This small text is the first in a list about the problems of hieroglyphic encoding in practice (actually, it could qualify as the second post on the subject, as my previous post about sign D396D396 could qualify too).
I was recently asked a question about the A90 A90 sign in the JSesh library. The A90 sign is directly adapted from the Hieroglyphica, and I would originally say that :
- it could be described as a man, standing and bending, with a curvy line flowing from his head, his arms twisted in a unnatural way;
- the use of A90B as code for A90B made me think of an "ennemy" sign in some royal epithet, something like "ḥw pḏ.t 9".
In the original Manuel de codage, the sign is classified below men bowing or praying, and looks like this (MdC, p. 55) :
Of course, this classication doesn't fit my assumption that the sign might be an ennemy or a prisonner. Let's explore a bit more.
Uses in encoded texts
Now, the part relevant to encoding practices. I had a look on my computer, and saw that the sign was refered by Bob Richmond as having a source in Faulkner's concise dictionary. A quick search on two files derived from Faulkner's dictionary (in particular Mark Vygus's list) gave me the following words :
- jꜣkby, mourners
- wnwn, to sway to and fro, to travel about (p. 61), with refs to TR 23, 69.70 (that is Lacau, Textes religieux égyptiens, which is an early edition of Coffin Texts)
- ḫfty, enemy
After looking in the actual Faulkner :
- in the case of jꜣkby, the sign hasn't go twisted arms ;
for wnwn, the drawing in Faulkner is very sketchy, and looks like a cursive hieroglyph. it's probably the same passage as DZA 22.380.330. Seing that is an alternative sign, the lines in front of the character might be his hair, not tears nor blood. There are other examples, where the subject is trees, which are performing a kind of adoration gesture for the Sun or for Amun (ex. Leiden I, 350, ro 2-6 V7:n-nw-w-iAm-iAm-iAm-nb-Hr:Z1-wn:n-wn:n-y:D54-n-Hr*Z1:V31 šnw nb ḥr wnwn n ḥr=k, all trees are swaying to and fro in front of you). Given that the free movement of plants is limited, the image is that of trees swaying and moving their foliage as if it was hairs. We might even translate the trees swing their foliage in your sight.
- ḫfty would be quite on the spot for an enemy sign, but this spelling is not in the original Faulkner - and I don't have a reference.
The sign is also known (thanks to R. Montfort for a reference in Michel Guay's GoT) in Rekhmirê, II, pl. CII, as determinative for a word twtj (also PT 591).
Horus has clad himself with his seshemet belt,
In the PT. version, the text is translated by Allen as:
Horus has arrayed himself with his malachite sporran that strides over his land in full
walking on his country as a twtj
In the Rekhmirê version, an =f has been added after nmtt, which is no longer a participle, and the A90 determinative has been added after the the last word. It doesn't fit with the translation "in full". The PT tend to avoid animate determinatives, but in this case, it's also possible that the New Kingdom scribes have reinterpreted the original text (as the addition of "=f" suggests).
Note : twtj seems quite an hapax. Could the m be an haplography, in which case it could be a nisbe on the line of "mwtti" = he of the dead, which would be a proper name for a mourner ?
Now, let's have a look at A88 A88. In the rendering used in JSesh, the man is crying (quite a lot). But, after this little study, I'm not sure at all the drawing is right. From what I know of Egyptian iconography, it's rather possible that the lines is indeed uncombed hair.
Now, without a proper description of the sign in the first place, it's next to impossible for a font creator to know what solution is the right one (or if both solutions are possible).
Link between A90 and A90B ?
Might the twisted arms remind of the ḥ sign in A90B?
Now, this is an interesting warning about the state of our encodings, and their uses. Basically, most of the encoded hieroglyphic files we find were not created with databases in mind, and even when they were, the level of care put in those files went into making the general text clear; the sign-level faithfulness was a secondary thought. Actually, the text creators, when they tried to render the text spelling, ended up using the visually closest sign they could find, even if the actual value of this sign was quite different from that of the original sign in the text.
This is a very general trend I have seen in texts encoders, and I don't blame them: it's relatively innocuous when used for printed texts, specially if you add some explanatory footnote.
The conclusion from this little exercice is that, as long as we don't have both a good hieroglyphic database, and very thoroughly typed texts, with a consistent view on how to encode the glyphs, computer searches of hieroglyphic database at the sign level can only be used if they are double checked afterwards on the original document.
Another conclusion is that this little study took me a few hours. And that's for one sign. It means it's very dangerous to add, without proper descriptions and second check, large lists of glyphs to Unicode.