What are the principles of Middle Egyptian spelling?

Question

In this question, the transliteration jdnw is given as the Middle Egyptian for the equivalent of "lieutenant". If I now want to render that in hieroglyphs, how would that work?

1. I could presumably 'spell' it out in uniliteral signs as reed/hand/wave/quail, with an appropriate determinant. I would assume that is 'possible', but not how a Middle Kingdom scribe would have done it.

2. I could look for the longest multi-lateral sign, which happens to be F21 (jdnw), and then add a uniliteral sign or two to complement the sound (to make sure it's not interpreted as the actual ear of ox), effectively spelling it jdnw d (n) plus a determinant

3. Assuming that there was no single multi-literal (as there happens to be in this example), I could go for the longest match; suppose there was a sign for jd, I could use that, and then repeat the process for the remaining nw until the word is completed; optionally supplementing the partial matches with uniliteral signs to again stress the phonological use over the literal one. Kind of spelling it jd j nw n.

4. I could try to look the word up in a dictionary and give up if I cannot find it.

I would think that (2) is the ideal case, but that (3) would be the most common fallback as there are more words than signs; (1) would be the last resort (probably for names or unknown/foreign words).

Is that how hieroglyph spelling works?

The background to my question is that I'm wondering about a transcription-to-hieroglyphs program (which might well use JSesh notation as an intermediate output format), and (2) could probably quite easily be implemented, given a list of signs with their transliterations.

When I look the word up in a dictionary (as idnw), there are two spellings given:

1. F21 N35 Z7 W24 (jdnw n w nw), and
2. F21 W24 N35 (jdnw nw n)

Here the second one is kind of what I'd expect, but the first one is a bit odd, in that the two uniliteral signs n and w are followed by a biliteral nw. But I assume these are attested spellings, so it would be a valid spelling?

Answer 1

Native words* are written in one of two ways. Either they're written with a single logogram, which represents the entire word, plus a stroke (Z1) that marks it as a logogram. Or they're written with a series of phonetic signs that represent the word's consonants, plus a "semagram" or "determinative" that hints at the meaning of the word (such as A1 for a person).

For this second option, there are a lot of ways to go about it. You can use purely uniliterals, if you want, or biliterals plus uniliterals for redundancy, or biliterals alone. Hieroglyphic writing was meant to look good, and these options give the scribes more flexibility in shaping it to their aesthetic preference.

The usual goal was to arrange the signs into "quadrats", or boxes, with as little empty space as possible. If you were trying to write jmn "(the god) Amon" horizontally, for example, reed-owl-water-god would not be ideal—the long flat "water" glyph is trapped between the tall "owl" and "god" glyphs, with empty space above and below it. Instead, you could use reed-game board-water-god, with the game board (mn) being long and flat; now you can put the board above the water, and everything fits together nicely.

If you were writing it vertically, on the other hand, reed-owl-water-god would be better: reed-game board-water-god would strand the tall narrow reed on its own.

* Foreign words can also be written with group-writing, but that's a separate issue.