Thursday July 22, 2004
Hard to Parse
The title of this post over at Transterrestrial Musings had me scratching my head. What the heck does "WHAT'S DELAY'S ANGLE?" mean—"What is the delay of the angle?"? TM is (mostly) a space blog, and in that context I was trying to make the title have something to do with the position in the sky of some celestial object (probably because I watched the ISS pass overhead earlier), and it doesn't make much sense that way, even given that it's in the telegraphic style used for headlines.
Turns out the title is actually "What's DeLay's Angle?", and it's about House Majority Leader Tom DeLay. (All caps: considered harmful.) It's interesting that simply removing the capital letters from his name—or rather, capitalizing all the lower-case letters—causes my written-word lexical lookup to find the high-frequency word delay to the exclusion of the low-frequency proper name DeLay. Another data point in favor of a Logogen-like model of lexical lookup, I suppose.
One of my professors has a favorite example of this sort of contextual priming in spoken language rather than text. He tells two brief stories whose last sentences are pronounced the same but which are actually made up of different words. However, listeners aren't confused when they hear the identical stream of phones because certain lexical items are more strongly primed by the preceding sentences, so the wrong reading doesn't even occur to them. I remember the key sentence, but not the stories, but it's easy to reconstruct plausible ones. Here are my versions, with the phonetically identical sentences in italics:
I have a friend who used to be a spy. He's been retired for a few years, but before he left he was privy to some highly classified secrets. I'm worried that some of them fall into the "I could tell you, but I'd have to kill you" category. That stuff he knows'll give him trouble.
I have a friend who's going to be interviewed on the radio tonight. As luck would have it, he woke up with a terrible cold this morning, and his sinuses are completely blocked. I think his voice is going to be hard to understand over the radio, and he'll be sniffing into the mic the whole time. That stuffy nose'll give him trouble.
It doesn't work in print, but read them aloud in your head (quit smirking, you know what I mean).
[Now playing: "Playground Love" by Air]
TrackBack URL for this entry:
Listed below are links to weblogs that reference Hard to Parse:
I am reminded of a puzzle created (as far as I know) by the late Isaac Azimov. He asserted that he could provide a simple and common written English word that could not be pronounced accurately.
Give up? The word was 'POLISH' (all-caps).
Lower case, it's the stuff you, uh, polish your shoes with.
Initally capitalized ('Polish') and it's the adjectival form of the nation of Poland.
But all-caps? You can't tell except from context, and divorced from a sentence, you can't decide how to pronounce it.
I read that when I was about 15 and it's stuck with me for 25 years. I can't pass a car-wash without wondering what "POLISH WAX" must be like...sounds dirty.
Funny, to me those are different sets of sounds. I don't know IPA, but I'll try to illustrate anyways.
stuffy --> Stu, then fee like "fee-fi-fo-fum", said w/o a pause.
stuff he --> stuff, then he, with h audible, or at least sensible
because there was a pause between the two words.
nose --> sounds like noze
knows --> sounds like no-wes, with emphasis on the no and the s
headed towards a z but still clearly not a z
I wish I knew the IPA so I could explain more clearly, or maybe had some spectrograms to show the difference.
Posted by: agm at Jul 23, 2004 11:37:24 AM
I also pronounce them differently. The one difference is easy: [stafi] v. [staf hi]. The [h] is normally pronounced in England in these function words. Also there might be a slight length difference, [stafi] v. [sta.f], but I won't swear to that.
The _nose'll_ v. _knows'll_ is puzzling, as I don't think I have any phonetic difference in _nose_ v. _knows_. Now in Scotland finally vowels are lengthened (_know_ = [no:]) and this is retained with inflection, so you get a contrast [no:z] ~ [noz]. But for me they're both just [n@Uz].
However in _nose'll_ they're subject + auxiliary, but in _knows'll_ they're unrelated words separated by a major phrase boundary, so I feel there might be some subtle length or intonation difference.