computers and Indology

Lars Martin Fosse lmfosse at ONLINE.NO
Fri May 8 10:05:21 UTC 1998

At 08:14 08.05.98 +1000, you wrote:
>On Thu, 7 May 1998, Anshuman Pandey wrote:
>> > 7) Saw elegant fonts for about 400 symbols used in Indus writings
>> > developed by Prof. Parpola. If used widely, many can suggest;
>> > may speed up the cracking the script, language.
>> Aid in decipherment? How?
>Obviously, one benefit is that of instant recognition of a character or
>word once it has been recognised initially by a computer.
>As far as decipherment is concerned, I believe that a statistical
>methodology may help. For example, from my research it is clear that from
>a statistical point of view, `function' words invariably occur quite
>frequently, `theme' words occur less frequently, and specific context
>words occur less frequently still. There is no reason to expect that
>(meaningful) text from the Indus civilization does not follow this general
>paradigm. A similar method may be applied to the phonology and grammar of
>the Indus script. Of course, all of this processing could be achieved by
>hand, but it is better done by a machine while the user sips a cup of tea.

Interesting point, David. However, a word of caution here: When I did
national service in the Norwegian army, I was taught to be a cryptographer.
We learnt that to make decipherment difficult for the enemy, manual codes
often had several representations of the most common letters, so that
statistics shouldn't be made too easy. (E.g.: the letter e is the most
common letter in the Norwegian language. Therefore this letter could be
represented in the code key by, say, x b and t. I think the same may obtain
with a script like the Harappan one. Several signs may represent the same
set of characters or concepts.

I think, by the way, that Parpola has used statistical techniques on the
Indus script, but I am not certain about it.

Lars Martin Lars Martin Fosse
Haugerudvn. 76, Leil. 114,
0674 Oslo

Tel: +47 22 32 12 19
Fax: +47 22 32 12 19
Email: lmfosse at
Mobile phone: 90 91 91 45

More information about the INDOLOGY mailing list