show a longitudinal shift, where as either of the potential identification shows primarily a shift in latitude.
Grasshoff also discusses the cases in which positive identifications should be made, but there are still large errors. This includes π Hya which is off by $-4.86º$ in latitude and $4.47º$ in longitude, as well as θ Eri which is pretty close in latitude (off by $0.37º$) but off by a whopping $6.81º$ in longitude.
The collection of manuscripts do not show any variance in the coordinates, so Grasshoff assumes there was an error in dating back to the original manuscript of the Almagest, or there was a serious malfunction with the instrument used to make the measurement3.
Ultimately, Grasshoff compiles his own star catalog in Appendix B of his work4.
- Which are $1º$, $30º$, and $4º$ respectively.
- Using the fourth edition which was published in $1982$. A fifth edition was published in $1991$.
- Having done a good deal of observation using my quadrant, serious malfunctions or misreadings strikes me as entirely plausible.
- The Toomer translation I’m using is the second edition which was released in $1998$ so it likely took into consideration Grasshoff’s identifications. However, I have not compared editions to see what changes were made.