If I want to describe the Hittite version of the DIŊIR cuneiform glyph, I could say "a double-headed horizontal, crossing a vertical". In other words:

DIŊIR glyph

This one's fairly straightforward, and it's easy to imagine encoding it in some convenient way (2h.v or whatever).

For a glyph like SAG, though…

SAG glyph

…it's harder to think of a good encoding for that.

However, I'm sure an unambiguous representation of the shape of cuneiform glyphs has been needed in the past; it might be a long-solved problem in the field.

Is there a standard for this? Note that I'm specifically interested in the shape of the glyph, not its meaning, in order to describe the differences between the old Sumerian star-shaped DIŊIR and the later Assyrian/Hittite one—so encoding it as DIŊIR (or U+1202D) won't work.

  • 1
    Since cuneiform scripts were used to transcribe a number of languages, I think this question is probably appropriate for this Stack Exchange. Commented Aug 6, 2020 at 0:09

1 Answer 1


I haven't heard about anything like that concerning cuneiform glyphs, but there's a very interesting paper, The Xixia Writing System (Bachelor of Arts Honours Thesis), 2008, by Alan Downes (downloadable here), in which the author proposes a very smart way to encode the Tangut characters which are far more complicated than cuneiform glyphs. The author's aim is to propose an easy way to index the Tangut characters and use the indices to find characters in a dictionary. The resulting code is a string of numbers enclosed into (), [], {} to show the relative position of the graphic elements within a character:

  1. Enclose a horizontal structure with brackets [. . .]
  2. Enclose a vertical structure with braces {...}
  3. Enclose a structure inside another structure with parentheses (. . .).

A sample recursive code (that's how the author calls it) is {1,3,[40,{11,1,[14,17,14]}]}, numbers standing for different graphic elements:

enter image description here

For more detail, see section 2.4 Recursive Index for Xixia, page 13 in that paper.

With a little effort a similar system can be easily created for the cuneiform script. For example, if we assume that a vertical wedge is denoted as 1 and a horizontal wedge is 2, and a crossed by b is (a, b), then the DIŊIR cuneiform glyph

enter image description here

has the code [2,(2,1)]

And the SAG glyph

enter image description here


Naturally, cuneiform glyphs can be split into more graphic elements than just 1 and 2 I used to encode your glyphs, for example, the cross (2,1) can be treated as a separate element encoded as 3, then DIŊIR is [2,3] and SAG is [{[2,1,{2,2}],3},1]. Also, there are slanted wedges, and the ones looking like <, etc., and the relative positions of the elements within a glyph are much more numerous than the three I used, still it looks like a good point to start, to create a way to encode every possible cuneiform glyph and then to propagate it so that others use it, too.

  • 1
    Two and a half years later and a system inspired by Downes's encoding is moving toward publication. I'd like to mention this answer, since without it I never would have found his work. How would you like to be acknowledged—by username or otherwise?
    – Draconis
    Commented Jan 17, 2023 at 3:21
  • @Draconis - Just the username is quite enough. Thanks, glad to help you.
    – Yellow Sky
    Commented Jan 17, 2023 at 15:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.