surprisal for dogs

today's Frazz comic:

i don't know if anybody has actually done research on dogs' abilities to learn frequency-based patterns (although we had cottontop tamarins not so long ago). and unlike the grammarpattern-sensitive monkeys, Mario didn't even wait to confirm the probability-based prediction, he just went for it.

AND??? and i hate you, congress.

why, why do i do things like try to figure out what has been going on in the Senate regarding the scazillion-dollar stimulus plan? it only a) gets my blood pressure up and b) confirms that our elected representatives are morons, or at least have stared at legislative doubletalk for so long that their judgements about English have been seriously compromised. exhibit 1: Senate Amendment 309, introduced by Thomas Coburn (R-OK)

At the appropriate place, insert the following:

SEC. __. LIMIT ON FUNDS.

None of the amounts appropriated or otherwise made available by this Act may be used for any casino or other gambling establishment, aquarium, zoo, golf course, swimming pool, stadium, community park, museum, theater, art center, and highway beautification project.

take another look. "…museum, theater, art center and highway beautification project"??? that is one hell of a project. in fact, i'm pretty sure you won't be finding any such mega-conglomerate initiative anywhere in the original bill. and they passed this amendment. what a waste of time. idiots.

of course, having someone proofread the damn thing and change and to or would have saved the nonsense that this will create in conference committee, in the courts if the bill is signed into law with this amendment in place, &c. &c.

Pearls Before Swine takes on English-only

sums it up pretty well, i should say (click for big):

flavors of English on Google

i was just looking through the site statistics for this here blog. one of the most interesting and useful bits of information that statcounter provides me are the search terms that people use. i would say that 99% of these searches are done on Google — we really have drunk the pagerank kool-aid. a lot of searches are pretty lengthy and specific (e.g. "kobe bryant interview in italian" or "who is the girl in the benny lava video?"). one recent search stuck out to me, though. somebody searched for just the word "whomever", and wound up at my previous post "The Office on whomever". i thought that was pretty remarkable. i clicked through on the link that statcounter provided me and saw that the search was made on google.co.uk, and that descriptively adequate was on the front page of results, at position number 6.

then, for whatever reason, i decided to re-run the search using google.com. my post was nowhere to be found on the first page. the results were entirely different. descriptively adequate finally showed up at #14 on the list of results. what's going on? certainly google hasn't written different versions of pagerank to deal with different localizations of English? as far as cataloguing search results goes, the fact that a bunch of Americans in California wrote the algorithm shouldn't adversely affect Brits and the like.

i couldn't stop there. i ran the search on all of the English Google localizations that i could think of, and got even more different results. i've also noted the number of total results that Google estimates, which also (oddly) vary by localization.

localization	#	total hits
google.com	14	7,480,000
google.co.uk	6	8,200,000
google.ca	7	8,180,000
google.com.au	10	8,190,000
google.com.nz	7	8,460,000

as i was compiling this table i remembered that Google mucks with your search results if you're signed in (which i of course had to be in order to access blogger, without which i couldn't be writing this post). i signed out, and on google.com the DA link rose to #4. i guess i should just be happy i'm on the front page on all of these searches. but there are still lingering, bizarre questions.

why does Google report different numbers of hits for different localizations?
no clue. (comments are open!)

what is causing the rank fluctuations even when i'm not logged in?
some clue. on all of the non-US localizations there is a feature "search pages from [country name]". perhaps i've got fewer australian sites linking to my blog, so my rank is slightly lower in australia than in the US or great britain.

why the hell is Google biasing my custom algorithm against my own damn blog?!
i mean throw me a bone here, guys.

and the baffler...
why do i get this on google.ca?
i mean, you're kidding, right? i'm sure that the frequency of whatever is much higher than that of whomever, but 8 million hits on a word that's in the dictionary should be enough data for google to not question my intent. and why only canadians, eh? this, of course, isn't the first time that i've seen weird spelling suggestions on Google. so perhaps they really do think they know something about English varieties that i don't?

Malaysian government fails to ban feature reconstruction

please, don't judge me about the inspiration for this post. the short story is "sometimes you just get bored, and who knows where you could end up on Wikipedia!" tonight it was crappy pop song articles. thence comes this quote from the "Controversy" section of the article for this summer's top hit, I Kissed a Girl.

In Malaysian radio stations, the song has been retitled 'I Kissed...' with the words 'a girl' silenced throughout the chorus in the song.

never mind the odd choice of preposition (as a native English speaker i've never heard a song in a radio station; on works fine). the fact of the matter is that this censorship is about as effective as bleeping the -hole in asshole. if you take the phrase "i kissed a girl" and eliminate "a girl", then in isolation it becomes completely open-ended. it could be "i kissed a man" or "i kissed my mother" or "i kissed a frog". too bad there are more lyrics in the song's refrain!

i kissed a girl / and i liked it / the taste of her cherry chapstick

oops! there's a gendered pronoun hanging out there, eight words later. and it needs an antecedent. and the only preceding nominals are i and it. i can't be the antecedent, because then she would have said my, and it is decidedly neuter. so it can only be...gasp! she didn't! chances are nobody's getting the wool pulled over their eyes either; Wikipedia also says that increasing numbers of Malaysians are identifying English as a first language. they can put the pieces of this not-so-tricky linguistic puzzle back together as quickly as i did. censorship falls flat again.

i think i've gotten more linguistic enjoyment out of the song than by listening to it. there's one other bit of the chorus that intrigued me. it's the other pronoun in those lines, namely it. i'm sure that the intended antecedent is "[the fact that] i kissed a girl", but i can't help but get an ambiguous interpretation where it could be topicalized and actually refer to "the taste..." is this a weird judgement? comments are always open here.