Monday, March 3, 2014

Ligatures in the Indus Script: Part I

Seal M-1921 with FORK & EF TOPPED DUBYA (center), a ligature.

In the hieroglyphic writing of ancient Egypt, on occasion two independent glyphs are written together, i.e., as a ligature.  In Gardiner’s list of Middle Egyptian glyphs, for example, D59 is a combination of the lower leg with foot (D58) and the forearm with hand (D36).  Separately, these represent the sounds transliterated b (D58) and (D36).  Together, the ligature represents these two sounds, ‘b, as in the word for “horn” (Gardiner 1976: 458).   

Egyptian glyph D59, a ligature of D36 (forearm) and D58 (leg).

 Following this example from a known script, we could hypothesize that Indus symbols form ligatures in the same way.  We may designate this type of ligature as “simple addition,” in which symbol A, with meaning x, combines with symbol B, with meaning y, to give the ligatured symbol AB, with meaning x + y.  The ligature AB would then serve the same function in the script as the sequence A + B.  More succinctly, the hypothesis is that ligature AB = sign A + sign B.  Korvink’s note on FORK & EF TOPPED DUBYA (M-1921) implies just such a hypothesis: “As a side note, POT and FORK-TOPPED POT, may be expressed in the abbreviated sign FORK & EF TOPPED DUBYA” (2008: 30).

Seal M-1908 with inscription: SINGLE POST (?) / FORK TOPPED POT / POT (over rhino).
To test whether this is the case, we will examine the inscriptions, seeking to compare sequences of A + B with occurrences of the ligature AB.  For the POT-HATTED BEARER (designated more briefly as POT BEARER), this means that this ligatured symbol should appear in the same contexts as sequences of POT + BEARER.  As it happens, the sequence POT + BEARER appears in eight inscriptions in the concordance of Koskenniemi and Parpola (henceforth referred to as KP; 1982: 23-24).  In five of these, the combination occurs at the end of the inscription (KP 5061, K-8, M-209, M-733, L-140).  In one case, this sign pair forms the whole inscription (L-184).  In the other two instances, the combination is followed by one to four additional signs (KP 5035 and the duplicates M-494 and M-495A).  This evidence tends to support the hypothesis.

Tablet M-495A with inscription (right to left): CIRCLED FORK / CRAB / HAIRY HUNCHBACK / POT / BEARER / TRIPLE CUPS / FORK (?).

Note further that there are no occurrences of either *DOUBLE POT or *DOUBLE BEARER (the asterisk indicates that it is not attested).  Thus, if POT BEARER = POT + BEARER, there should be no occurrences of *POT + POT BEARER or *BEARER + POT BEARER.  In fact, POT BEARER appears at the end of 51 inscriptions.  In the other 35 inscriptions, this sign is followed by one to four additional signs (KP 1982: 24-25).  Among the inscriptions that include POT BEARER, there are, as expected, no sequences of *POT + POT BEARER or *BEARER + POT BEARER.  This evidence is consistent with the hypothesis that ligature AB (here, POT BEARER) is equivalent to the sequence A + B (POT + BEARER). (Gulf seals bearing Indus signs may show DOUBLE POT: BM1883_1116_1 ends with this doubling; cf. M-1792 with three non-contiguous POTS along with the singleton DUCK IN POND.)

Gulf seal BM 1883.1116.1 with inscription: PLOW / CRAB / SINGLE QUOTE / DOUBLE POTS.

But other apparent ligatures do not provide the same evidence.  MAN HOLDING QUOTE appears to be a ligature of MAN + SINGLE QUOTE.  MAN appears most often at the end of an inscription (60 times out of 102 inscriptions or 59% in KP 1982: 29-30).  This fact leads Korvink to analyze MAN as a terminal sign (2008: 28-29).  In contrast, SINGLE QUOTE functions as the “constant” and final part of a prefix (2008: 22).  As such, it seldom occurs in final position in inscriptions (11 instances in KP 1982: 84; about 6% of the total of 191 occurrences 2011:166).  Since the medial section of an inscription usually separates prefix and terminal, sequences of MAN + SINGLE QUOTE are not expected.


When these two signs are ligatured, forming MAN HOLDING QUOTE, the result is neither a prefix constant nor a terminal.  Of 16 inscriptions featuring this ligature, the sign appears in final position only once (KP 2751; although Kd-9 includes it as one of only two signs, so that it might be in final position, depending on the direction of reading).  It appears as part of the variable portion of the prefix in 7 instances (M-71, L-2, KP 9551, Laursen 26/KP 9907, M-1052, L-1, and duplicates M-519 through M-521 and M-1470 through M-1472).  In the other 8 inscriptions, it appears as part of the medial segment (C-40; Kd-9; H-660; duplicates M-547-549 and -1555-1560; duplicates M-2053 and -599; M-425; M-403; and M-428).  

Seal M-1918 with partial inscription: FORK / MAN / SINGLE QUOTE / ... / TRIPLE BRICK.

Prefixes, medial elements, and terminals appear in distinct contexts, so it is not surprising that the same kind of evidence as in the first example is lacking (POT BEARER as POT + BEARER, all three terminals).  Indeed, MAN and SINGLE QUOTE rarely occur in sequence: in M-1918, MAN immediately precedes SINGLE QUOTE, while in KP 9701 SINGLE QUOTE immediately precedes MAN (inscriptions in the KP 9000 series are Gulf seals with Indus symbols, a group that shows a number of differences from true Indus inscriptions, as noted in Laursen 2010: 96-134).  This is insufficient evidence to support the hypothesis of “simple addition” for MAN HOLDING QUOTE.
Similarly, MAN HOLDING POST (37 occurrences; e.g. H-72 and M-282) is medial, occurring 6 times in initial position, twice as the variable portion of the prefix (H-241 and M-1686), never in final position.  MAN does not appear alongside SINGLE POST, so there is no data for comparison.  It may be significant that all the modifications of the basic MAN are medial signs except for the BEARER types (BEARER, CHEVRON-HATTED BEARER, POT-HATTED BEARER).  This complicates the test of the “simple addition” hypothesis.  

Seal H-72 with inscription: MAN HOLDING POST / DOUBLE GRIDS.

But the change from terminal to medial sign suggests that modification itself may have a significance other than simply the addition of the meaning of the ligatured elements.  In other words, virtually any modification of MAN being modified changes it from a terminal to a medial sign.  Does this happen to other terminal signs?  This does not seem to be the case for the BEARER.  The changes in the top of this sign from a straight vertical line to a chevron (CHEVRON HATTED BEARER) or to POT (POT HATTED BEARER) do not change these signs’ “syntactic” function.  They remain terminals.  So it is not a general rule that ligatures must be medial signs.  
Tablet H-967 with inscription (right to left): ODD STACKED / 2 POSTS / POT / COMB.

But modifications to POT do change its function.  The addition of a SINGLE QUOTE, BI-QUOTES, THREE QUOTES, or FOUR QUOTES changes the terminal POT to a medial sign (POTTED ONE, POTTED TWO, etc.).  Korvink is somewhat vague on this point.  He notes that inscriptions may contain more than one terminal sign, in the following sequence: (1) FORK-TOPPED POT, (2) COMB, (3) SPEAR, (4) POT, (5) MAN, (6) CHEVRON-HATTED BEARER, (7) POT-HATTED BEARER, (8) BEARER, (9) COMB (and this final COMB may be duplicated) (2008: 29).  (Note that every sign in this sequence is optional and no inscription contains all of these.)  To determine the relative placement of MAN and the various types of BEARER in this sequence, Korvink uses what he calls “variants” of MAN, including MAN BETWEEN POSTS, MAN WITH EAR, and MAN HOLDING POST (2008: 31).  He explains that he uses the term “variant” here not to suggest that these signs all mean the same thing but “to infer a similar syntactical function of the sign” (2008: 30).   
Tablet H-1934 with inscription (right to left): QUADRUPED / POTTED 1 / HAIRY HUNCHBACK / POT / MAN.
While it is clear why MAN should be considered a terminal sign, it is not so clear with these proposed “variants.”  MAN BETWEEN POSTS always immediately precedes a terminal sign, either one of the BEARERS or the COMB (although in KP 1141 the sign or signs following MAN BETWEEN POSTS are illegible).  Thus, it might be analyzed as a terminal itself.  But such an analysis would make a fair number of inscriptions anomalous in containing nothing but a terminal of two signs (M-716, M-915, M-1816, M-1839; M-197; M-830, and the duplicates H-543 and 544).  MAN WITH EAR occurs three times (out of a total of 6 inscriptions) in the prefix, which would be unusual for a terminal sign (though not without parallel).  Further, while MAN HOLDING POST often immediately precedes a terminal (17 times), it also occurs immediately after the prefix (8 times), as part of the prefix (3 times), and often in combination with DOUBLE GRIDS (14 times), which is not a terminal.  Thus, it appears to be a medial element rather than a terminal.
In contrast to this subsuming of MAN’s “variants” in his discussion of terminal signs, Korvink clearly differentiates POT, as terminal, from its apparent variants, POTTED ONE, POTTED TWO, POTTED THREE (2008: 35, fig. 18).  

POT (terminal)
Positional Distribution of POT, POTTED-ONE, -TWO, and -THREE (Korvink 2008: 35).

Here, it is clear that POT most often occurs in final position.  When it appears in medial position, it is usually because another terminal follows it (e.g., H-967: ODD STACKED EIGHT / TWO POSTS / POT / COMB).  In other cases, POT occurs medially as the end of the first unit of information in a long inscription, followed by another unit of information (e.g., H-58: FISH UNDER CHEVRON / DOT IN FISH / FLANGE-TOPPED POT / POT / PLOW / CIRCLED FORK / POT BEARER).  The insertion of one or more “quotes” changes the POT to a medial sign.  As such, the ligatured sign may occur in initial position as part of the prefix or as a medial sign where there is no prefix.  In rare cases, it ends an inscription when there is no terminal.  Inscriptions containing both POT and POTTED ONE/POTTED TWO reveal the difference in these signs most clearly (e.g., H-1934 with POTTED 1; M-304 with POTTED 2; for the latter, compare M-258 with the sequence POT / TWO POSTS).  Thus, Korvink concludes: “Using pictographic similarity to infer a similarity in meaning is a highly speculative approach” (2008: 36).

Seal M-304 with inscription: MAN HOLDING STOOL (?) / FAT STOOL / POTTED 2 / FISH / POT // MAN.

In other words, the hypothesis that ligature AB = sign A + sign B is not confirmed as far as modifications of MAN and POT are concerned.  Now, although I began this post by referring to an Egyptian ligature, this type of “simple addition” does not apply to all Egyptian ligatures either.  In Egyptian, some glyphs are phonetic symbols.  They represent one, two, three, or occasionally more consonants.  The ligature of two phonetic glyphs does follow the “simple addition” pattern.  In addition to the example cited at the beginning of this post, there is S 30, a combination of S 29, a folded cloth, and I 9, a horned viper.  Phonetically, s + f = sf.  But not all glyphs are phonetic.  A great many serve as ideographs or determinatives, or both.  For example, the most common anthropomorph is A1, a man seated with his left leg under him and the right knee up, both arms bent at the elbow.  As an ideograph, it may occur alone or with the vertical stroke (which indicates that a glyph is an ideograph), meaning “man” (Gardiner 1976: 442).  Also ideographically, it may represent the first person pronoun, “I, me.”  In either case, it may be accompanied by phonetic glyphs, s for “man,” and i, wi, ink, or kwi as “I, me.”  When the seated man holds a basket steady on his head, this is A9, determinative in 3tp “load,” f3i “carry,” or k3t “work” (1976: 443).  When the seated man holds an oar in his right hand, this is the determinative in sķdw “sail.”  So it goes.  Modifications of the basic man are almost always determinatives and/or ideographs.  An exception, an upright man with legs spread, is A27, phonetic in “by” (1976: 445).

Chinese characters from oracle bones: bao yi (upper left); bao (upper right); bao bing (lower left); bao ding (lower right) (after Keightley 1985: 185).

Chinese characters from oracle bones: two variants of wang "king" (after Keightley 1985: 216).
 In Egyptian writing, the determinative comes at the end of a noun or verb.  In this way determinatives are similar to the Indus terminals.  However, virtually all phonetic glyphs in Egyptian can also act as determinatives or ideographs (or both).  So the two classes of glyph are not mutually exclusive.  In fact, all originally were ideographs and their phonetic function was due to the rebus principle (the same principle by which English speakers use the numeral 2 as an abbreviation for “to” in texting). 
Egyptian glyph A1, the basic seated man (after Gardiner 1976: 544).

Glyphs A9 (left) and A10 (right), which resemble ligatures of A1 with a basket (W10) and oar (P10).
Similarly, in the earliest Chinese writing on oracle bones, there are ligatures, combinations of two or more elements in a single character.  For example, there is a squared off “C” shape similar to the square bracket “]” with longer horizontals, bao.  The insertion of the character yi, similar to the Indus ESS, inside the bracket bao creates the name Bao Yi (Keightley 1985: 185).  If, instead, a small square is inserted in the bracket, this is the name Bao Ding.  One estimate is that 27% of the oracle bone characters are made up of two components, a radical and a phonetic (1985: 68, f. 49).  

In some cases, modifications of these early Chinese characters are not meaningful, representing simply variant writings.  For example, the character wang “king” appears sometimes as a stick figure upon a horizontal base, at other times as the same stick figure on the same base but with an additional horizontal across the top, and in still other cases as three stacked horizontals joined by a single vertical (1985: 216).  Other modifications are indeed meaningful.  The variants of wang appear to be modifications of the simple stick figure, which represents da “big, great” (1985: 119).  The first variant of wang resembles da “big” plus yi “one” as the base.  However, in this case the meanings of the two characters apparently ligatured in wang (“big” + “one”) do not provide its true meaning (“king”).  The addition of a horizontal across the top – which seems to place the stick figure between the strokes of erh “two” – does not derive the ligature’s meaning from its apparent components (“big” + “two”) either.  Thus, as Korvink noted concerning Indus symbols such as POT and POTTED ONE, etc., assuming a semantic relationship between symbols based only on graphic similarity is speculative and may be misleading.  

Thus far, it seems that one cannot automatically assume that a complex symbol is a ligature simply because it looks like a combination of two (or more) simpler symbols.  Nor can one assume that an apparent ligature serves as an abbreviation for its component parts.  In other words, it is only sometimes true that ligature AB = sign A + sign B in a sequence.  I will return to this theme in my next post.


Gardiner, Sir A. 1976 (1927). Egyptian Grammar: Being an Introduction to the Study of Hieroglyphs. Oxford: Ashmolean Museum and Griffith Institute.

Joshi, J.P. and A. Parpola. 1987. Corpus of Indus Seals and Inscriptions. 1. Collections in India. Helsinki: Suomalainen Tiedeakatemia.

Keightley, D.N. 1985 (1978). Sources of Shang History: The Oracle-Bone Inscriptions of Bronze Age China. Berkeley: University of California.

Korvink, M.P. 2008. The Indus Script: A Positional Statistical Approach. Gilund Press (Amazon).

Koskenniemi, K. and Parpola, A. 1982. A Concordance to the Texts in the Indus Script. Helsinki: Department of Asian and African Studies, University of Helsinki.

Parpola, A., B.M. Pande, and P. Koskikallio. 2010. Corpus of Indus Seals and Inscriptions. 3. New Material, Untraced Objects, and Collections Outside India and Pakistan. Helsinki: Suomalainen Tiedeakatemia.

Shah, S.G.M. and A. Parpola. 1991. Corpus of Indus Seals and Inscriptions. 2. Collections in Pakistan. Helsinki: Suomalainen Tiedeakatemia.

Wells, B.K. 2011. Epigraphic Approaches to Indus Writing. Oxford & Oakville: Oxbow Books.

No comments:

Post a Comment