In an effort to make sure the various manuscripts of the star catalog that I have copied into my spreadsheet are accurate, I have attempted to located images of the original copies when possible to review each value even when transcriptions are available. Each manuscript has its own unique eccentricities so, as I go through them, I’d like to dedicate a post to discussing the various things I notice.
To start with, I’ll discuss the Paris $2389$ manuscript, beginning with an overview of how to read the Greek. As a note, I’m not getting into translating the descriptions which is far beyond my skill level. Rather I’ll just discuss the coordinates and magnitudes since the catalogs generally keep the same order1.
Reading of the Greek Texts
Constellations
First off, the names of the Greek constellations are, well, Greek. A table is below:
| English | Greek | Abbreviation |
| Aries | Krios | kri |
| Taurus | Tavros | tau |
| Gemini | Didimoi | did |
| Cancer | Karkinos | kar |
| Leo | Leon | leoh |
| Virgo | Parthenos | par |
| Libra | Zygos | zug |
| Scorpio | Skorpios | skor |
| Sagittarius | Tokostis | toc |
| Capricorn | Aigokeros | Aig |
| Aquarius | Ydrohoos | uAr |
| Pisces | Ihteis | ixq |
While the first three characters are almost always present for the abbreviation, there are instances in which we see four characters or sometimes more.
Longitude/Latitude
Both of these values are expressed with a whole number as well as fractions, if necessary. However, instead of having unique characters for numbers, the Greeks reused letters. Furthermore, the Greeks didn’t think of the tens place the same way we did. Rather, they had separate symbols for the number of tens to which the ones would be added. First, let’s look at a table, and then I’ll provide an example:
| Number | Symbol |
| $0$ | o (Often with a line over it) |
| $1$ | a |
| $2$ | b |
| $3$ | g |
| $4$ | d |
| $5$ | e, |
| $6$ | ϛ2 |
| $7$ | z |
| $8$ | h |
| $9$ | q |
| $10$ | i |
| $20$ | k |
| $30$ | l |
| $40$ | m |
| $50$ | n |
| $60$ | c |
| $70$ | o |
| $80$ | p |
So if we wanted to say $55$ we would write ne.
To denote fractions, some had unique characters, but many were a number with an accent mark above it to indicate the inverse of that number. For example, d is $4$ but d‘ is $\frac{1}{4}$.
| Fraction | Character |
| $\frac{1}{2}$ | L’ (often with the horizontal branch trailing downwards) |
| $\frac{1}{3}$ | g‘ |
| $\frac{1}{4}$ | d ‘ |
| $\frac{1}{6}$ | ϛ‘ |
Another special case is $\frac{2}{3}$ which was written like the $\frac{1}{3}$ but with a circle inside the g.
For other fractions, such as $\frac{3}{4}$, the fractions were added together. For example, $\frac{3}{4}$ is Ld‘ and $\frac{5}{6}$ is Lg‘.
This system can cause quite a bit of problem. To see why, consider $10 \frac{1}{6}$ which is written as iϛ‘. If the accent mark is lost, this becomes iϛ which would get read as $16$. In comparing various manuscripts, this exact sort of error is extremely common.
The particular hand (uncial) is also problematic. I’ve specifically used a font here to best replicate the hand used in the Paris $2389$ manuscript but it isn’t perfect. However, consider the difference between a, d, and l. These characters all have a very similar shape and are easily mistaken for one another. Fortunately, the scribe for this document tends to make the horizontal stroke on the delta (d) extend past the legs and adds serifs to the end, but if the ink has faded, it can still be hard to tell.
North/South
To indicate which hemisphere of the celestial sphere the latitude is referring to, a simple abbreviation of bo (short of Borealis) is used. Similarly, for south we see no (short for Notus).
Magnitude
In general, the magnitudes are simply given as a number with the accent mark (again, indicating the inverse of the number). It’s interesting to see it expressed this way as generally magnitudes are thought of as whole numbers in modern astronomy. But expressing them as fractions makes it more intuitive as to why the magnitude decreases as the number (in this case, in the denominator) gets larger.
However, there are a few oddities. For example, Ptolemy sometimes indicates that the brightness of a star is brighter than or fainter than the indicated magnitude3. In this case, we get the same notation for the magnitude, but followed by a modifier.
I haven’t exactly figured out what the modifiers are supposed to mean and they seem to vary in how they’re expressed. But generally, if I see an em (often on top of one another and sometimes with other characters after it) it is brighter than. If I see a el (sometimes with the l as a superscript and sometimes with other characters after it), then it’s fainter than.
Also, for nebulous objects we read nef and for faint ones we get amau.
Paris $2389$ Manuscript
Now that I’ve covered the basics of the Greek manuscripts, let’s talk a little about things that I’ve seen in the Paris $2389$ manuscript.
History
This manuscript is believed to be one of the oldest known. Halma dated its origin to the $7^{th}$ or $8^{th}$ century. However, Knobel notes that this may only be part of the story. I’ll discuss it more below, but the handwriting changes notably half way through. The hand in the second half is more characteristic of this period, but the first half is more characteristic of the late $9^{th}$ century, according to Peters.
This manuscript was likely previously housed at the Laurentian library in Florence when it was bought by Catherine de Medici ($1519-1589$) who brought it to Paris. It them found its way to the Bibliotheque Nationale. It now bears a stamp in gold of Henry IV ($1553-1610$).
The Hand(s)
While the entirety of the text is in a uncial hand, there appears to be a change in the handwriting at the beginning of Book VIII when the text begins the southern zodiac.
There are several noticeable changes. The most obvious to me is that the o is notably rounder, almost circular whereas the previous hand is distinctly oval.
For the character q, the first hand had the horizontal bar extending well past the borders of the circle, terminating in serifs. The second hand ends them at the circle as is typical.
The characters are more consistent in size relative to one another. It’s like the later hand was written on graph paper and each character fit to the box. Previously, each character was its own size. The second is much more methodical.
Another oddity of the second half is the second character seen below:
This character is a variation on a delta (d) and evidently can be found as far back as the second century. It is not clear why it gets interspersed with the other form.
The, abruptly as this second had arrives, once we get to the constellation of Lepus, it returns to the original hand.
This variation in handwriting has led to debate over whether or not two scribes were involved.
Alternate Readings
One of the odd things I’ve seen in the Paris $2389$ manuscript is that values for the longitude and latitude are sometimes given two values. Here’s an example:
Here we can see the scribe clearly wrote kd ($24;00$) but above the d ($4$) they wrote a ($1$). This implies an alternate reading of $21;00$.
My initial thought was that this was a correction. However, there’s a few things that have changed my mind on this.
The first is star $432$ on folio $221r$, for which g ($3;00$) is replaced with g‘ ($\frac{1}{3}$).
Had the scribe truly intended to make a correction they could simply have added the accent mark. No need to rewrite the same character with the mark.
Second, there is evidence of what this scribe does for actual corrections such as for star $396$ on folio $220v$:
Here, we can see the scribe scratched out a d and replaced it with e. If the scribe’s intent is to make a correction, then it is likely they would have left no ambiguity similar to how they did here.
Lastly, the values that are the “corrected” ones, are frequently not in line with what critical editions have accepted as the most likely value. Rather, frequently, the original value frequently is the one that is generally accepted. So why would the scribe have “corrected” something to an “incorrect” value.
I think the answer for what the scribe intended is that this was their way of expressing ambiguity. In the first two examples above, these are instances of the two types of issues that can come as a result of the handwriting. In the first, the source they were copying from likely had an ambiguous character: d vs a. The scribe couldn’t discern which was correct, so they passed both on to us.
Similarly, for the second example, the scribe was likely uncertain if g‘ or g was the correct reading and offered us both.
However, there are some counterexamples to this hypothesis. Consider star $394$ on folio $220v$:
Here, the scribe has replaced an a with a b which is a distinctly different character. I find it unlikely that the source from which the scribe was copying was so muddled that they could not discern between these two characters.
Similarly, with star $444$ on folio $221r$:
Here, the scribe replaced L ($\frac{1}{2}$) with ϛ ($\frac{1}{6}$) which are distinctly different as well. Again, I find it difficult to believe that a scribe would have mangled them so badly that they could not be told apart.
However, Peters & Knobel offer another solution, suggesting that this may indicate “the scribe copied from more than one manuscript.” I find this extremely plausible as it can account for the cases in which the nature of a character should not be in question.
- There are occasionally small deviations in the order, often with two stars being switched. But this is easy to discern as you’ll see the coordinates switched.
- This character is called a digamma and it is drawn more like a modern lowercase delta in this manuscript.
- Toomer and Grasshoff both use $>$ or $<$ signs for this which I find troublesome. This is because the signs are attached to a magnitude. Thus, we should expect $>4$ to be something like magnitude $4.3$. But it’s not. What the signs are supposed to represent is the brightness which works inversely to the magnitude. Thus, $>4$ would be something like magnitude $3.7$. This strikes me as a very poor choice on their part and I only realized this when I compared to Peters & Knobel who would express it as $4-3$.