Friday, April 13, 2012

Comparison of Indus script with Luwian and Chinese oracle bone scripts

Block with Luwian hieroglyphs (from an old Turkish postcard).

I own a single resource with texts in Hieroglyphic Luwian (also known as Hieroglyphic Hittite) (Çambel 1999).  None of the texts in this volume come from seals – the main source of Indus inscriptions – although Luwian hieroglyphs do appear on seals.  This makes a direct comparison of statistical measures between the two scripts invalid.  In addition, this group of texts comes from a single location and is not necessarily representative of the script in general.  So any statistical measures cannot be generalized with any confidence.

Nevertheless, it may be of some value to compare sign frequencies between this corpus of Luwian (all on stone monuments) and that of Indus script (all on small objects).  The reason is that Indus script is sometimes characterized as a logo-syllabic writing system (e.g., Parpola 1994: 85; Wells 2010: 92-93).  It is worthwhile examining how the Indus script compares with a known logo-syllabic system, of which Luwian is one example.  Luwian is, of course, incompletely understood, but far better understood than Indus script.

I made a count of the number of times each Luwian sign occurred in the corpus of this single volume, with the exception of the highly fragmented portions.  With a combined total of 2,511 symbols, 133 potential signs occurred between 1 and 232 times each.  I say potential signs because I found it impossible to identify perhaps two dozen with absolute certainty due to broken or abraded surfaces.  Frequencies are summarized in the following table (data on Indus script is from Possehl 2002: 133).

Table 1.  Frequencies of Luwian hieroglyphs and Indus script.

Frequency of sign
No. of signs with this
frequency in
Luwian hieroglyphs
No. of signs with this
frequency in
Indus script
83 (62%)
219 (45.4%)
19 (14%)
100 (20.7%)
14 (10.5%)
86 (17.8%)
13 (9.8%)
46 (9.5%)
100 or more
3 (2.3%)
31 (6.4%)

To simplify the information in the table, we can combine the frequencies of all signs occurring fewer than 100 times, for which any statistical measures are likely to be invalid.  For this sample of Luwian hieroglyphs, that would yield 129 signs making up 96.3% of the total.  For the corpus of Indus signs, this yields 451 signs making up 93.4% of the total.  In both cases, only a very small proportion of signs occurs more than 100 times (3 Luwian signs making up 2.3% and 31 Indus signs making up 6.4%).  This broad similarity in frequency is one piece of evidence pointing toward the Indus script being a logo-syllabic script (along with the total number of signs).

In addition to these very general similarities in frequencies, both Luwian and Indus symbols demonstrate regularities in position.  For example, in the Indus inscriptions, POT is the most frequent sign by far, the only one that occurs over a thousand times.  When it appears, it is most often in final position.  In the Luwian inscriptions, the symbol for the syllable mi occurs more often than any other, multiple times in virtually every inscription and almost always at the top of each register (or row) of glyphs.  The precise total for mi should be approximately half the number I counted since “IC” occurs twice to make up this syllable.  Thus, instead of appearing 232 times, this syllable occurs 116 times.  But since I could not always see IC duplicated – i.e., I did not always find an even number of IC symbols – my counted total is probably inaccurate anyway. 

One of the main reasons that researchers consider the Indus script likely to be logo-syllabic writing system is the number of total signs.  Alphabets have fewer than 100 symbols; a syllabary may have 100 or so depending on the type of syllables represented; logographic writing such as Chinese has thousands of symbols.  Thus, the mixed type termed logo-syllabic writing is the best match with the Indus script, with symbols numbering in the hundreds.  In such a writing system, the syllabic (or other phonetic) symbols typically occur more frequently than the logographs.  This is clearly demonstrated in a previous post where I discussed Linear B (Mycenaean writing).  However, there may be some overlap in frequencies, with a small number of syllabic signs occurring rarely on the one hand, and a small number of specific logographs appearing especially frequently. 

The function of the most frequent logographs differs between writing systems and, indeed, between genres of inscriptions.  Logographs that represent a few commodities are the most common of this category in Linear B.  This is because the Linear B tablets are exclusively economic in nature, the texts being essentially names of commodities and numerals measuring them, with some proper names of people and places involved with transactions.  On Egyptian scarabs, glyphs that apparently represent wishes (such as ankh “life,” nefer “beauty, goodness” and djed “stability”) appear to be the most frequent logographs, while certain phonetic symbols are the highest in absolute frequency, especially the bread loaf t, zigzag n, and chick w (I have not yet completed an exact count of glyphs from a corpus of this type).  The prevalence of the ankh and similar signs, in this case, is probably due to magico-religious reasons.  Inscriptions that can be read mostly consist of personal names and titles or descriptions of professions.  Numerals are rare.  The Luwian texts I examined comprise a third type, neither economic nor onomastic.  Names of rulers do occur with some frequency, but the texts do not simply give names and titles, but also include descriptions of military and religious actions.  Numerals are rare in this corpus as on Egyptian scarabs.

Now let us examine the frequencies of Chinese characters as they appear in a small sample of oracle bone texts (Keightley 1985).  Since Chinese is the logographic system most commonly referenced in discussions of writing, it should prove illuminating.  Of course, modern Chinese could just as easily be described as “logo-syllabic” given the latitude with which that term is used (such as for Egyptian hieroglyphs, which do not actually represent syllables but only consonants).  That is, both semantic and phonetic information appear in most modern characters.  Often an element of the meaning resides in the “radical,” most often on the left side of a character, while a phonetic clue appears on the right in the form of a simple character whose pronunciation is indicated (or something close). 
A Chinese oracle bone inscripion: guei-mao [date] crack-making, [the] king divined:
[in the next 10-day] week, [there will be] no disaster.  In [the] 5th month, jia-chen
[date], yong ritual [performed to ancestor] Da Jia (Keightley 1985: 115, rephrased).

However, this type of combination character is not nearly as frequent in the early stage of writing on oracle bones (Keightley 1985: 68).  According to one study, 23% of oracle bone characters are pictographs or simple pictures, 32% are compound ideographs (such as the combination of the woman and child characters to form the word “good”), 27% combine radical and phonetic, 11% are “loans,” and 2% are abstract symbols.  The so-called loans are other types of characters (pictographs or compound ideographs) that have a second function representing an unrelated but homophonous word.  That is, where modern Chinese uses the radical + phonetic to distinguish characters representing homophones, oracle bone script often uses the same simple character for two different words that happen to sound alike.  Context distinguishes the two functions.  Of an estimated total script “vocabulary” of 5,000 or so characters, about half are understood by modern scholars (1985: 59). 

In the small, unrepresentative sample of texts included in this work, frequencies of characters are as shown in Table 2.

Table 2.  Frequencies of Chinese oracle bone characters.

Frequency of sign
or character
No. of signs with this
frequency in Chinese
1 occurrence
49 (38.6%)
2 occurrences
30 (23.6%)
3 occurrences
18 (14.2%)
4 occurrences
9 (7.1%)
5 occurrences
5 (3.9%)
6 occurrences
2 (1.6%)
7 occurrences
4 (3.1%)
8 occurrences
1 (0.8%)
9 occurrences
1 (0.8%)
10 or more
8 (6.3%)

Since no character appears more than 17 times in this small sample, the total numbers of occurrences are much smaller than for the Luwian or Indus samples.  Again, though, one can see that the majority of characters occur infrequently, with many singletons.  In these texts, the content is considerably more varied than in the texts of Egyptian scarabs, Near Eastern seals, or Linear B tablets.  Some of the Chinese texts are economic: the inscriptions may enumerate the turtle plastrons or animal scapulae brought in and prepared for divination.  But this is a relatively minor aspect.  Numerals are quite frequent – at least smaller ones – because the scribes write down the number of cracks made in the surface in the process of divination.  That is, each crack typically has a number.  Proper nouns are also common, both in the inscriptions describing preparation of oracle bones and in the main texts.  The main texts describe the charge – who performed the divination, when, and for what purpose – as well as often indicating the result (typically verification of the result that was foreseen).  Topics of divination include sacrifices, military campaigns, hunting expeditions, journeys, weather, agricultural activities, illness, childbirth, the source of various troubles, meanings of dreams, settlement building, tribute payments, and divine assistance or approval (Keightley 1985: 33-35).

The most frequent characters include the character referring to making cracks (bu 14 occurrences), or divining (zheng 17 occurrences), and those used in dates (e.g., jia 13 occurrences).  Dates often contain no numerals, instead combining the 12 celestial stems and 10 earthly (or horary) branches.  Some dates are more frequent than others because the people evidently performed divinations at the beginning of each 10-day “week,” to determine whether the coming period would be auspicious.  Many of the singletons in this sample are the names of the diviners, though the character wang, “king,” is the most common diviner referenced (14 occurrences).  In a more complete sample, particular diviners would appear frequently, each being involved with many divinations.

It is highly improbable that the Indus seals and tablets bear content comparable to the oracle bones.  But it is certainly possible that the inscriptions are magical and/or religious in nature.  In fact, magical and religious elements also appear within the restricted context of Near Eastern cylinder seals.  And as noted above, Egyptian scarabs often have magical or religious symbols alongside or instead of regular hieroglyphs.  Some scholars maintain that Indus signs are not writing, but magico-religious symbols (Farmer 2004).  While this may not be true of all the symbols, parallels from other Bronze Age cultures make this a distinct possibility.
Tablet H-316A with inscription (right to left): HAIRY HUNCHBACK // POT / COMB.

Specifically, Farmer suggests that the POT is a schematic representation of a tree, the so-called ritual stand in front of the unicorn on so many seals also representing a tree, pipal or fig (2004: 16).  He further states that the high frequency of POT in inscriptions indicates it was the central symbol of Harappan society (2004: 22).  He also sees the POT HATTED BEARER as the result of syncretic merging of a sacred tree and a likely divinity of water.  
Tablet H-308A with inscription (right to left): HAIRY HUNCHBACK // POT / COMB.

The common sequence of HAIRY HUNCHBACK / POT / COMB, in this author’s view, is a representation of a human sacrificial victim sitting before this same sacred tree (2004: 40).  This sign sequence is mostly found on miniature steatite seals, inscribed on both sides.  The reverse typically contains CUP alongside an apparent numeral (SINGLE POST, TWO POSTS, THREE POSTS, less frequently FOUR POSTS, and in a couple of rare cases, SIX POSTS).  This group of inscriptions (CUP + “NUMERAL”) may indicate an economic record of sacrifices (2004: 29).  The miniature inscribed objects might have served as something like vouchers for seasonal offerings.  For example, each “citizen” may have been required to present a measured amount of a particular commodity at a harvest festival.  In return, the “citizen” would receive one of these small seals showing that he had given his share.  These interpretations are speculative, of course, but provide some interesting possibilities to consider.


Çambel, H. 1999. Corpus of Hieroglyphic Luwian Inscriptions. Vol. II Karatepe-Aslantaş. The Inscriptions: Facsimile Edition.  Berlin and New York: Walter de Gruyter.

Farmer, Steve. 2004. “Mythological function sof Indus inscriptions: Eight conclusions arising from the nonlinguistic model of Indus symbols,” Sixth Harvard Indology Roundtable, 8-10 May 2004.  Available online at .

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

Parpola, A. 1994 and 2009. Deciphering the Indus Script. Cambridge, MA: Cambridge University Press.

Possehl, G.L. 2002. The Indus Civilization: A Contemporary Perspective. Lanham, MD: AltaMira Press.
Wells, B.K. 2010. Epigraphic Approaches to Indus Writing. Oxford and Oakville: Oxbow Books.

No comments:

Post a Comment