Language Log

The first report I have seen concerning the language skills of the imprisoned children involved in the horror story coming out of Amstetten, this Daily Telegraph story by Nick Allen [hat tip: Matt Austin], is headlined Dungeon children speak in animal language. I suppose we should have expected it: the usual headline-writers' nonsense. Animals do not have language, and these children do not communicate like animals. The story says:

Stefan Fritzl, 18, and his brother Felix, five, learned to talk by watching a television in the dungeon where they were held with their mother Elisabeth Fritzl, 42. But their form of communication is only partly intelligible to Austrian police officers.

Police chief Leopold Etz, 50, who has met the two boys, said: "It is only half true that they can speak. They communicate with noises that are a mixture of growling and cooing."

"If they want to say something so others understand them as well they have to focus and really concentrate which seems to be extremely exhausting for them."

Being able to say anything at all to other people, however exhausting the process, makes them already quite different from any non-human animal species on earth. And semi-private speech modes used by children with siblings (e.g., identical twins with mostly or only their twin for company) are well known to developmental psycholinguists.

I find the Amstetten story almost unbearably appalling. I am viscerally affected by the story each time I think about it, which is many times each day. A point about reporting on their language seems almost too trivial to make. But perhaps it is worthwhile to say just this much: let us all try to ensure that the terrible psychological damage done to these poor children and their mother by the monster who imprisoned them is not now amplified by the promulgation of sensationalist nonsense likening them to animals.

I mentioned recently here on Language Log that the people who live in terror of splitting infinitives appear never to have looked inside the handbooks that they claim to be respecting. I came upon a remarkable instance of this the other day while looking for something else.

Punctuality Rules! is advertised as "A blog devoted to writing, grammar, good manners, and basically trying to save Civilization, one punctuation mark at a time." In this post last year the proprietor, who identifies herself as "Deb", wrote about her beloved copy of Strunk and White's The Elements of Style (of which she actually provides a photo):

Now, Strunk and White (as it's commonly called) is quite strict about some of its rules: don't end sentences with a preposition, never start one with a conjuction [sic], don't split an infinitive. All rules which common usage mostly lets slip these days. (How many non-writers do you know who even know what an infinitive is?) Its reputation is almost stodgy. A long list of rules and commands by two old, old men, you might think . . . and then you open it and start to read.

She loves her copy of S&W's third edition, of course, and she says that she reads it: "the quality of the writing is superb", and it is "possibly the very best place you can learn the rules", she thinks. I think the exact opposite is true. But never mind that. My point here is that as far as I can see, Deb hasn't actually paid attention to what The Elements of Style says.

Just in case you haven't already seen it, you should go check out WALS Online:

The data and the texts from The World Atlas of Language Structures, published as a book with CD-ROM in 2005 by Oxford University Press, are now freely available online.

WALS Online is a joint project of the Max Planck Institute for Evolutionary Anthropology and the Max Planck Digital Library . It is a separate publication, edited by Martin Haspelmath, Matthew S. Dryer, David Gil and Bernard Comrie (Munich: Max Planck Digital Library, 2008).

I was scanning the staider newspapers this morning looking for items of linguistic interest for our readership when I encountered a story in the Daily Telegraph concerning one Lord Laidlaw, a British peer who was recently discovered by a tabloid newspaper of the British Isles to have been fairly extraordinary quantities of money on flying prostitutes from Britain to a $12,000-a-night presidential suite at a Monte Carlo hotel where the girls "drank champagne and fine wines before taking part in lesbian and bondage sex acts." The puzzling part for me, given my remarkably sparse experience of such champagne-fueled sex acts, was the only linguistically relevant remark in the story:

One recent party was said to have involved a Vogue model, three prostitutes, a male gigolo and a trilingual bisexual.

What on earth, I wondered, was the relevance of the bisexual participant's ability to conduct business in three languages? I would have thought it was rather difficult to speak even one language when one's mouth is full (and I am told that at events of the sort Lord Laidlaw enjoyed, one could hardly be said to be participating fully if one didn't have one's mouth full). Are trilingual bisexuals well known to be in special demand among devotees of the lesbian/bondage scene? Are there perhaps special agencies where peers of the realm go to rent them? Or do they post advertisements in the "Personal Services" section of free newspapers? "Versatile male, trilingual (English/Spanish/German) and bisexual (French/Greek), seeks generous House of Lords member for party work in Monaco area…"

The recent gift of a staggering $100,000,000 by a single person to Harvard University — the largest gift from an alumnus in Harvard's history — has just been announced, in prose that suggests no matter how much money they may raise, the development and public relations staff at Harvard are afflicted by ancient irrational terrors:

David Rockefeller, a member of the Harvard College Class of 1936 and longtime University benefactor, has pledged $100 million to increase dramatically learning opportunities for Harvard undergraduates through international experiences and participation in the arts.

What are "dramatically learning opportunities", you might ask? We'd normally expect an adjectival rather than adverbial modifier on "learning opportunities"; is it a typo for "dramatic learning opportunities"?

No. The writer of this newsletter item (see this link) was in the grip of unreasoning fear, too petrified to consider using a normal and fully grammatical construction of Standard English that is acknowledged as grammatical in even the most conservative reference works, and never was ungrammatical at any time in the entire history of the language. Rather than use this much-dreaded construction, the writer blundered into something that actually is ungrammatical, and put an adverb in a position that actually is syntactically blocked. It truly makes me wonder whether intellectual progress is possible for a tribe as prone to panic and primitive superstition as modern educated Americans.

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?

According to Dana Milbank, "Still More Lamentations From Jeremiah", Washington Post, 8/29/2008

The Rev. Jeremiah Wright, explaining why he had waited so long before breaking his silence about his incendiary sermons, offered a paraphrase from Proverbs yesterday: "It is better to be quiet and be thought a fool than to open your mouth and remove all doubt."

Barack Obama's former pastor should have stuck with the wisdom of the prophets.

Milbank focuses on the fact that at the National Press Club yesterday,

Wright praised Louis Farrakhan, defended the view that Zionism is racism, accused the United States of terrorism, repeated his belief that the government created AIDS to extinguish racial minorities, and stood by his suggestion that "God damn America."

We'll leave those issues to the political blogs. But Rev. Wright's recent divagations have extended into linguistic territory as well. And the results are mixed at best.

Hate speech laws, in my opinion, are in general offensive and counterproductive, but the Alberta Human Rights, Citizenship and Multiculturalism Act contains a provision that really takes the cake. Section 3(1) provides:

No person shall publish, issue or display or cause to be published, issued or displayed before the public any statement, publication, notice, sign, symbol, emblem or other representation that

(a) indicates discrimination or an intention to discriminate against a person or a class of persons, or

(b) is likely to expose a person or a class of persons to hatred or contempt

because of the race, religious beliefs, colour, gender, physical disability, mental disability, age, ancestry, place of origin, marital status, source of income or family status of that person or class of persons.

You can't "expose a person … to hatred or contempt because of …source of income"‽ I'm sympathetic to the goal of discouraging the idea that, say, toilet cleaners or leather workers are inferior to other people, but aren't there some occupations that are legitimately and more-or-less universally held in contempt? Are there ways to describe a pimp, a torturer, a pirate, or a slave trader that don't expose them to hatred or contempt? I hope not.

Earlier today in a speech about his relationship with Louis Farrakhan, the leader of the Nation of Islam, at the National Press Club, Jeremiah Wright, the controversial pastor of the church that Barack Obama attends, said (transcript) (video):

Louis Farrakhan is not my enemy. He did not put me in chains. He did not put me in slavery. And he didn't make me this color.

Let me get this straight. Putting someone in chains is bad, right? Putting someone in slavery is bad, right? So "making me this color" is also bad, right? Personally, I don't think that there's anything wrong with being black. I'm dismayed that Rev. Wright does.

Some time ago Geoff Pullum wrote about the connection between the song "Happy Birthday" and linguistics, via Archibald Hill, Professor of Linguistics at the University of Texas, who inherited part of the rights to the song. The convoluted story of the copyright to the song has now been sorted out by Robert Brauneis in a paper available here. Professor Brauneis is guest-blogging on the topic today at the Volokh Conspiracy.

As promised a few days ago by Victor Mair, Amber R. Woodward's senior thesis has now been published: ("A Survey of Li Yang Crazy English", Sino-Platonic Papers No. 180, April 2008).

[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?

Last week, I posted a couple of example sentences that had given me pause:

1. I'll never forget how he must have felt. (overheard)
2. Aren’t you glad you archived instead of deleted? (over-read)

I promised I'd get back to these, so here I am.