signs do not occur haphazardly, but follow certain rules. Some signs are limited to the end of a sequence, even when such a sequence occurs in the middle of an inscription, while other signs are usually found in the beginning of a sequence, some others are never found there, and so on.
It must be admitted that it is very difficult to construct even parts of the Indus grammar on this basis.18) Nevertheless, the positional se- quences can be profitably exploited to analyse the Indus texts syntacti- cally, to define the textual junctures, and to classify the signs into phonetically or semantically similar groups. Such analyses can be car- ried out with automated methods.19) Data accumulated in this way will certainly be useful in decipherment once a decisive breakthrough has been achieved — in other words when the language has been identified and some signs have been read phonetically in a convincing manner. But analyses of this kind are themselves unlikely to provide that break- through.
Conclusion Perishable archaeological material being involved, and taking into con- sideration the very limited amount of surviving monumental art, nega- tive evidence is not sufficient to prove wrong the hypothesis that the Harappans wrote on palm leaves or on cloth. Richard Sproat, the computer linguist of the Farmer team, admits that by statistical means it is not possible to distinguish a logo-syllabic script of the Mesopotamian type from non-linguistic symbol systems.20) The question of whether the Indus signs are script or not, ultimately depends on whether one can
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Cf. Parpola 1994: 86–97.
See Parpola 1994: 97–101.
My colleague Kimmo Koskenniemi, who is Professor of Computer Linguistics at
the University of Helsinki and has participated in research on the Indus script, asked by e-mail Dr. Richard Sproat the following question: “It appears that we agree that plain statistical tests such as the distribution of sign frequencies and plain reoccurrences can (a) neither prove that the signs represent writing, (b) nor prove that the signs do not represent writing. Falsifying being equally impossible as proving. But, do I interpret you correctly?” In an e-mail sent to Kimmo Koskenniemi on Wednesday 27 April 2005, Dr. Sproat answered to this question with one word: “Yes.”
〔 42 〕