With Linear A, there have been many attempts to decipher, and many failures. And computer attempts as well. In some ways, it would seem that Linear A would be a comparable task, although problably easier that this Indus one... which seems to consist of mostly 5 symbol objects according to Farmer.
****
Using English as an analogy for people to understand is fine. I understand the criticisms raised by Farmer, though.

In looking for patterns, you start big and get smaller. If you see the phrase:
"themanisinthehouse", 18 letters, in 5 different texts, then you have something. You still don't know that this is 6 words, or what each of these words mean, or how they fit together.

Or something like "theball" vs. "aball" vs. "baseball"... you might find 10 "theball"'s, and 50 "ball"'s, and 100 "the"'s. How do you interpret this? Tricky stuff. But yeah, you should be able to create a "potential word/phrase list", from this kind of analysis. Certainly there will be mistakes, and subjective guessing, and programming glitches and other problems. And this doesn't provide any context, either. It is still better to have this analysis then to not have it.
*****
For instance, in Linear A, many texts have been decoded(incorrectly)... and then somebody finds these words in other texts and proves what gibberish this attempt was.

If you had a "perseus-tufts" type of tool that showed frequency of each word, then perhaps there wouldn't be so many bad/failed attempts.

So I do see value and merit in this type of study/analysis, but I'm not holding my breath waiting for results.



Edited 1 time(s). Last edit at 08/04/2009 11:50AM by rich.