Check out this post from Brad Feld:
Other than reading a bunch of blog posts about the Google / China thing, I hadn’t paid much attention to it. Early today, I got an email that fits the proverbial “picture is worth a thousand words” cliche.
I wasn’t able to try this using Chinese characters (er – I don’t know Chinese and I’m tired and lazy after running a marathon yesterday) so this might be a spoof. But – if not – enough said.
It's no spoof. I can type both Chinese and English on my computer and here were my results typing in Chinese on both the English and Chinese Google portals:
This means that Google is censoring not only the Chinese site, but also searches conducted on the English site using Chinese characters. Very interesting indeed.
Also, check out the discrepancies when the word "Mormon" is searched for:
Quite remarkable, and a bit unsettling. Nevertheless, I think the access to information provided by even a censored Chinese Internet will help to assist in accelerating the tipping point as far as China's long march towards democracy. As they say in Chinese: 慢慢来。

Very interesting!! I had no idea google was different from country to county.
Posted by: kristenq | 2006.01.31 at 15:32
Hey I read an article yesterday that relates to your post, but I can say that I heard it from you first! Very interesting, though; it's kind of a quandry these companies find themselves in: promote free speech or make more money. Interesting to see what they're choosing.
http://www.msnbc.msn.com/id/11178913/site/newsweek/
Posted by: Claire | 2006.02.10 at 08:17
"This means that Google is censoring not only the Chinese site, but also searches conducted on the English site using Chinese characters. Very interesting indeed."
Not necessarily true. Google presumably indexes text around images (including title= and alt= tags) in order to do the search.
It would not be unreasonable to assume that many of the site employing Chinese characters are in China and are therefore already censored.
In other words, the source may be censored, rather then the presentation by Google.
You have a similar issue with your second test.
Brad's test, on the other hand, is valid. The two engines have exactly the same source data, so any difference in presentation is the result of a decision by Google.
Posted by: Allan | 2006.03.06 at 13:32