Language Log

Over on the American Dialect Society mailing list, we've returned to the topic of illeism, the use of third-person expressions to refer to oneself (treated on Language Log last year), in particular, illeism in speech to or from young children, as in:

[mother to child] Mommy has to go now.

[from child named Kim] Can Kim have ice cream?

As Larry Horn noted, such illeism seems to be a way of coping with the difficulty that young language-learners have with first- and second-person pronouns, which famously are "shifters", with reference that shifts from context to context. Ordinary proper names (like Kim) and kin-terms used as proper names (like Mommy) have a reference that doesn't depend on context the way the reference of first- and second-person pronouns does. Horn recollected:

I recall a Sesame Street episode when our own children were at the appropriate tender age that attempted to "teach", or at least play on, such issues involving the proper use of "I"/"you", "my"/"your", etc.

Carrying the Sesame Street theme in a different direction, I added that Elizabeth Daingerfield Zwicky reported to me some time ago that toddlers' use of their names for self-reference comes up repeatedly on parenting discussion sites, usually in the context of blaming Elmo for it. Elmo refers to himself as "Elmo", and parents reason that their kids picked up their illeism from Elmo. Where else could it have come from?

There's a suppressed premise in that reasoning, and when it's exposed we can see that this way of looking at things is pretty much backwards. And that it ties in with other widespread beliefs about what happens in child language acquisition.

In my post "Uptalk anxiety", 9/7/2008, I tried to comfort an American parent who was worried about a daughter's use of rising pitch accents on statements. As part of the recommended cognitive therapy, I observed that there are regional varieties of English, known as "Urban North British", in which rising pitch accents on statements are more common than not.

But Bob Ladd, who ought to know, commented that "It's important not to confuse the rises in Belfast, Glasgow, etc. with uptalk. They're phonetically and functionally very different."

I responded that "There's no question at all that they're *functionally* different. In terms of sound, though, I think that the issue is less clear." I asked Bob whether he thinks that the pitch contours are systematically different, and in particular whether he could "tell the difference, on short phrases whose F0 and amplitude contour was used to modulate a non-speech oscillator, in the mode of the example e.g. here?"

Bob answered: "If you make some examples, we can do the experiment, but the short answer is that I think I could as long as there is a "tail", i.e. unstressed syllables after the nucleus (last main stress) – in the sound example I posted, there are two postnuclear syllables, -mond and mine. The main phonetic difference between classic North American / Antipodean uptalk and the "Urban North British" statement rises is that the latter rise at the nuclear syllable and then level out or trail off, whereas in uptalk the pitch just keeps on rising."

We're not ready to do the experiment yet, but I can offer some evidence-based suggestions about how it's likely to turn out.

For our first lolcat on New Language Log, here's a phonological one (passed on by Lise Menn):

That is, insert cat between cushun and cushun. (This one is along the lines of a linguistic lolcat suggested by Laurel MacKenzie in an earlier Language Log posting on lolcats.)

[Andrew Carnie writes to tell me that there's a LOLPhonology group on Facebook. Very entertaining. There are 82 photos there at the moment, including this one.]

Mark Liberman's recent posting on the Cobot elicited some comments about talking in slogans. And now along comes Zippy, in catch-phrase dialogue with Griffy:

And from a while back, Dingburgers conversing in George W. Bush quotes:

Because I seldom listen to the news on the radio or watch it on TV, most of what I know about happenings in the world is gleaned from print media:  books, magazines, papers, and so forth.  Consequently, I occasionally adopt a "reading pronunciation" for the name of a person or place that is at variance with the actual spoken pronunciation of the name.  Such is the case with a place name that is currently prominent in the news:  Ossetia.  In my mind, and even when speaking to others, I have been blithely and happily saying Os-sĕ-ti-a.  After all, I thought to myself, the people who live there are Os-set(e)s, and their language is Os-sĕt-ic or, so I thought, Os-sĕ-ti-an.

From the LiveJournal of lord_whimsy, a social report, with amateur dialectology. The setting:

Last week, the Missus and I attended Interview's relaunch party, held on the top floor of the partially-completed Standard New York, a retro-brutalist sort of structure which towers on tall stilts over the Meatpacking District.

and now the observation:

we followed the gaggle of impossibly tall, thin models and sundry gay boys through the construction site to the elevators, whose walls were still bare plywood. We literally came up to the waist of some of these striking extraterrestrials. I calculated the lisp per capita ratio in the elevator to be an astounding 3:1, which had a similar aural effect as a swarm of summer locusts. My ears literally hurt from the insectoid crispness of the diction being volleyed overhead. I've long suspected that there's a third dimension to regional dialects: not just geographical, but vertical. Someone should do a linguistic field study of New York elevators that lead to media offices: A much overlooked micro-dialect is thriving in elevator shafts all over Midtown Manhattan. 

Some of this — in particular, the hyperbolic "my ears literally hurt from the insectoid crispness" — is just routine disdain for the gay voice (similar to the intense disdain many people freely express about the speech of young women, various social and geographical dialects, and so on). But there's a small chance that lord_whimsy was on to something about the vertical dimension in this particular case.

Every so often we post here about baffling headlines — baffling to readers who don't have the real-world knowledge needed to interpret them. Most recently, Geoff Pullum posted about

Detective attacks jailed canoe wife who lied to sons

(which he used as a springboard for a discussion of noun-noun compounds like canoe wife). Today's delight (from Bruce Webster, who came across it on Dave Barry's blog) is

All Blacks lock rubbishes Wallabies poor form line

Without some context, this is impenetrable — unless you something about rugby (especially in the southern hempishere) and some British slang. It's significant that the headline comes from a New Zealand rugby site. And that All Blacks and Wallabies are capitalized; they refer to rugby teams (the All Blacks are the national team of New Zealand, and the Wallabies of Australia). So lock is not a verb, but a noun referring to a rugby position and the person playing it. And so on.

The slang is the verb rubbish, which means 'criticize severely and reject as worthless' (NOAD2) in British English and varieties influenced by British English (including at least Australian, New Zealand, South African, and Indian English).

(Hat tip to Bruce Webster.)

A recent PartiallyClips strip illustrates a technique that we might call "associative dialogue":

(Click on the image for a larger version.)

The tailor's algorithm, obviously, is to use the customer's questions as queries into a database of quotations.

This technique has been used to great advantage by generations of chatbots. The one whose design I know most about is the "extended chat" mode of Cobot, described in C. Isbell, M. Kearns, D. Kormann, S. Singh, P. Stone, "Cobot in LambdaMOO: A Social Statistics Agent", AAAI 2000:

Any utterance directed towards Cobot that is not recognized as a request for social statistics becomes a candidate for the following process. Words in the incoming utterance are matched to words appearing in sentences in the documents, assigning to each sentence a weight based on the number of matching words. […] Cobot randomly chooses a sentence to utter according to the distribution defined by the weights.

The Presidential campaign of the past few days provides us with not one but two examples of false claims about candidates' statements. The first is the now widespread claim that Republican Vice Presidential candidate Sarah Palin believes that the Iraq War is divinely ordained because she said that:

our national leaders are sending them [the troops] out on a task that is from God

The title attribute for the most recent xkcd strip has the value "Supercollider? I 'ardly know 'er", with apostrophes in place of the two initial h letters. This is a cultural mistake, a rare thing from Randall Monroe, who is usually pitch perfect.

Commenting on our posts about Business Week's credulous coverage of the SpinSpotter software release, Trent defended his former profession:

When we are knowledgeable in a particular field, we notice errors made by outsiders. […] Because the typical journalist at a newspaper is a generalist, and because he or she may have to write 10 column inches within 20 minutes about something unfamiliar, there are bound to be errors — some substantive, some not so. […] Demanding that a newspaper hire experts in all fields is just … unreasonable. Demanding that a journalist spend hours researching the material — well, you can get it perfectly accurate, or you can get it fast. Newspapers are in the business of being fast. Journals are in the business of being rigorous.

I've heard versions of this excuse many times over the years. And with respect, I believe that it's irrelevant to the case under discussion, and largely nonsense in general.

Our recent adventures with the vaporware/demoware SpinSpotter (here and here), which purports to detect passages of untrustworthy spin, reminded me of last month's software delight, apostrophree, which, it was said, automatically and silently

corrects common errors of spelling, punctuation, grammar, and usage in blogs and especially comments and discussion forms.

(this from a Typical Programmer interview with apostrophree's founder John Scogan).

What is spin? According to the OED's 1993 additions,

2. g. fig. A bias or slant on information, intended to create a favourable impression when it is presented to the public …

What is SpinSpotter? According to Claire Cain Miller in the NYT ("Start-Up Attacks Media Bias, One Phrase at a Time", 9/8/2008), it's a Web tool that "scans news stories for signs of spin".

The Spinoculars find spin in three ways, said Mr. Herman. First, it uses an algorithm to seek out phrases that violate six transgressions that the company’s journalism advisory board came up with based on the Society of Professional Journalists’ Code of Ethics. They are personal voice, passive voice, a biased source, disregarded context, selective disclosure and lack of balance. […]

SpinSpotter’s algorithm also uses a database of common phrases that are used when spinning a story. Finally, readers can flag instances of spin. Other SpinSpotter users can see these flags, and the reported phrases will enter the spin database.

The guy being quoted is  "SpinSpotter founder and chief creative officer, Todd Herman". Other stories about SpinSpotter — and there are quite a few of them — give a similar picture.

But here's another definition, offered by me in comments on a weblog post yesterday:

This might be an unusual type of demoware …, one that is released for general use in the hope that enough people will submit their proposed spin-spots to give the company enough free training data to actually develop some of the technology that they pretended to have in the first place.