Language Log

More shabulengdengde poetry, from the menus of Chinese restaurateurs who put their trust in bad machine translations (as usual, click on the images for larger versions):

After all, who'd want to eat the hair blood of an impoverished person? Or a cowboy bone fried by a merely pentangular germ?

We were recently introduced to the delicacy known as "Braised Enterovirus in Clay Pot", which led to an edifying discussion about the possible role of viruses in food processing. I never would have imagined that, just a few days later, Ori Tavor would send me a photograph of a menu offering "Sautéed Wild Bacteria."

Sara Scharf sent in a link to a Chinese menu picture from the Bad Translations group on flickr:

Gov. Arnold Schwarzenegger of California appeared on ABC's "This Week" on Sunday, and he got some press attention for his stated willingness to serve as an energy and environment czar in a hypothetical Obama adminstration, even though he has endorsed his fellow Republican John McCain. In the recaps of his interview, I was struck by one sentence in particular: "I'd take his call now, and I'd take his call when he's president — any time." This use of when instead of if struck me as unfortunate, and my first thought was that it might be the result of interference from Schwarzenegger's native German, where wenn can serve as the equivalent of English if or when. To check up on my hunch, I emailed the perspicacious polyglot Chris Waigl (who has bailed me out before on German-English translation conundrums), and she replied in her typically thoughtful and nuanced manner. Her response follows below as a guest post.

That's what they call it, over at the Palais des Congrès in Paris:

Do you suppose that the Académie Française made them stick in the extra c? Anyhow, there are quite a few of these signs — I think I saw four, and probably there are more.

From Victor Mair:

The Chinese characters are CAN1TING1 餐厅 ("dining hall")

[Source of photograph: Facebook; uploaded by Samuel Osouf; taken on the Beijing-Taiyuan expressway in June, 2008. Link sent to Victor by Ori Tavor.]

Language Log readers may appreciate the following classic example of writing in technical terms from the perspective of the technician or engineer rather than from a standpoint that would seem useful to the customer or reader. I was engaged in reserving a rental car on the web, and got the date syntax wrong. Instead of having one of those little click-on-the-calendar widgets, the site required entry of the date of the reservation in a blank box using a strict syntax that it did not explain until the reservation failed and an error message was displayed. By strict syntax I mean (i) slash must be used as separator (continuous numbers will not work), (ii) day must be before month (American month-day-year syntax will not work), and (iii) the date must be four digits (two-digit year indications won't work). I'm not worried about having to know the date syntax of the culture I'm in; I can deal with that. And although the 4-digit year is a bit crazy (since time travel is impossible, the first two digits will be 2 and 0 for all reservations for the next 92 years, so they are not carrying much information), that piece of programming stupidness is not my concern here. The classic bit was the error message that popped up in red:

Please select a valid pick up date (DD/MM/YYYY) greater than today.

Have you ever said "I could do lunch any day next week greater than Tuesday"? Or "It would be helpful if you could deliver it greater than the 27th"? Or "I'm younger than my wife because my birthday is eight days greater than hers"? Of course you haven't.

Going through the latest batch of Chinglish offerings that friends sent to me last week, my eye was caught by this striking bilingual sign:

Normally, I do little more than marvel at the mistranslations and ungrammatical constructions that are characteristic of Chinglish. Seldom do I undertake deeper research into how they came about, since the causes of the bloopers and blunders are usually painfully obvious. It is only when Chinglish expressions — whether humorous or not — are hard to explain that I make a special effort to analyze them and figure out how they occurred.

But this one was different.

Alex Baumans sent in this quote from a press release about a Chinese Exhibition on solar energy:

3rd 2008 Asia Solar PV Exhibition attracted many companies come from more than 20 countries and regions, such as Germany, France, Switzerland, the United States, Hungary, with their ears, Italy, Japan, Singapore, South Korea,China Taiwan,and so on. [emphasis added]

There may be Language Log readers whose reaction to the wonderful series of posts started by Victor Mair (the ones with the photographs of signs in Chinese supermarkets saying things like "Fuck to spread the fruit"), now classified under our Lost In Translation category, was that they constitute cruel mockery and humiliation of Chinese people. Not our intent, of course; these disastrous mistranslations are a linguistic phenomenon crying out to be explained. If we merely wanted to laugh at Chinese grocery store managers we would want the signs left up uncorrected. We would never mention them; we would just pass the photos round by email, and snigger privately. Anyway, it's not about Chinese or the Chinese people, though various factors conspire to make their errors salient (there are over a billion of them, and many of them never meet a native speaker who could serve as an informal translation consultant, and the language has a very high degree of non-systematic polysemy). Long-time Language Log readers may recall that Spanish attempts at constructing English prose can be just as capable of being unintendedly hilarious.

Indeed, yesterday I was in a hotel in La Coruña, a bustling port city of northern Spain, and I found that the room service menu offered, among other things, Kettledrum of potatoes with broken eggs and creaking ham. I'll leave comments open below for people to try their hand at figuring out from Spanish/English dictionaries how this little translinguistic catastrophe could have happened. ("Creaking" was not a typo; "creaking pizza" was also on the menu.) It might be charged that there is much less excuse for such zany mangling of English in the case of a port city across the Bay of Biscay from England: La Coruña is just 90 minutes' flight time from Heathrow, and hundreds or thousands of native speakers of English come through every week. All a hotel manager would have to do would be to sit down with one guest for five minutes to learn that a kettledrum is a poor choice of cooking utensil, and that the (utterly delicious) ham of northern Spain does not creak any more than pizza does.

Um, for whatever my what?

(Click on the picture for a larger version.)

That picture was taken on a NYC subway car by Aaron Davies. Could it be that America's peanut farmers have outsourced their sign creation to China?

[Executive summary: Who knows?]

This follows up on a series of earlier posts about the comparative efficiency — in terms of text size — of different languages ("One world, how many bytes?", 8/5/2005; "Comparing communication efficiency across languages", 4/4/2008; "Mailbag: comparative communication efficiency", 4/5/2008). Hinrich Schütze wrote:

I'm not sure we have interacted since you taught your class at the 1991 linguistics institute in Santa Cruz — I fondly remember that class, which got me started in StatNLP.

I'm writing because I was intrigued by your posts on compression ratios of different languages.

As somebody else remarked, gzip can't really be used to judge the informativeness of a piece of text. I did the following simple experiment.

I read the first 10⁹ or so characters from the xml Wikipedia dump and wrote them to a file (which I called wiki). I wrote the same characters to a second file (wikispace), but inserted a space after each character. Then I compressed the two files. Here is what I got:

1012930723 wiki
2025861446 wikispace
314377664 wiki.gz
385264415 wikispace.gz
385264415/314377664 approx 1.225

The two files contain the same information, but gzip's model does not handle this type of encoding well.

In this example we know what the generating process of the data was. In the case of Chinese and English we don't. So I think that until there is a more persuasive argument we should stick with the null hypothesis: the two texts of a Chinese-English bitext are equally informative, but the processes transforming the information into text are different in that the output of one can be more efficiently compressed by gzip than the other. I don't see how we can conclude anything about deep cultural differences.

Note that a word-based language model also would produce very different numbers for the two files.

Does this make sense or is there a flaw in this argument?

Henry Kenrick wrote:

I read your posts on Google translations with interest, as just a few days ago I came across a similar example when translating Russian to English.

Below I have copied the text and the Google translation. It's clear there's still some way to go, but some phrases aren't bad. The most interesting artefact was the translation of свердловских (genitive case plural of the adjective for Sverdlovsk region) into Heilongjiang.