Indus script

Seal impression showing a typical "inscription" of five "characters".

Collection of seals

The term Indus script (also Harappan script) refers to short strings of symbols associated with the Indus Valley Civilization, in use during the Mature Harappan period, between the 26th and 20th centuries BC. In spite of many attempts at decipherments and claims, it is as yet undeciphered. The underlying language is unknown, and the lack of a bilingual makes the decipherment unlikely pending significant new finds.

The first publication of a Harappan seal dates to 1873, in the form of a drawing by Alexander Cunningham. Since then, well over 4000 symbol-bearing objects have been discovered, some as far afield as Mesopotamia. In early seventies Iravatham Mahadevan published a corpus and concordance of Harrapan writing listing about 3700 seals and about 417 distinct sign in specific patterns. The average size of writing is five signs and largest text in a single line is 14 signs. He also established the direction of writing as right to left.^[1]

Some early scholars, starting with Cunningham in 1877, thought that the script was the archetype of the Brahmi script used by Ashoka. Cunningham's ideas were supported by G.R. Hunter, Iravatham Mahadevan and a minority of scholars continue to argue for the Indus script as the predecessor of the Brahmic family. However most scholars disagree, claiming instead that the Brahmi script derived from the Aramaic script.

Inscription corpus

Early Harappan

The script generally refers to that used in the mature Harappan phase, which perhaps evolved from a few signs found in early Harappa after 3500 BC.^[2] However, the early date and the interpretation given in the BBC report have been challenged by the long-term excavator of Harappa, Richard Meadow.^[3] The use of early pottery marks and incipient Indus signs was followed by the mature Harappan script.

Mature Harappan

The Harappan signs are most commonly associated with flat, rectangular stone tablets called seals, but they are also found on at least a dozen other materials.

Late Harappan

After 1900 BC, the systematic use of the symbols ended, after the final stage of the Mature Harappan civilization.

Late Indus script found on pottery at Bet Dwarka dated to 1528 BC based on thermoluminescence dating.

A few Harappan signs have been claimed to appear until as late as around 1100 BC (the beginning of the Indian Iron Age). Onshore explorations near Bet Dwarka in Gujarat revealed the presence of late Indus seals depicting a 3-headed animal, earthen vessel inscribed in what is claimed to be a late Harappan script, and a large quantity of pottery similar to Lustrous Red Ware bowl and Red Ware dishes, dish-on-stand, perforated jar and incurved bowls which are datable to the 16th century BC in Dwarka, Rangpur and Prabhas. The thermoluminescence date for the pottery in Bet Dwaraka is 1528 BC. This evidence has been used to claim that a late Harappan script was used until around 1500 BC. ^[2] Other excavations in India at Vaisali, Bihar ^[3] and Mayiladuthurai, Tamil Nadu ^[4] have been claimed to contain Indus symbols being used as late as 1100 BC.

In May 2007, the Tamil Nadu Archaeological Department found pots with arrow-head symbols during an excavation in Melaperumpallam near Poompuhar. These symbols are claimed to have a striking resemblance to seals unearthed in Mohenjodaro in the 1920s.^[4]

In 1960, B.B. Lal of the Archaeological Survey of India wrote a paper in the journal Ancient India. It carried a photograhic catalog of megalithic and chalcolithic pottery which Lal compares with the Ancient Indus script.^[4] Ancient inscriptions that are claimed to bear a striking resemblance to those found in Indus Valley sites have been found in Sanur near Tindivanam in Tamil Nadu, Musiri in Kerala and Sulur near Coimbatore.^[4]

Script characteristics

The writing system is intensely pictorial. The script is written from right to left,^[5] and sometimes follows a boustrophedonic style. Since the number of principal signs is about 400-600,^[6] midway between typical logographic and syllabic scripts, many scholars accept the script to be logo-syllabic^[7] (typically syllabic scripts have about 50-100 signs whereas logographic scripts have a very large number of principal signs). Several scholars maintain that structural analysis indicates an agglutinative language underneath the script. However, this is contradicted by the occurrence of signs supposedly representing prefixes and infixes.

Attempts at decipherment

Over the years, numerous decipherments have been proposed, but none has been accepted by the scientific community at large. The following factors are usually regarded as the biggest obstacles for a successful decipherment:

• The underlying language, if any, has not been identified.
• The average length of the inscriptions is less than five signs, the longest being one of only 17 signs (and a sealing of combined inscriptions of just 27 signs).^[8]
• No bilingual texts (like a Rosetta Stone) have been found.

The topic is popular among amateur researchers, and there have been various (mutually exclusive) decipherment claims. None of these suggestions has found academic recognition.^[9]

Dravidian hypothesis

The Russian scholar Yuri Knorozov surmised that the symbols represent a logosyllabic script and suggested, based on a computer analysis by his team, an underlying Dravidian language as the most likely linguistic substrate.^[10] Knorozov is perhaps best known for his decisive contributions towards the decipherment of the Maya script, a pre-Columbian writing system of the Mesoamerican Maya civilization. Knorozov's investigations were the first to conclusively demonstrate that the Maya script was logosyllabic in character, an interpretation now confirmed in the subsequent decades of Mayanist epigraphic research.

Knorozov's suggestion was preceded by the work of Father Henry Heras, who suggested several readings of signs based on a proto-Dravidian assumption.^[11]

The Finnish scholar Asko Parpola led a Finnish team in the 1960s-80s that vied with Knorozov's Soviet team in investigating the script using computer analysis. Based on a proto-Dravidian assumption, they proposed readings of many signs, some agreeing with the suggested readings of Heras and Knorozov (such as equating the "fish" sign with the Dravidian word for fish "min") but disagreeing on several other readings. A comprehensive description of Parpola's work until 1994 is given in his tome "Deciphering the Indus script".^[12]

The discovery in Tamil Nadu of a late Neolithic (early 2nd millennium BC, i.e. post-dating Harappan decline) stone celt adorned with Indus script markings has been considered to be significant for the Dravidian identification.^[13][14] However, their identification as Indus signs has been disputed.

Iravatham Mahadevan, who supports the Dravidian hypothesis, says, "we may hopefully find that the proto-Dravidian roots of the Harappan language and South Indian Dravidian languages are similar. This is a hypothesis [...] But I have no illusions that I will decipher the Indus script, nor do I have any regret."^[15]

Massimo Vidale ^[16] says, "I actually think that the Indus script was probably a protohistoric script, somehow conveying the sounds and words of one or more still unidentified languages."

Kalyanaraman, philologist, who supports an Indian linguistic area, notes in a monograph in July 2009 ^[17] that Indus script encodes mleccha language of smiths using pictorial motifs and glyphs as hieroglyphs read rebus. Kalyanaraman considers, in a paper presented in Madrid workshop of International Congress on the Archaeology of the Ancient Near East, 2006, two pure tin ingots found in a shipwreck in Haifa to be rosetta stones of the script, denoting the words, ranku 'tin' and datu 'mineral' ^[18].

Script vs. Ideographical Symbols

If the Indus signs are purely ideographical, they may contain no information about the underlying language spoken by their creators, i.e., they would be just logographic script, or pictograms.

In 2004, Steve Farmer, an independent scholar,^[19] computational linguist Richard Sproat ^[20] and Indologist Michael Witzel published an article^[21] asserting that the Indus Script symbols were not coupled to oral language. Witzel had earlier presented his "Para-Munda" Hypothesis, that the spoken language of the northern Indus civilization was distantly related to the Austro-Asiatic family, though not identical with Proto-Munda (Witzel 1999)^[22].

In their 2004 article, Farmer, Sproat, and Witzel present a number of arguments in support of their thesis that the Indus script is nonlinguistic, principal among them being the extreme brevity of the inscriptions, the existence of too many rare signs increasing over the 700 year period of the Mature Harappan civilization, and the lack of random-looking sign repetition typical for representations of actual spoken language (whether syllabic-based or letter-based), as seen, for example, in Egyptian cartouches.

Asko Parpola, critically reviewing the thesis in 2005, opines that these arguments "can be easily controverted".^[23] He cites the presence of a large number of rare signs in Chinese, and emphasizes that there is "little reason for sign repetition in short seal texts written in an early logo-syllabic script". He notes that Sproat, "the computer linguist of the Farmer team" agreed in a personal communication that by plain statistical tests such as the distribution of sign frequencies and plain reoccurrences, it is not possible to prove or disprove that the signs represent writing.^[24] The latter point, however, was already clearly made in the original Farmer-Sproat-Witzel paper of 2004: "Statistical regularities in sign positions show up in nearly all symbol systems, not just those that encode speech." (cf. also Sproat and Farmer, in: CSLI Studies in Computational Linguistics, Stanford University, 2006, p. 10 ^[25])

Revisiting the question in a 2007 lecture ^[26], Parpola takes on each of what he considers the 10 main arguments of Farmer and his colleagues, presenting what he considers to be counterarguments for each. He states that "even short noun phrases and incomplete sentences qualify as full writing if the script uses the rebus principle to phonetize some of its signs". He gives the example of "rebus-punning" in the protoliterate phase of Sumerian and Egyptian writing in the 32nd to 31st centuries BC, citing the Narmer Palette as a good example of what is "already a writing system even if the texts are on average shorter than the Indus texts!"^[27]

Joining the debate in a newspaper article, Iravatham Mahadevan has called the Indus "non-script" a non-issue, listing a variety of archaeological and linguistic arguments in support of his thesis.^[28]

In an article entitled "The collapse melts down: a reply to Farmer, Sproat and Witzel", Massimo Vidale opines, based a detailed evaluation of the archaeological context, why he thinks "their thesis is not acceptable."^[16]

However, Farmer, Sproat, and Witzel have responded to their critics in a paper entitled "The Collapse of the Indus-Script Thesis, Five Years Later: Massive Nonliterate Urban Civilizations of Ancient Eurasia," at a large Indus conference in Kyoto, Japan, on May 29-31, 2009; see pdf: ^[29]. They opine that Vidale's paper "like others in its class" is polemical in nature, badly distorted the long list of arguments in Farmer et al. 2004, and did not in fact discuss even a single Indus inscription—despite the fact that the 2004 paper was based on detailed analysis of the Indus corpus. The new Kyoto paper also promises that its final version will discuss all other critiques by "traditional script adherents" in detail and present a general FAQ on developments in the field in the past five years. As announced at Kyoto, this paper will be jointly published in full early in 2010 by Harvard Oriental Series (Opera Minora) and the Research Institute for Humanity and Nature (Kyoto).

A computational study conducted by a joint Indo-US team led by Rajesh P N Rao of University of Washington, consisting of Iravatham Mahadevan and others from Tata Institute of Fundamental Research and the Institute of Mathematical Sciences, was published in April 2009 in Science.^[30] They conclude that "given the prior evidence for syntactic structure in the Indus script, (their) results increase the probability that the script represents language".^[31] An explanation of the Rao et al. paper can be found here: ^[32].

Farmer, Sproat and Witzel published a quick two-page in reaction to the study, ^[33] and announced a more detailed criticism would follow.^[34] Two specialists in computational linguistics, Mark Liberman ^[35] and Fernando Pereira ^[36] have also criticized the paper.

Responding to their critics, Rao and colleagues have pointed out that the counterexamples proposed by Liberman, Sproat and others are based on the assumption that conditional entropy by itself is a sufficient condition for language, a claim not made in their paper (for the full response, see pdf: ^[37]). They define their methodology as Bayesian in nature (see ^[38]) focusing on uncovering the necessary conditions for language, which they enumerate as “Zipfian frequency distributions, syntactic structure such as the clear presence of beginners and enders, preferences of symbol clusters for particular positions within texts etc. and finally, similarities in conditional entropy”. They cite the fact that the Indus script also satisfies a growing list of these properties of natural languages as evidence in favor of the linguistic hypothesis.

At the Kyoto Indus conference of May 2009, new data presented by Farmer, Sproat, and Witzel, (and again by Sproat at a conference in Singapore in early August ^[39]), suggest that the entire methodology of so-called conditional entropy cannot even identify languages in the same language family, let alone nonlinguistic from linguistic symbols; see pdf: ^[40]

However, speaking on the issue of whether language families can be identified using conditional entropy, Rao and colleagues already state in their Science paper that “answering the question of linguistic affinity of the Indus texts requires a more sophisticated approach, such as statistically inferring an underlying grammar for the Indus texts from available data and comparing the inferred rules with those of various known language families”^[41] (more details at ^[42]).

Notes

See also

• Proto-Elamite script
• Undeciphered scripts
• Bronze Age writing

References

External links

• Indus Script (ancientscripts.com)
• Indus Script (http://www.shangrilagifts.org/hp/indus.html - Comparison of Indus Valley Harappan 哈拉帕 and Ancient Chinese Jia-Gu-wen 甲骨文 "Bone Script")
• "Discovery of a century" in Tamil Nadu ("Discovery of a century" in Tamil Nadu )
• The Indus Script (From harappa.com)
• BBC - 'Earliest writing' found
• How come we can't decipher the Indus script? (from The Straight Dope)
• Iravatham Mahadevan, Towards a scientific study of the Indus Script
• Script Image;Article
• Collection of essays about the Indus script (Steve Farmer)
• WIRED.com (WIRED.com)