13 June 2005

ironic

One of the ironic things about this post is that by publishing it, I may be exposing my own misunderstanding of the term irony. Indeed, it has sat unpublished for six weeks partly because of this fear (and it was a propos of nothing until today). My interest is rekindled by this story on CBC.ca by Andre Mayer, about the 10-year anniversary of Alanis Morissette's CD Jagged Little Pill and her decision to release a new acoustic version of the entire album to mark this event. I have no objection to Mayer's claim that the new release, in essence, is yet another public exercise in narcissism on Ms. Morissette's part. However, Mayer includes a sideways jab that the lyrics in the song "Ironic" are actually devoid of irony:

For all its bluster, Jagged Little Pill’s best song was also its meekest; despite its basic misreading of irony.

I have seen this claim regarding Morissette's grasp of irony made before in stronger terms, in Jorge Cham's PhD. Comics:

1995. Alanis Morissette releases the hit single, "Ironic". Ironically, none of the song lyrics actually describe an ironic situation.

OK - looks like we're getting into lexical semantics here. I think what both Mayer and Cham are referring to is a distinction between literary or dramatic irony and a lay sense of the word. Ironic, in the lay sense, refers to a more general notion of unfortunate timing. I sometimes (ironically) refer to the lay usage as "Alanis Morissette irony". But I actually think that some of Ms. Morissette's vignettes are ironic in the literary sense.

Here's the relevant part of the Merriam-Webster definition of irony, which Mayer's article also links to:

3 a (1) : incongruity between the actual result of a sequence of events and the normal or expected result (2) : an event or result marked by such incongruity b : incongruity between a situation developed in a drama and the accompanying words or actions that is understood by the audience but not by the characters in the play -- called also dramatic irony, tragic irony.

And here's what Oxford has to say:

2. fig. A condition of affairs or events of a character opposite to what was, or might naturally be, expected; a contradictory outcome of events as if in mockery of the promise and fitness of things. (In F. ironie du sort.)

These definitions seem fairly consistent. Now, a logician will tell you that all you need to prove a universally quantified statement to be false is to provide a single counterexample. "None of the song lyrics actually describe an ironic situation" is a universally quantified statement, so I only need to show that at least one situation in the song is indeed ironic. For the sake of completeness, I will evaluate all the situations below.

An old man turned ninety-eight.
He won the lottery and died the next day.


Expected result: old man gets to enjoy riches of lottery.
Actual result: old man dies before he can do so, despite his longevity.
Evaluation: ironic. This would be more ironic if the knowledge of owning a winning ticket is what caused the old man's death, but even without it, there's an incongruity here.

It's a black fly in your Chardonnay.

Expected result: The chardonnay is for you to drink.
Actual result: The chardonnay is for the fly to drink.
Evaluation: ironic. Presumably the black fly would also find this ironic, since it would rather be consuming you.

It's a death row pardon two minutes too late.

Expected result: the pardon will come in time to save the inmate.
Actual result: two minutes is long enough for the injection to take permanent effect.
Evaluation: somewhat ironic. More like plain bad timing.

It's like rain on your wedding day.

Expected result: nice weather on your wedding day.
Actual result: it rained on your wedding day.
Evaluation: ironic in the drought-prone summers of the Ontario of Alanis's (and my) salad days, where wedding days between May and September are very reliably sunny.

It's a free ride when you've already paid.

Expected result: had you known, you could have got a free ride.
Actual result: you didn't know, so you bought a non-refundable ticket.
Evaluation: somewhat ironic.

It's the good advice that you just didn't take.

Expected result: knowing how good the advice was, you take it.
Actual result: your narcissism leads you to ignore the advice.
Evaluation: This seems to be more indicative of stubborness than irony. It might be more ironic if the same advice had been given and followed in the reciprocal direction in the past.

Mr. Play-It-Safe was afraid to fly
He packed his suitcase and kissed his kids goodbye
He waited his whole damn life to take that flight
And as the plane crashed down he thought
"Well isn't this nice..."
And isn't it ironic...dontcha think


Expected result: Mr. Play-it-Safe arrives at his destination using the "safest form of travel".
Actual result: Mr Play-it-Safe dies the horrible violent death he always feared.
Evaluation: ironic. I think Mr Play-it-safe has a reasonable expectation of a safe arrival.

A traffic jam when you're already late.
Evaluation: This seems to be an example of poor planning rather than irony, especially if the traffic jam is a recurring one.

A no-smoking sign on your cigarette break.
Evaluation: Maybe this sign is a "sign".

It's like ten thousand spoons when all you need is a knife.

Evaluation: Assuming an even distribution of knives, forks, and spoons in the world, the chances that a sample of cutlery with 10,000 pieces would not contain a knife are too small for my calculator to render, but are something like (2/3)^10000. In this case there is a reasonable expectation that the set of cutlery would contain at least one knife. Unless, of course, it's a shipment of spoons. Otherwise, this is a bit ironic.

It's meeting the man of my dreams
And then meeting his beautiful wife


Evaluation: Maybe she shouldn't have written that other song about Mr. Duplicity, or spent all that time with him.

Overall, it seems like the situations do become less ironic over the course of the song, but, as I've argued, some of them illustrate an incongruity between an expected outcome and an actual one. I think what might have happened is that some sort of meme to the effect that "'Ironic' isn't ironic" got started and continues to get repeated, without deep analysis of the claim.

Again, I'm perfectly open to poking fun at Morissette's lyrics, but I think the Ironic thing is off the mark. Will I get Jagged Little Pill Unplugged? Probably not. I only own the original because a friend passed off an extra copy to me after her mother, ironically, gave it to her for Christmas two years in a row.

11 June 2005

coding for lexical neighbors

A few weeks ago on phonoloblog, I posted some thoughts on lexical neighborhoods for languages with contrastive segment length. The issue is that calculating the number of neighbors for a given item presumably would net different results based on how you conceive a geminate: is it a pair of segments, or is it a single segment?

The clarifying example I used is from Italian: if geminates are considered two segments for neighborhood purposes, then occo and osso are not neighbors, but inna [6/25/05 - I meant inno] and anno are. But if geminates are considered one segment, occo and osso are neighbors.

The basic intuition for the two-segment procedure is like this: your comparison procedure takes two words and decides if they're neighbors. A separate procedure keeps a tally of neighbors for each word in your list. If the words are not identical, and have a difference in length of no more than one segment, then the procedure knows they may be neighbors.

So then you need some contingencies. If the target word is shorter by one segment, then the program needs to watch for a segment addition in the compared item. If the target is longer by one, the procedure needs to watch for a deletion in the compared item . If they are the same length, there may be a substitutiuon. So the procedure checks equivalently positioned segments, looking for mismatches. If it finds more than one, the words are not neighbors.

In a basic measure of computational complexity, this requires 50 lines of code in Java, 9 ifs, 13 pairs of curly brackets, and three while clauses. I have placed the procedure into a larger program that asks for a source data file to read and writes to a different output file.

But adhering to the concept of a geminate as a single segment becomes much more computationally cumbersome. The identical-length comparison still just needs to check for substitutions, but must be updated to tolerate an instance of one geminate subbing for another (e.g., occo otto). It also needs to avoid being tricked by forms like a hypothetical pair occo ooco, which are orthographic neighbors, but not phonological ones (because they differ in the length of two segments).

When words differ in length, you need more contingencies of situations to watch for... a difference in overall length could foretell an addition, or a length contrast. Imagine your pair is (pak, paak). When you reach the third-segment comparison, there is a mismatch. The program needs to know that the a in the second item is the second half of a geminate, and that really the comparison should now be between the 3rd segment of pak and the 4th segment of paak - so tally up a score of one mismatch so far. Simple enough, but then what if you're comparing pak and pook? You already have a mismatch in second position; when you see another in third position, you need to know this time not to make the mismatch score 2.

Now my coding skills are not supreme, and I'm further thrown by the index of the first segment in a string being 0. The code I have for this type of neighbor detection is much more complicated ... 139 lines (including comments), 43 curly brackets, 40 ifs, and 3 whiles. There may be some ways of making it neater, but this works, which I've determined by using a small artificial lexicon. As I show on phonoloblog, my original intuition seems to be confirmed: most words have more neighbors when you allow geminate substitutions.

The last time I blogged about my coding exploits, it was about an alphabetizing script I'd written that was taking several hours to sort about 360,000 words. A commenter kindly pointed me toward a sort command that would automatically alphabetize elements in an array, changing the hours to milliseconds. I'm dreading finding out the same is true for neighbor detection, like the following imaginary script.

public String[] neighbors(String [a]) {
String b[] = new String[2];
b = a.getNeighbors(b);
return b
}

I bet it couldn't handle the geminates the way mine does, though.

07 June 2005

take {the} 101

A fun little shibboleth that marks an LA resident or native is the use of a definite article before a highway number. E.g., in the Bay Area you might take 101 or 880, but in LA, you take the 101 or the 405. Only having been in California for a year, I wasn't aware of how salient this was until this quarter, when my students all vehemently agreed that it's a pretty blatant SoCal marker.

A crude Google search backs it up. Search terms like [take-101] and [take-the-101] register lots of "directions" web pages. (6 of the first 50 for [take-101] do not).

Only 3 of the first 50 hits for [take-the-101] involve Bay-Area pages, with perhaps 6 Phoenix sites - not the same 101, clearly - and the rest in the LA area including Santa Barbara. [take-101] nets 37 or so pages from the Bay Area, along with 7 for the LA region (and some of these sites seem to be tersely stripped of definite articles in general).

At this point I can't remember clearly what we did with Ontario highways. Limited-access highways in Ontario are numbered with 3 digits, the first of which is 4, like 417 and 401. I can't reliably say whether I used "417" or "the 417". I do recall referring to other (non 400-level) highways as "Highway N" (e.g., "take Highway 7 through Peterborough"). Perhaps another google project will help clear it up.

[Update: a similar search with "417" reveals astonishing results. Of the first 50 ghits for [Take-417], 25 refer to a highway in Florida near UCF, and 25 refer to the higway from Ottawa to Montreal. Meanwhile, of the first 50 ghits for [take-the-417], 5 refer to the Florida highway and 45 refer to the Ottawa highway. This might be why "the 101" never sounded odd to me.]

04 June 2005

you take point

Several months ago I was playing Medal of Honor, a first-person shooter game set in WW2 Europe. I was on a mission through the Ardennes, and found myself on a snowy country road lined with steep hedges when an artificial-intelligence controlled colleague turned to me and said, "You take point". And my first thought was, what, we're on a power play?

A little digging revealed that taking the point (and thus being the point man) has several distinct meanings, one of which is military (derived from ranching/herding), and one of which is sporting. It's also not clear that the sporting use is directly applied from the military (or ranching) one.

Here's the military usage of point man from OED: The soldier who walks at the head of a patrol. With apologies for the epithet, the earliest citation is this: 1944 Yank 4 Feb. 9/1 The Jap point man was on the scene before any camouflaging could be done.

Oxford also cites point as the region on a hockey rink just inside the attacking zone and close to the boards:

An area just in front of the opposing team's blue line close to each edge of the rink, where players on offence are often positioned, esp. during power plays. 1953 Globe & Mail (Toronto) 27 Nov. 24/3 He missed a pass-out from Gord Howe and the puck went to Pronovost on the blue-line point.

I once heard the reason was that the blue line and boards form a right-angle, i.e. a point.

The point man is the player on the attacking team who holds this region during a power play (a man advantage, with an opponent serving a penalty). The point man intercepts clearing attempts made along the boards, ensuring the puck stays in the attacking zone, and time permitting, will also take shots and pray for a lucky bounce through the throng of players crowding the goalie. It's a position of trust and responsibility, but not leadership.

It would be easy, but perhaps inaccurate, to cite hockey's point man as a metaphorical extension of the military point man. The first 100 or so web sites that come up for {point-man} as a google search term all refer either to the military use of it, a metaphorical extention of the military use (e.g., leading a trade delegation), or to a religious movement. Only one mentions the hockey use ... of all places, the Mavens' word of the day, in a discussion of "point man":

Point in the sense 'the leading party of a military advance guard' is first recorded at the beginning of the 20th century. The point is also the position at the head of a column or V-shaped wedge of troops. Point man meaning 'the lead soldier of a patrol' dates from the mid-1940s; like the cowboy, he is said to "ride, walk, or take (the) point." (Actually he is sometimes positioned at the back of the rear guard, also a dangerous position.)


Later in the same discussion we see this:

In ice hockey, the point man is the player in the offensive position inside the attacking zone.

This is not quite precise enough. Despite the qualifier in the Mavens' entry, it seems like the dominant (non-hockey) use of point man is the lead position of an attack. However, in a power play situation, as I describe above, the point man is inside the zone, but trails the lead attacking group. And this is not quite bringing up the rear, in the sense of watching for counter-attack from behind.

I propose the trajectory of point man in hockey to be like this: sometime after the introduction of blue lines to manage offsides, the area of the rink came to be called the point. The player working this area became the point man. This may have happened independently, or it may have been encouraged by the frequency of point man in news media during the war. If that is the case, the metaphorical extension brought the phrase and its sense of a responsible member of an attacking group, but not its sense of leader of an attacking group.

Oddly enough, I have seen the leadership role in power-plays referred to as the quarterback. Another example of terminology crossing over sports.