Implications of Spacing and Line Breaks for Sign Identification

Chinese characters all fit into square blocks of constant size, so the distinction between a single character and a sequence of characters is easy. The distinction between a character and a component of a character is easy. There can be no division of a single character across line breaks.

Unlike Chinese, Cuneiform signs are of variable width. This page concerns types of evidence supporting the distinction between a single sign (a sequence of sign components) and a sequence of signs.

Spacing

Under some circumstances, spacing between one sign and the next is greater than spacing between components of a single sign. This is not always the case, if the between-sign spacing is too small. And it may differ for different classes of signs. We need to accumulate full empirical evidence, and where we do not yet have that, best estimates.

Kerning

There can also be Kerning of particular sign sequences so they fit more closely together, when they ARE adjacent on the same line. So questions of closeness of spacing are in principle separate from questions whether a given glyphic image represents a single sign or a sequence of signs. Examples which may or may not fit into this category include signs whose first component looks like IGI or SAL, or sign sequences in which the first sign is IGI or SAL. For lists of these, most of which I believe are single signs, please click here.
These two different analyses are precisely what is in question here. There are surely other examples. The borderlines may not be absolutely sharp in every case between Kerning (automatic whenever two signs appear adjacent) and Ligaturing (where two signs are joined into a single glyphic image).

Divisions across Line Breaks; "Hyphenation" rules for cuneiform?

Cuneiform signs are sometimes divided into glyphic fragments across line breaks. This is in substantial measure motivated by spacing, to fill out a line before beginning the next line. It is surely also governed by the scribes’ sense of sign identities. These two can interact, so the number of divisions of felt single signs across line breaks may be more frequent with wider signs than with narrow ones, and more frequent when both fragments of a divided sign are substantial than when one of them is very narrow (one or two wedges only) and the other wide.
For lists of signs with first component U (=SHU4) or SHU2, please click here.

The frequency of such divisions might also depend on whether the resulting fragments look like existing signs, and may be arbitrary and conventional, predictable only separately for each sign. There could be conventional patterns of such line breaks analogous to what we know as "hyphenation". These could be complex in the same way that a highly sophisticated English hyphenation will include as part of it also a fixed list so that "hot-house" is divisible where "th" is not hyphenated in "brother". Unambiguous single signs might even, to permit filling out a line, be replaced by a sequence of signs or glyphs which do not look exactly like the glyphic parts of the single sign. Has there been any published study of these questions?

Historical Change

Cuneiform script covers an enormous span of time. There are repeated instances of sign combinations or sequences which fused into single signs (or into Container x Infixed structures of components), or which were later regularly ligatured. There are repeated instances of single signs which were lost and replaced by sign sequences, or reanalyzed into similar-appearing (Container-x-Infixed) structures which were not etymologically valid. So the handling of all matters discussed here can change from one era of cuneiform to another, and can be different in scribal schools of different areas.

Examples of Spacing and Component Size
Giving evidence on single signs vs. sequences of independent signs.
See also the web pages mentioned above.

The numeral sign "4" is not the same as a repetition of the sign "2", nor is "8" a repetition of "2" four times, nor a repetition of "4" twice, nor a sequence of "6" and "2". Relevant single signs are properly named and distinguished as TAB, LIMMU2, ASH4, USSU2. USSU2 is not the same sign as ASH4.TAB, etc.

U.ZAG vs. a single sign. Borger’s new sign list includes number B700a for what is usually described as "U.ZAG", This notation presumably implies a sequence of independent signs. In the Ellermeier / Studt catalog for the "ASSUR" fonts, the sequence of two signs "U.ZAG" is visually distinct from the single form which is given the number B700a. Presumably the analysis as a single sign should be represented by the reading given there, "BUZUR3", and not by "U.ZAG", but since single readings are normally given for text units which really are sequences of independent signs, we must not confuse the two. We may consider that there are almost certainly many other examples of this type, where sign lists simply do not cover them.

MUSH3 in two different versions. Borger’s new sign list assigns the number B153 to a form of MUSH3 in which the final vertical wedge is immediately adjacent to the preceding omponent which appears like SAR. There is also a form in which the spacing is greater, but the sign number is not placed adjacent to that illustration. The two are distinguished in extended Deimel / Labat numbering as L103 vs. L103v, with only the latter equated to B153.

AN+MUSH3 ligatured. The Ellermeier / Studt ASSUR fonts contain two glyphs, one in which the element "GAR" appears at full size, one in which the element "GAR" appears at reduced size, as it does in signs of the form Container-x-Infixed. The latter, because of the reduced size of a component, would be unambiguously a single sign if it were not a ligature. These are recognized by Borger as ligatures and numbered B153lig and B153lig2.

PA.GESH2 (ugula.gesh2) vs. GUR. Steinkeller (ZA 69 no.2 pp.176-187) uses character spacing to identify what is a single sign vs. a sequence of two signs, and as a result to reject a reading GUR.DA in favor of ugula-gesh2-da. The sign PA (one vertical crossing two horizontals) when followed very closely by GESH2 '60' (one vertical wedge) looks like the sign GUR. Steinkeller writes: "However, there are also clear cases where 60 is placed far apart from PA, thus leaving no doubt as to the reading. See, e.g., YOS 4,93:3. [and] RA 8 (1911) 185-187. ... the photograph (... p.187) .... the same conclusion can also be reached from the study of several examples of "GUR.DA" where, despite the crowding of the signs, the vertical wedge which follows PA is distinctly 60, since it is noticeably larger and thicker than the regular vertical wedges (DISH) occurring in the very same text." Notice that the occurrence of some cases where PA+GESH2 are written very close together does not negate the evidence from cases where there is space sufficient to establish the nature of the signs. Steinkeller also uses semantic contexts to establish the reading (ugula-gesh2-da 'in charge of 60 men').

Infixation into single signs

UZ3 x BALAG = BALAG x UZ3. These two Container-x-Infixed signs are very important examples. The reduced size of BALAG in UZ3 x BALAG shows that the combination is a single sign, not a sequence of MA2, BALAG, and KASKAL, which could not represent the sign in unformatted "plain text", and which would misrepresent it even in formatted text, because the parts at the extremes, appearing like the glyphs of MA2 and KASKAL, are not here separate signs, they are parts of a single sign, by Borger given the number B203. The infixation is actually evidence of that single-sign status, which may or may not be supplemented by other lines of evidence, about line breaks or spacing. The other form, BALAG x UZ3, infixes the combination UZ3 into BALAG, and here there are no glyphs permitting a misrepresentation of the single sign as a sequence of glyphs. These two examples are in the Ellermeier / Studt catalog of the ASSUR font on pages 37 as UZ3 x BALAG and 77 as DUB2 or BALAG x UZ3, with extended Deimel/Labat numbers L122n and L352n2. Borger assigned them no numbers. Many other signs which are like these can be recognized and named as of the type Container-x-Infixed. Please click here and see under the heading "Containers which in later cuneiform surround infixed component(s) on left and right (so the container component appears split). Some of these in older cuneiform surround infixed component(s) completely."