Google Blogoscoped

Forum

Overview of Sites Google Agreed to Censor in China  (View post)

Bálint [PersonRank 1]

Thursday, January 26, 2006
18 years ago

This just got me thinking: can Chinese people access google.com by setting their browsers to english? (I know it does not matter much, but anyway...)

Philipp Lenssen [PersonRank 10]

18 years ago #

I would like to know too. Can someone from China test this?

Seth Finkelstein [PersonRank 10]

18 years ago #

Ah, I think I finally figured out why I couldn't reproduce the results at first.

The search [falun gong] seems to lock the right-most button, but NOT [Falun Gong] (case sensitive?)

Can you confirm?

Philipp Lenssen [PersonRank 10]

18 years ago #

Ah, you're right! (That could be the reason why I was heavily confused during writing yesterday's article... I couldn't believe my eyes as the results behavior seemed to switch in front of me while I was checking them!)

Both [Falun Gong] and [falun gong] are censored though, according to the page count and the notice at the bottom. But [Falun Gong] returns more results when checked with the left radio button.

Wuming [PersonRank 1]

18 years ago #

Yes, we can access Google.com for now.

GrinGod [PersonRank 0]

18 years ago #

Why don't you repeat this with google.fr and search for references to nazis? It seems that google is also filtering search results for France. China isn't the only coutry to be receiving special treatment.

Philipp Lenssen [PersonRank 10]

18 years ago #

> Why don't you repeat this with
> google.fr and search for references to nazis?

GrinGod, please see here:
http://blogoscoped.com/archive/2005-01-15-n50.html

Philipp Lenssen [PersonRank 10]

18 years ago #

Another censored search is for [mao tse-tung] – missing around 1 million pages, including Google.com's top result of a Geocities page.

When keeping in mind Google often screws up big numbers when providing their page count estimate, the following are no clear estimates to me but might point at something going on:

- All 19 million pages for [site:geocities.com] are missing.
- On Google's Blogger, roughly half the pages are showing, but over 19 million are missing here too.
- A search for [site:com] in Google.cn ("all websites" option) results in 10,420,000,000 missing pages when compared to Google.com, with no notice of censorship on Google.cn.
- A search for [the] results in 20,420,000,000 results on Google.com, but only 4,870,000,000 on Google.cn (with no censorship notice).

Jason [PersonRank 0]

18 years ago #

The last time I was in China (3/2004) when I tried to hit google.com I was bounced back to google.cn. The computer I was on was an English version of XP. I believe the bouncing is IP based.

Also, there are a few things to note which might bias these results:

1) Are you sure that google.com and google.cn use the same index? Even within google.com you can get different results as queries are routed to different shards that are in different states of indexing. In short, a side-by-side comparison of query results might not be so much a sign of overt censorship as simply indexing anomalies.

2) I have no proof of this, but I believe the "your query returned x pages" thing is an estimate.

2a) Again, the index google.cn uses may simply be smaller.

For what it's worth I was able to easily search for "tibet freedom" and the first page returned discussed "Chinese genocide" – which I'm guessing would be considered inflammatory by the Chinese central government.

Finally, I'm not a fan of restricting the free flow of information, but I have to ask myself which is better: Google genuflecting to a dictatorship and enabling censorship, or Google not doing business in China at all?

Philipp Lenssen [PersonRank 10]

18 years ago #

> Are you sure that google.com and google.cn
> use the same index?

For the page count I did in the comment above, I will take that into account. So, like I said, these might be estimates we can ignore (we don't know). However, on the comparison chart page, you can see that all result pages except the one for "playboy" indicate (at the bottom) that the results are censored. And even for Playboy.com, I find it very unlikely that this site just happens to be missing in the index...

Michael Leddy [PersonRank 0]

18 years ago #

You've done an excellent job of documenting differences. I found some interesting results yesterday morning searching for "dalai lama." See here:

http://mleddy.blogspot.com/2006/01/google-and-censorship.html

Brian M. [PersonRank 10]

18 years ago #

The Chinese government is already blocking these web sites. Google is now blocking the meta data of these web sites. Why is it necessary? They are being strongarmed by the government. Block access to the metadata and deny that these websites even exist. If you do not we will instead deny you access to millions of potential searchers.

Google has taken the top-down view that hey, if we don't pretend that some websites don't exist, we can't even tell these searchers that /any/ websites exist.

But wait a second – Google Corporate says that Google's mission is to "organize the world's information and make it universally accessible and useful."

Google has organized this set of information. However, they are not making it universally accessible because they are pretending some of it does not exist. The layman's reading of Google's mission, which is the intended reading, leads to the thought that Google is violating it's core ethos. That Google is willing to accept two wrong's /can/ make a right.

Of the comparison filtered queries Philipp showed us, the one that strikes me as most awful is "human rights." When you pretend that some human rights don't exist, you can count on the fact that you have done something evil.

Apple [PersonRank 1]

18 years ago #

For every search I've done, I have found much less cn pages.

for example, try searching "Apple"

1. The chinese search gives about half the number of results than those from google.ca.
2. Notice how the results are different.

I'm not saying that China doesn't censor anything, but it it obvious that your comparison is flawed.

Philipp Lenssen [PersonRank 10]

18 years ago #

Sorry Apple, what's your point? That Google.cn has less pages than Google.com, Google.ca et al? Of course they do, that's the problem! They censor millions of pages, and many of them contain "harmless" words (like "apple") as well.

Do a site search for Geocities, which is at this moment completely banned in Google.cn, and you will see there are already 238,000 censored pages there which happen to include the word "apple" as well.
Search for
   site:geocities.com apple
or
   site:geocities.com
whichever you prefer. You will see zero results on Google.cn, but instead this notice (approximate translation):

"According to local laws and policies some search results are not showing"

You can even do a search for ["Welcome to Apple II Oasis"] to see that disclaimer... because the top result would have been a Geocities page!

anonymous [PersonRank 0]

18 years ago #

Where is your screenshot for "xenu"?

RMX [PersonRank 0]

18 years ago #

The Google Image Search for Tiananmen is even more obvious. Please add screenshots of it.

   * http://images.google.com/images?q=tiananmen [google.com]
   * http://images.google.cn/images?q=tiananmen [google.cn]

Scary.

Anonymous [PersonRank 0]

18 years ago #

Your "f*ck" screenshot is just stupid. Whoever took it is searching for "f*ck" in Chinese-only pages, and FYI, Chinese has ITS OWN ALPHABET and "f*ck" is NOT A CHINESE WORD.

Put up a screenshot with the results for "xenu". Hmmmm? Where are they?

Rocky [PersonRank 0]

18 years ago #

I'm distressed by Google's stance here. I understand why they're doing it but I'm thinking of another approach which seems to meet China's govn't's needs while reducing Google's complicity in the govn't propaganda. Either 1. they've missed this approach, 2. they have a good reason not to take this approach, or 3. they have a bad reason not to take this approach (money, market share, unnecessarily appeasing the Chinese govn't).

Proposed approach:
For keyword searches for which the govn't wants google to restrict access to metadata (i.e. search terms such as "Tibet", "Falun Gong", etc where search results promoting one side of the argument don't show up), Google should show _no results_. That is, Google could politely provide a notice that local laws restrict displaying many search results for that keyword and suggest looking elsewhere, without displaying the govn't manufactured results.

The moral corrosion of "Google Certifying" a profoundly non-democratic presentation of key political, economic, and environmental issues (notwithstanding Google's boilerplate disclaimer) outweighs the negligible advantage of the searcher being able to access those propaganda sites.

I ask Google, "what is the ostensible usefulness of providing links to sites demonizing Falun Gong?" Chinese websurfers can get that propaganda from every other domestic media outlet. If China isn't demanding that Google _actively_ distribute propaganda--and there have been no published indications that this is the case--why is Google doing it?

Tintin in the Tibet [PersonRank 0]

18 years ago #

Anonymous says : < Your "f*ck" screenshot is just stupid. Whoever took it is searching for "f*ck" in Chinese-only pages, and FYI, Chinese has ITS OWN ALPHABET and "f*ck" is NOT A CHINESE WORD.
Put up a screenshot with the results for "xenu". Hmmmm? Where are they? >

Response: I don't agree with you, Anonymous! Many Chinese guys know what "F*ck" means, how to say it in Mandarin and how to write it in English and Cantonese. And if some don't, they'll know soon as they'll travel in the non-censorship world before the next century, I guess...

Tintin in the Tibet

Anonymous coward [PersonRank 0]

18 years ago #

The link you provided about (former) Google principles has dissappeared.

Dave [PersonRank 0]

18 years ago #

I think Google have done the right thing. Better to have a Google in china then not one at all. Chinese users are going to better off than they have been, and once Google is in the strongly market, I'm sure it will use some of it's strength to try and break down barriers.

Philipp Lenssen [PersonRank 10]

18 years ago #

> Your "f*ck" screenshot is just stupid. Whoever took it
> is searching for "f*ck" in Chinese-only pages

Anonymous, read the introduction note – I made it *bold* so it can't be missed easily:
http://blogoscoped.com/censored/

Sterling [PersonRank 0]

18 years ago #

>For keyword searches for which the govn't wants google to restrict access to >metadata (i.e. search terms such as "Tibet", "Falun Gong", etc where search >results promoting one side of the argument don't show up), Google should >show _no results_.

I don't think the block the Chinese government used is keyword based but site based. Multiple different keywords could get to a same site, some of those keyword might even be benign.

For example of Falun Gong.
Benign searches like organization, government, groups, etc might contain Falun Gong, and they'll still have to block it resulting in some pages being blocked. So I doubt using blanket keyword blocking might work. And of course, there's always a chance that a few sites might slip through...

Winkyboy [PersonRank 0]

18 years ago #

I think it would have been better to have no Google in China rather than a filtered one; this is the first I've heard about Google breaking its principles, and it has me now thinking about what other search engines would be better.

Kubo [PersonRank 0]

18 years ago #

Please note Google's press release on this matter: http://googleblog.blogspot.com/2006/01/google-in-china.html

While you are right for raising the issue, it seems to me that gradual opening of a communist country is better than doing nothing about it – even if this "opening" means temporary compliance with that country's antidemocratic regulations.

Cheers,
Kubo

There's many problems with these examples [PersonRank 0]

18 years ago #

Some of these examples are good, others are ludicrous. Using Tiananmen Square in Google Images as an example is ridiculous. Tiananmen Square is a place that's a hugely popular tourist attraction. It's like Google Images searching for New York City and seeing nothing but pictures of Osama bin Laden and the World Trade Center collapsing.

If you search for tiananmen square in Chinese characters instead of in Pinyin, you'll see identical results. Google's image search will generally convert Pinyin into Hanzi. If you search for &#22825;&#23433;&#38376; on images.google.cn and images.google.com you'll see the exact same images. Google will convert all Pinyin into Simplified Chinese. Try typing in any Chinese word as Pinyin, and watch it get converted into Simplified Chinese.

Searching for words in English between the two sites just is not a good comparison. Of course there will be differences in ranking and results returned from a site whose searches are optimized for words and phrases that would be different for a native English speaker.

The really telling results came from when the A:B comparison was done using particular sites as an example. However searching for something like Tiananmen or English words such as "f<b></b>uck" just isn't a good or very valid example.

I'm not defending Google's actions at all, but some of these examples use poor research techniques.

Philipp Lenssen [PersonRank 10]

18 years ago #

> Searching for words in English between the
> two sites just is not a good comparison. Of
> course there will be differences in ranking
> and results returned from a site whose searches
> are optimized for words and phrases that would
> be different for a native English speaker.

No, this was not about ranking – the tanks were missing form *all result pages* using that search. This is not simply country-specific optimization behavior, either: Google themselves at the bottom of that image search result admitted, with their disclaimer, that the results are censored.

George Hedfors [PersonRank 0]

18 years ago #

What I just realized working on a project in Germany is that Google seem to filter some content here too. Upon disabeling the "SafeSearch" feature of images.google.de, the following message was displayed proofing my statement.

"Note: You have selected the least restrictive filter for your results. Images not compliant with German law will still be filtered. You may, however, see some images that you consider inappropriate for minors. If minors are using your browser we suggest enabling strict filtering."

What else does google filter that are not communicated to the users?

george.hedfors[put at-character here]gmail.com

Philipp Lenssen [PersonRank 10]

18 years ago #

George, Google says that everything added to the censorship blacklist before March 2005 does not reveal itself as such. A more complete list of what was or is censored in Germany can be found here:
http://blogoscoped.com/archive/2005-01-15-n50.html

Travis [PersonRank 0]

18 years ago #

But is Google just censoring China?

http://www.travolisblog.com/is-google-censoring-in-the-united-states-33.html

"This video is not playable in your country."

Philipp Lenssen [PersonRank 10]

18 years ago #

Someone at Nathan's blog hinted that this might be because you can choose the countries you want to support when you upload user content.

Olivier [PersonRank 1]

18 years ago #

An image search of [tibetan flag] on both google.com and google.cn is quite obvious..
However, a search using Chinese characters [西藏国旗] (translated with Google's "language tools") on google.cn shows some pictures of the Tibetan flag. I just wonder if these are forgotten pictures that will sooner or later be censored, or if this is deliberate..

Philipp Lenssen [PersonRank 10]

18 years ago #

What differences do you see for [tibetan flag] Oliver? Though I see the "censored" disclaimer, I don't see much obvious difference in the first few images...

Olivier [PersonRank 1]

18 years ago #

Oh.. I just figured out i had a '&cr=countryCN' flag in the research addresses I found in my history... I must have clicked on the 'Chinese websites' radio button (the last), but I'm pretty sure I didn't : I wonder if this has been automatically selected during the numerous searches I did..

Philipp Lenssen [PersonRank 10]

18 years ago #

Olivier, yes, it is automatically selected for some searches... and then, you can't put it back to "all sites" for that search. Very confusing and misleading.

Forum home

Advertisement

 
Blog  |  Forum     more >> Archive | Feed | Google's blogs | About
Advertisement

 

This site unofficially covers Google™ and more with some rights reserved. Join our forum!