Of dreams and inspiration


Due to chemotherapy I was having strange dreams and, as I woke up, hallucinations over the weekend. A friend, Pam B, prayed that my dreams would not be random but would be helpful. Here's a short summary of a dream I had that seems to be an answer to her prayers.

You know that both the Q'uran and much of the New Testament were originally orally transmitted. As a result, mild variations in content and form occurred over time. In my dream we discussed that, by the time of the 4th Caliph, Uthman, the differences were enough that Uthman decided to destroy all copies of the Q'uran and replace them with his own official version, which is what endures today.

Christianity took a different path. We have footnotes marking differences and uncertainties, etc. In the dream, I pointed out the above facts, and then proposed an experiment (which I'd love to try sometime): let's take a document and suppose that each word has a 10% probability of an error (spelling, changes, additions). That's a ridiculously high probability, but never mind. Now, the question is: how many corrupted copies of these documents do you need until you can be 99% sure you've recovered the original uncorrupted document?

The answer is staring at us today: we have an enormous number of old scrolls, fragments, etc., and we can easily spot spelling errors and even places where someone might have deliberately beefed up a story. We can recover the "original" Bible with great reliability.


Note to my engineering buddies: yes, this does pretty much look like a conditional expectation problem with lots of data taken; the E[MSE] of the result should be less than the E[MSE] of a single measurement.

Posted: Wed - July 9, 2008 at 07:29 AM           | |

©