| more __ | less __ | Totals |
| chocolatey | 3,060 (97.3%) | 84 (2.7%) | 3,144 |
| minty | 671 (87.8%) | 93 (12.2%) | 764 |
| gamey | 745 (73.3%) | 272 (26.7%) | 1,017 |
Although this is hardly an exhaustive list of flavor types, it's interesting to notice the asymmetry between positive and negative scalar comparisons. It's hard to say whether or not this is merely a frequency effect relating positive vs. negative scalar uses more generally, or if this is something specific about flavor-related language.
Assuming the latter, where might it come from? "Flavor" by itself is hard to define a scale for -- does Miller taste more like hops than Budweiser does like barley? Questions like that don't even make sense to ask, which makes it hard to make sense of a claim that one beer has "more flavor" than another. Even if we interpret that to mean that Miller's characteristic flavor is more intense on its scale than Budweiser's flavor on some other scale, it's not clear that the drinkers polled also meant that Miller's flavor was more pleasant than Budweiser's. Of course, in the absence of some solid statistics about the relative frequencies of "more" and "less", this speculation might merely rank high on the crackpot scale.
Mark Liberman discusses one of the obituaries for philosopher Sidney Morgenbesser provided by Arts & Letters Daily today. Since your host must confess that the first time he hear of Prof. Morgenbesser was when he opened up AL Daily this morning, he didn't expect that this would be the cause of any postings. It's clear enough from the obituaries that Prof. Morgenbesser was something of a wiseass, which is actually a compliment when it comes from this source.
In reading the New York Times' obituary, though, your host was caught by a sentence with amusing semantics:
In an interview yesterday, Noam Chomsky, the linguist at the Massachusetts Institute of Technology who agreed with Dr. Morgenbesser about some things and not others, called him "one of the most knowledgeable and in many ways profound thinkers of the modern period."
It's not Prof. Chomsky's quote that's interesting, but rather the definite description assigned to him: "the linguist at the Massachusetts Institute of Technology who agreed with Dr. Morgenbesser about some things and not others". Unpacking the ambiguity here is a great example of why all journalists should have to sit through a lecture on quantifier-raising before going off to spout their nonsense.
I can get four readings out of this description. Since notating them logically is a bit of a challenge without a decent equation editor, I'll put them in (hopefully) unambiguous prose:
1) There is only one linguist at MIT. That linguist is named Noam Chomsky. Noam Chomsky agrees with Prof. M. about some things, and disagrees with him about others.
2) There are many linguists at MIT. One of those linguists is named Noam Chomsky. For all linguists who work at MIT, if their name is not Noam Chomsky, then they either agree with Prof. M. about everything, or disagree with him about everything. Only Noam Chomsky agrees with Prof. M. on some things while disagreeing with him about others.
3) There is only one linguist at MIT. That linguist is named Noam Chomsky. At each university containing linguists, there is exactly one linguist who agrees with Prof. M. about some things and disagrees with him about others. If there are other linguists at other universities, they must either agree with Prof. M. about everything, or disagree with him about everything.
4) There are many linguists at MIT. One of those linguists is named Noam Chomsky. For all linguists who work at MIT, if their name is not Noam Chomsky, then they either agree with Prof. M. about everything, or disagree with him about everything. Only Noam Chomsky agrees with Prof. M. on some things while disagreeing with him about others. If there are other linguists at other universities, only one at each university may agree with Prof. M. on some things but not others; all remaining linguists at each university must agree about everything or disagree about everything.
Now, none of these strikes me as being precisely what the author actually meant. That would be:
5) Of all the people I could have called for this story, Noam Chomsky is about the only name that any of my readers will recognize. Chomsky is a linguist at MIT. Chomsky agrees with Prof. M. about some things and not others. No implications ought to be drawn about the number of other MIT linguists, or their views on Prof. M. Ditto for linguists at other universities.
The problem, of course, is that (5) is not actually something that can be extracted from the original sentence as written. Had the author written "Noam Chomsky, a linguist at MIT, who agreed with Dr. M. about some things and not others", it would have been quite clear from the indefinite article that there could be more than one linguist, and from the separation of descriptions provided by the comma that there are no implicational relations between Chomsky's status and other people's views about Prof. Morgenbesser.
It so happens that this isn't hard to figure out because the average linguist or philosopher knows better than to assume that ordinary language is really being used in the highly technical senses meant in their professions. In fact, most people probably won't ever be disturbed by readings 1-4, because they run so counter to ordinary usage. But you don't have to be a semanticist to be bothered by the sloppiness of that "the", with its unnecessary implication that some aspect of what follows is unique to Chomsky. Even if one doesn't take the time to work through all the possible readings, a description like that looks wrong on an intuitive basis. I wouldn't have thought twice about it if I hadn't read it and thought: "did this guy call everyone at MIT?". Of course he didn't -- but it wouldn't have taken much more effort for him to write clearly enough to avoid raising such questions in the first place.
SC and Mrs. SC went to dinner this evening at a trendy pan-Asian restaurant. To be frank, it was just OK. But as your host devilishly remarked as he signed the bill, "I may not have gotten a good meal out of this, but I'm getting a post!". What caused this comment?
Mrs. SC only ate about half of her meal, and requested a box for the remainder. When she did so, the waiter replied, "I can box it for you". Mrs. SC then told him, "That's fine". Upon the waiter's return to the table, he handed her a box and left. Mrs. SC was surprised.
"Wasn't he going to box it for me?", she asked. "No, you told him not to", replied SC.
The confusion stems from the ill-defined polarity of the phrase "that's fine"; we've addressed this before with regard to the question, "do you mind?". There are two possible interpretations, which we might expand like so:
1) That's fine, I would appreciate it.
2) That's fine, I can do it myself.
To your host, the second one is the more natural interpretation of the isolated "that's fine", but that intuition is certainly open to debate. Similarly, it seems to me that a falling intonation indicates (2), but again, introspection is not necessarily a reliable guide. A flat or rising intonation doesn't strike me as conveying (1) particularly strongly.
Pursuing my own intuitions further, it seems likely that if I wanted to say (1), I would actually say "that would be fine", and that this latter phrasing would never mean (2). But (1) is not an unreasonable interpretation of Mrs. SC's original utterance; had she said "that's fine with me", (1) would be clearly understood, and it's easy to imagine that she could have produced "that's fine" through ellipsis.
Of course, there's an easy way to avoid problems of ambiguity, whether one word is at issue, or the syntax of the entire sentence: just say something that disambiguates it. But if we all spent enough time trying to compute the potential interpretations of what we want to say before we say it, I'd be writing about the waiter who ran off in frustration rather than the ambiguity of words without explicit yes/no polarity.
In today's Wall Street Journal, there's an article on the annual Harris Interactive poll of "best brands". Since the article requires a subscription, we'll just reproduce the list, and then continue:
1. Sony
2. Coca-Cola
3. Dell
4. Kraft
5. Toyota
6. Ford
7. Honda
8. Procter & Gamble
9. General Electric
10. General Motors
The question that generated this list is simply: ""We would like you to think about brands or names of products and services you know. Considering everything, which three brands do you consider the best?". For completeness, here's the article's description of the polling technique:
Methodology: This poll was conducted online in the U.S. between June 10 and 16, 2004, among a nationwide cross section of 2,136 adults. Figures for age, sex, race, education, region and household income were weighted where necessary to bring them into line with their actual proportions in the population. Propensity score weighting was also used to adjust for respondents' propensity to be online. In theory, with probability samples of this size, one could say with 95% certainty that the results have a sampling error of ±2 percentage points of what they would be if the entire adult population had been polled with complete accuracy.
Although SC is often critical of specific facts about advertising, this is ultimately what it's all about. Every name on that list conjures up a whole host of properties, a few of which might even accurately be ascribed to them. Having taken a number of marketing surveys as part of a regular panel of volunteers, it's not hard for your host to imagine some of the word associations thatmight come up: "trustworthy", "innovative", "reliable", "quality", "a good value". Every one of these associations is worth billions, and has been hard-won.
In many ways, it would be quite erroneous to confuse this list with representing companies that are the best at what they do. While Kraft is one of the world's largest producers of processed cheeses, even the most casual cheese eater would likely admit that the equally mass-production Tillamook or Land-O-Lakes cheddars taste better than Kraft 2% milk singles. For the price, though, you could easily do worse. And so it goes through a variety of categories: Dell is simply the lowest-cost assembler of mostly generic parts (the obvious exceptions being the processor and video card), Coca-Cola produces a drink that most people think is worth no more than 60 cents a serving, but is good enough that they'll hand over $2.00 in the right contexts, and as for General Electric's presence on this list, suffice it to say that the poll considers GE an electronics company, and they haven't even made any of the clocks, radios or TVs sold under their own name in years. It wouldn't be wrong to say that this list represents the world leaders in "good enough".
That's not to say that each of these companies hasn't demonstrated a capability to produce products which are world-class in other respects (nor should the manufacturing capabilities that allow such cheap mass production be dismissed as trivial accomplishments). One of the terms of art frequently used by marketers is the notion of a "halo" product, something incredible that gives the brand status on the assumption that everything else in the product line is somehow related to it. This is especially an issue in the automotive industry, which is why Ford resurrected the Thunderbird, and then decided to bring back the GT-40 when the new T-bird didn't quite live up to that role in the press. It's why Toyota needed to create Lexus, and why Honda created Acura. It's why Sony originally created their "ES" electronics, even though they then quickly brought them right back down to the manufacturing standards of their regular products. Not everyone needs a "halo"; Dell gets by just fine without one, but that's as much because nobody would have any idea what the difference was between a luxury computer and a regular one as anything about the company.
Another study which focuses more on specific products, and which also is considered to be a benchmark for positive name associations, is the annual J.D. Power vehicle dependability survey, the 2004 edition of which just came out a week and a half ago. In some ways, the methodology is absolute garbage: although the company collects more detailed data from consumers, the reported statistic is "problems per 100 vehicles", which aggregates transmission leaks, buttons that fall off the controls, and comsumer disappointment that they're not getting quite as high gas mileage as they expected into a single nebulous category (a fact which explains Hummer's perennial ranking at the bottom of the initial quality survey). The results demonstrate that "dependability" is only really a factor in reputation for brands which have little else to go on. Lexus tops the most recent list, BMW is only slightly above average, and Mercedes-Benz is far below average, and yet the evidence from annual sales indicate that consumers are willing to in fact tolerate a lot more of the same problems out of the German brands listed here than they will out of domestic Lincolns or Cadillacs. The associations of other positive properties with the BMW propeller or the Mercedes star -- better performance, more luxurious interiors -- don't even have to be true at present because of the enormous contributions of history and advertising to the meanings of the names.
The problem with studies like these is that what's being measured is very slippery. In the case of J.D. Power studies, "problems" clearly have very little correlation with brand prestige, and fare only somewhat better in relation to sales rankings. And the Harris question provides absolutely no basis for justifying the opinions expressed by the people being surveyed. But these facts merely help us to understand that the relationship between experience and language is complicated. Is Sony better at cranking out CD players than Coca-Cola is at making soda? SC would wager that the people who ranked them 1-2 have no idea how they would answer that question, and might not even recognize that they had made such a statement. Similarly, statistical evidence of the sort found in the J.D. Power survey appears to have no impact whatsoever to people who insist on the superior quality of imported over domestic vehicles. Advertising and the historical experience of the 1980s are adequate to maintain those associations well after the facts that supported them have changed. Somewhere within the complicated mix of experience and advertising lies an opportunity to make a name as valuable as the products themselves, and the Harris poll shows that, this year, Sony has done it better than anyone.
Since the fastest way to ruin a joke is to explain it, we'll start off by quoting a line that opens an editorial in the Wall Street Journal today. Enjoy:
New Jersey is the butt of many jokes, and some of them are even unfair.
Your host brings this up because it reminds him very much of the unusual review of "Farenheit 9/11" penned by Christopher Hitchens recently, and discussed by Mark Liberman.
The technique Hitchens attempted to apply is illustrated by the line:
To describe this film as dishonest and demagogic would almost be to promote those terms to the level of respectability.
As Prof. Liberman noted, "The rhetorical trope in play is a routine one: 'To describe X as P is an understatement', where P is some scalar evaluative predicate." He goes on to note that P can be negative (i.e., "To describe this split as acrimonious would be an understatement.") or positive (i.e., "To describe this report as timely is an understatement.") The oddness of Hitchens' line derives from the fact that the implication in either normal version of the rhetorical device is that X exemplifies P to an extreme degree, but that clearly isn't how Hitchens has written things.
SC was prepared to dismiss it as "overnegation", a la Liberman, until he saw the Wall Street Journal line. The humor in it comes from the fact that the line, at least in a more charitable essay, would normally go, "New Jersey is the butt of many jokes, and some of them are even fair." That would also be intended as a punchline; what makes the Journal's line more biting is that it inverts the scale being used for comparison, rather than simply being negation of the scalar predicate.
Finding more examples like this isn't easy, as Googling for "some of them are even" yields plenty of irrelevant hits. But here's another one, a quote from a Guardian story regarding disgraced BBC reporter ([a redundancy -- ed.]) Andrew Gilligan:
Andrew gets great stories, and some of them are even true.
This one only works by inverting the scale; ending the line with "some of them are even false" would be merely anomalous.
The problem with Hitchens' attempt at using this inversion technique is that the syntax of the sentence makes it twice as challenging to interpret as it would be otherwise. Had he written it as something more like:
Farenheit 9/11 is dishonest and demagogic, and those descriptions even almost make it sound respectable.
it would have been a lot more clear what he was up to. SC gives him points for trying, but it's a stylistic flub from somebody whose craftsmanship is usually much better than that.
Following up in the SC tradition of comments on holidays, particularly Mother's Day, an observation about fatherhood and possible worlds.
As with Mother's Day, SC went shopping for cards, and was reminded of how incredibly fine-grained the categorizations of mothers could be. This time around, though, he noticed one category in particular that led to some additional thinking: "father-to-be".
Ostensibly, such a category is intended to apply to men who know that their wives/girlfriends/one-night-stands-with-good-legal-counsel are pregnant and expecting delivery in the foreseeable future. Another possible use -- one which SC is very leery of -- would be to inform said men that their wives/etc. are now pregnant and expecting. But an optimistic Hallmark product planner might go further than that, and hope that it would also appeal to those women who believe that their men will be fathers at some time in the future. Going even further, they might introduce a new category: "father-in-potential", which could be justifiably sent to any man who, in some possible world not necessarily their own, could be a father. This would expand the universe of possible card recipients to be nearly all men (minus a few very unfortunate types), and because of the additional revenue potential, SC is frankly shocked that Hallmark hasn't beaten him to this one.
Of course, there's one scenario in which your host can even imagine for sending Father's Day cards to women. In the event of a single mother taking on the roles of both mother and father, perhaps the really aggressive Hallmark card writer could imagine a card saying "Thanks for all you do, on both Parents Days", maybe with an appropriately feminine color scheme. And then they could get around to doing the same thing the next year for single fathers; SC suggests they might want to try introducing some Mother's Day cards featuring power tools, and see if it works.
On an unrelated note, those of you looking for "Huntington contra Nunberg" will have to wait until tomorrow, while SC takes the rest of his Father-in-Potential Day off.
Although there hasn't been much reason to mention it, SC is a big fan of the Detroit Red Wings. Those of you immediately suspecting gravy train/bandwagon-jumping tendencies should be aware that this has been true for some 14-15 years, and that your host in fact was in a fairly serious accident in 1999 as a result of driving from San Diego to Los Angeles to watch them play the Kings in the playoffs.
All this is by way of explaining how it is that SC occasionally reads the Detroit News (also an invaluable resource if you're an automotive buff), and so he was quite amused by this story regarding comedian Jimmy Kimmel. It seems that Mr. Kimmel said something which SC regards as a noncontroversial statement of fact:
“They're going to burn the city of Detroit down if the Pistons win,” Kimmel said during halftime of Game 2, using his customary snarky tone.
ABC broadcaster and Ann Arbor resident Mike Tirico objected immediately, saying, “Hey, hey, hey, be careful. That's my home state.”
Kimmel looked a bit stunned at Tirico's objections, then backtracked a bit by saying analyst Tom Tolbert's eye-popping plaid suit should instead be burned.
Needless to say, fans were outraged. The paper cites a local talk show host as follows:
Callers to Baligian's show were mad that Kimmel slammed Detroit without cause.
“Caller after caller, and people who were e-mailing me, were mad because what Kimmel said is simply untrue,” Baligian said. “It's not true. Detroit does not burn when teams win. It's a lie, and people are mad that somebody would say something so terrible about Detroit on national TV.
Unfortunately for the sports fans of Detroit, other people have long memories, and in spite of the city's relative calm after hockey championships (SC maintains that hockey fans are actually the most civil), the story was quite different after the Tigers' 1984 World Series win (which SC took particularly hard, having possessed one of the innumerable "Mow Down Motown" T-shirts then adorning San Diego). In spite of the fact that the web didn't exist even in anyone's mind for another half-decade, Google returns 260+ hits for "+1984 +"World Series" +Tigers +Detroit +riot". Among the evidence, a column from the Detroit News just two days ago which states:
On a more unfortunate note, Los Angeles picked up on the worst of Detroit traditions after the Lakers won NBA championships in 2000, 2001 and 2002. The post-game riot was originally made famous by the celebration that followed the Detroit Tigers’ World Series championship in 1984.
Legally, proving that a statement is true is a valid defense against a libel suit. Mr. Kimmel might be in some trouble here, on account of the fact that, while considerable fire damage accompanied the riots, the statement "the city of Detroit burned down" is at least arguably false, depending on how one defines "burn down" (does it have to be complete?). It's not clear that anybody's actually going to sue Jimmy Kimmel, but the libel accusation is just a legal formalization of the claim that a given proposition is a lie.
SC applauds the instinct of people to defend their home, and it would be inappropriate to expect that Detroit residents should flood the phone lines instead with self-criticism. But your host nevertheless has to question how one can simultaneously think both 1) a riot including much fire damage occurred in Detroit as a consequence of a sporting victory, and 2) anyone asserting (1) is a liar. It's a paradoxical mental state, and SC doesn't envy the people straining to maintain it.
[Edited at 1:24 p.m. on 6/10/04 to correct typo.]
A tragic story from Minnesota this morning, about a group of teenagers who died from carbon monoxide poisioning while exploring a cave near St. Paul. Apparently, when people start fires in the caves -- which this group did not do -- carbon monoxide builds up in the cave's atmosphere, and leaks out slowly at best.
It would seem desirable, given the peculiarly poor ventilation of these caves, to simply seal them off. In dismissing that possibility, the local fire chief said:
"There are entrances and exits that we don't even know of."
That's an interesting statement to make about an "known unknown" (we discussed the related phenomena of "unknown unknowns' previously here and here). It's a stronger claim than "there are probably entrances and exits that we don't even know about".
Your host finds himself quite confused about how to properly analyze the truth conditions of such a statement, in either probabilistic or "absolute" form. On the one hand, the statement strikes him as empirically unverifiable -- even if one went out in search of previously unknown entrances, and found a few, both versions would still be open to further testing.
And yet, it also strikes SC as intuitively erroneous to dismiss the claim as meaningless. It's quite compatible with a possible-worlds interpretation along the lines of "in all possible worlds, there exist entrances not in our model of each world". The probabilisitc version might even be tautologically true, softening the truth conditions to "there exists a possible world with entrances not in our model of this world" (although, as connoisseurs of possible worlds know, in principle no such world must necessarily exist).
Perhaps "unknown unknowns" and their half-known cousins are really just the rest of the world catching up to the philosophical discussion of counterfactuals. A brief search along these lines turned up an abstract that points toward a solution. It might be the case that positive statements about "unknown unknowns" are really just ceteris paribus claims with the necessary operators being implicit. If the fire chief's statement is rephrased as "all other things being equal, there are entrances/exits we don't know about", then we can more readily imagine how things might not be equal -- more time spent searching for entrances, maybe a little infrared imaging, maybe a couple tons of dynamite to get rid of the caves (note that SC isn't advocating that, it's just one way that our knowledge of the caves' condition might change). On the other hand, as Alice Drewery points out in the abstract linked above, we can judge claims like "there are entrances we don't know about" without knowing the full list of things that might change our belief. Perhaps it's not so much important that we have ideas about what sorts of things we don't know, as that we know what might go wrong with the things we do.
As your host has mentioned on a number of occasions before, he relies on taking a couple of trains to commute to his office. Frequently, in the course of waiting to change trains in the morning, he stops at the Burger King in the station to get a carton of orange juice, and if he's feeling hungry, an order of BK's not-at-all-good-for-you french toast sticks.. This morning, as with many others, the same dialogue played out:
SC: One orange juice and an order of french toast sticks, please.
Cashier: For here, or to go?
SC: For here, please.
Cashier: OK, your total is (X). [exchange of money, passage of time]
Cashier: [puts order in bag] Here you go.
SC: [shrugs]
It's not that the store doesn't have any trays. It's not even the case that the cashier has a pathological aversion to handling the trays. It's just that in this particular context, the question "for here or to go?" is only part of a script (not necessarily even one mandated by the restaurant, just socially). It's more efficient for the cashier to simply put everything into bags, which lets her move faster than trying to pay attention to what the customers actually asked her to do with each order. If she didn't go through the motion of asking, though, customers might be more annoyed by the disruption of the routine than they are by the matter of whether or not their request was actually acted on.
This same thing happens in other commercial contexts. In many grocery stores, when the bagger starts handling your order, they'll just start putting things in plastic bags, and mutter "plastic OK?", hoping that the customer will just ignore the usual dialogue, again because plastic bags are much more convenient for the task at hand. Occasionally, SC is moved to enough peevishness to wait for them to put in a few items and then say, "I'll take paper, thanks", but only if the bagger is acting especially lazy.
Even in social contexts where people are speaking as equals, and not as customer/employee or some other formal relationship, there are plenty of exchanges where it's a mistake to respond to a request literally. When somebody says "how are you?", the last thing they actually want to hear about is that you put in 15 hours at the office yesterday and your goldfish just died ([these are not actually biographical details of SC, at least not recently -- ed.]). They want to hear "fine", "good" or some other short, positive response. Of course, the degree to which this is true varies by relationships -- the people SC happens to see on the train platform each morning are more likely to prefer a short, information-free response than Mrs. SC.
So the next time you feel like nobody's listening, just remember -- they're only doing it because they're trying to serve you. Or at least it's nice to tell yourself so.
As part of SC's new life as a marketer, he has to pay careful attention to the wording of proposal solicitations. It's important to understand what your prospective customer is actually asking for, which may not always be what it sounds like they're asking for. But sometimes, it's best not to take their requests too seriously.
Consider this passage from a current DARPA solicitation:
How can plausible but not logically valid reasoning be used in a consistent and pragmatic way to get to reasonable conclusions? Can we build a real-world-scale cognitive system that has a principled foundation for its representation and reasoning capabilities?
It's hard to determine what coherent meaning ought to be assigned to this statement. Intuitively, your host wants to agree that there is such a thing as "plausible but not logically valid reasoning" -- if he says, "It's a sunny day out; therefore, I'll spend the day in front of my computer", it ought to be a reasonable inference to anyone who knows he's a geek. But it's hardly logically valid, unless we're starting from a world model including the axioms: "If it's sunny out, vampires will die"; "Spending the day in front of a computer will prevent an individual from being exposed to sunlight"; "SC is an individual"; and my personal favorite, "SC is a vampire" ([the only evidence for that is your screwy sleeping schedule, bat-boy -- ed.]).
It's not at all clear to SC, though, that "plausible but not valid" meshes well with "consistent", at least not where mechanical reasoning processes are involved (although the hardcore philosophical determinist will reply that this just means the world model isn't sophisticated enough). Humans make statements that at least look like inferences all the time, but which aren't really logically coherent. That's what the DARPA folks are trying to get at, but SC's not sure how desirable it really is to have machines emulate that sort of behavior, especially not if they're making life-and-death decisions. Taken as a whole, the statement above reads: "Make us a principled system that makes unprincipled conclusions". Brilliant!
Brian Weatherson passes along news that a blog penned by someone named Noam Chomsky has recently started up. Brian expresses surprise, though, that the author refers to himself by full name; since SC likes the way he puts it, we'll quote:
I wouldn't have guessed that Noam Chomsky calls Noam Chomsky "Noam Chomsky", but if it's good enough for Rickey Henderson I guess it's good enough for the Noam.
SC is not at all convinced, though, that the statement "Noam Chomsky has started a blog" is true. Although the signature after each post says "Posted by Noam Chomsky", your host thinks the opening post indicates something rather different:
This blog will include brief comments on diverse topics of concern in our time. They will sometimes come from the ZNet sustainer forum system where Noam interacts through a forum of his own, sometimes from direct submissions, sometimes culled from mail and other outlets -- always from Noam Chomsky.
In other words, the content will always be statements made by or articles written by Prof. Chomsky. However, if by "Noam Chomsky's blog", we mean some entity which is a member of a set defined as "a web publication where Noam Chomsky personally has sat down at a computer and used the Movable Type interface to post an article", then SC submits that the set is empty and the phrase "Noam Chomsky's blog" refers to an entity much like "Pegasus" and "the present King of France". While your host knows a thing or two about writing in the third person about himself, he reads the quote from the blog to mean that it's a clipping service run by a third party (apparently Z Magazine) with Chomsky's endorsement.
Over the weekend, the invaluable Arts & Letters Daily publicized a story by a woman who apparently was assigned a cell-phone number previously belonging to comedian Chris Rock. Predictably, she couldn't resist the urge to commit a little mischief when famous people called; SC loves this story:
LAURA: [Curious, and ready to tackle the unknown] Hello?
CALLER: Is Chris there?
LAURA: [Inquires politely] Who's calling?
CALLER: It's Spike.
LAURA: [Mischievously inquisitive] From...?
CALLER: [Blurts out, in an annoyed tone] It's Spike Lee.
LAURA: [Speechless, stunned, mouth frozen open. Guess wasn't quite ready enough to "tackle the unknown." Takes longer than usual to respond, and when does, does so very slowly] Uh... well... actually... you have the, uh... wrong number.
Linguistically, the story raises a couple of interesting questions. Most of the vignettes feature people calling on behalf of someone else, like so:
LAURA: Hello?
CALLER: Hi, I'm calling from Jerry Seinfeld's office. Jerry would like to get two tickets for Chris' show in L.A. this weekend. Would that be possible?
Clearly, the caller wasn't actually expecting to get Mr. Rock (SC feels a bit like the New York Times, famously referring to Meat Loaf as "Mr. Loaf"), or he would have reacted with surprise to hear a woman's voice. So when the author says that she was assigned "Chris Rock's cell phone number", it's ambiguous between two interpretations: 1) a number belonging to someone in Chris Rock's entourage, or 2) Chris Rock's personal cell phone. Not being fortunate enough to be acquainted with Cell Phone Etiquette of the Rich and Famous, your host cannot reasonably speculate on how often callers expect that Mr. Rock personally answers "his" cell phone.
But there's a specific case which your host finds especially interesting:
LAURA: Hello?
CALLER: Hi, is Chris there?
LAURA: Who's calling, please?
CALLER: It's Jack Nicholson.
In the next millisecond, frozen in time, I nearly dropped dead, but then I thought, "Wait a minute." Adam Sandler totally sounded like Adam Sandler. I quickly put two and two together here, and they didn't add up. With a hefty dose of skepticism, I boldly commented, "This doesn't sound like Jack Nicholson." The caller replied, "I'm his assistant."
Previously, we've talked about rigid designators, which are explicitly referential expressions. Although for most people's names, we know that there are actually plenty of people that the name could refer to, celebrity names surely must approach rigid designator status in most people's mental models. When one says "Madonna", it's far more likely that they're referring to Louise Veronica Ciccone than to anyone else. Similarly, your host suspects that this person is the usual referent of "Jack Nicholson".
So, strictly speaking, "It's Jack Nicholson" was a false reply to the question of who's calling. One response might be that in fact, the caller's statement was merely shorthand for "it's a designated agent of Jack Nicholson". But such talk is: 1) cumbersome, and 2) reflective of a more sophisticated philosophy of language than the Hollywood crowd has any real interest in/need for. Perhaps, then, in a social circle where the names are all understood to uniquely refer to particular individuals, we can analyze names as being rigid designators which can be assigned by an act of the individuals they refer to. It would be absolutely disastrous for credit card companies to accept this sort of behavior in discourse with hundreds of thousands of clients, but when the social group is a little more intimate, flexible designation makes perfectly good sense.
By now, the news of the bombing in Madrid is common. As in the first day or so after September 11th, the question of responsibility is open, and subject to considerable speculation.
Geoff Pullum, writing about a discussion on the BBC World Service, quotes a Prof. Halliday of the London School of Economics as saying "For a start they speak a language that no one can learn." This comes in reference to the Basques, an ethnic group living near the French/Spanish border. The language is famous for being wholly unrelated to anything in geographic proximity (or elsewhere, really) which would account for its origins. Prof. Pullum explodes the notion that "no one" who isn't a Basque can learn the language. But Prof. Halliday ought to have been vaguely embarrassed to even have originally put it like that.
SC assumes a logical form for the statement which is something like (using A for the logical "all" operator):
AxAy( (Person(y) & Is_Basque(y) & Speaks_Basque(y)) & (Person(x) & !(Speaks_Basque(x))) )
If we wanted to be really rigorous about it, we could add terms making explicit the relationship between speaking Basque and learning it, but SC will assume that anyone who speaks Basque managed to successfully learn it. Otherwise, I've tried to capture the part that Basques speak Basque with the first group of terms, and that "no one" speaks Basque with the second group. Regardless, assuming our variables range over the set of individuals, some x and y values will be identical -- i.e., Basques -- and the statement is a contradiction. Not that one really needs to have written that out in logical terms to have seen the problem.
It's not at all clear, though, that ETA, a Basque terrorist group, is responsible for the attack. Semioticians, the sort of people who are interested in figuring out why particular language items carry symbolic meaning (as opposed to semanticists, who worry about how), will be quick to note that today is the 11th of the month. And the evidence is somewhat mixed in regard to possible suspects. Investigations of this sort of thing ought to turn on real evidence, not the speculations of linguists, but it strikes SC as likely that whoever is responsible didn't pick the date accidentally.
Two weeks ago, negation was the hot topic of linguistics blogs. SC's still out looking for examples (not too consciously, just taking notes when he finds them) and came across an interesting one this morning.
From Rob Neyer, SC's favorite baseball writer:
"Oh, and this doesn't really fall under the purview of today's column, but what are the odds against Esteban Loaiza winning 21 games and posting a sub-3.00 ERA again? Pretty long, I would guess."
In the context of the column, it's clear that Mr. Neyer's intent is to express the opinion that Esteban Loaiza will revert to his usual unfortunate state of Esteban Loaiza-hood, and put up 10 wins and a 4.5-ish ERA ([which will be 10 more wins, a real number of runs allowed, and several million dollars more than you'll be seeing, pal -- ed.]). However, as written, Mr. Neyer has claimed that it would be unlikely for Mr. Loaiza not to repeat the performance, as the odds against him are what are being called "long".
So the SC research staff checked the relative frequency of some "odds" phrases in Google:
"odds of *": 860,000 hits
"odds for *": 765,000 hits
"odds against *": 94,700 hits
A cursory examination of the results suggests that "odds of" is pretty much used correctly, as a description of the probability of something (albeit often metaphorically, and not with actual calculations). "Odds for" is similar to "odds of". "Odds against", though, has two vaguely opposing uses:
From a recent MSN Money article: "Odds against an audit are good"
From the Houston Chronicle, June 28, 2003: "Rudy T's doctors like odds against cancer".
From CNN, September 5, 2003: "Olympic swimmer defies odds against diabetes"
In the first case, the parse is something like NP[odds against an audit] VP[are good], where the meaning of odds is exactly like the for/of cases discussed previously, and the odds in question are low. However, for the latter two, it's something more like S'[Rudy T's doctors like odds] PP[against diabetes], and the expression is meant to signify that the odds being discussed are highly likely, and favor the overcoming of whatever's against them. Here, the negation isn't carried in the lexical meaning of "odds"; it's buried in the syntax.
Returning to Rob Neyer's sentence, it's fairly clear that what happened is a confusion between the two parses because of subject-auxiliary inversion. "What are the odds against X?" looks, on the surface, like the second parse of "odds". Had he first written "The odds against Esteban Loaiza repeating his 2003 performance are long", it would have been immediately obvious what his mistake was, and he would have instead written: "What are the odds against Esteban Loaiza repeating? Pretty good, I'd say." In case Mr. Neyer should happen to come across this post, I only point it out because I know he cares about getting these things right.
Yesterday, SC wrote about the frustration of watching someone violate the terms of using an express lane in a particularly egregious manner. This morning, he realized he has a better analysis to offer.
Recalling that the woman said "I only have ten items", when she had more like double that, SC originally wrote "it's difficult for SC to decide if her statements in fact qualified as lies". This was due to the fact that she could have meant "I only have ten items in each order". Your host has changed his mind -- she was a liar.
The basis for deciding this is not mere peevishness ([although you'll have a hard time convincing anyone who knows you, pal -- ed.]). It's the fact that she said "only". Larry Horn provides this definition: "The semantics of only says this: it asserts that no proposition from the set of relevant contrasts C other than the one expressed by its sister sentence [alpha] is true". I'm not sure if he's attributing that formulation to Kai von Fintel based on the way the abstract is posted; if so, consider the credit assigned. Regardless, we can represent her statement as a conjunction of logical statements:
Ex((Rude woman(x) & (Possesses_10_items(x)) & ~(Possesses_11_items(x) & ~(Possesses_12_items(x)...)
SC is sure the speaker who prompted this posting would disagree with the formulation's first term, but SC's not terribly interested in hearing her complaints right now.
Given that her statement was "I only have 10 items", the above formula is only true if there exists a person x who is a rude woman and possesses exactly 10 items. If she possesses 11 items, then the term Possesses_10_items remains true, but now Possesses_11_items(x) is also true, and so its negation -- the actual term in the formula -- is false. Since "and" statements are only true when all of their conjuncts are true, her statement was false, and she was a liar.
But wait! What about the pragmatic claim that she might have actually meant "I only have 10 items in each order"? Let's resolve that by recourse to Grice's maxims. Readers encountering them for the first time may think that Grice was lecturing about how people should speak. In fact, Paul Grice offered these as his formulation of the rules which appear to guide our approach to conversations. They're an approximation of our expectations about how other people communicate with us, as well as how we try to formulate our own statements (although some people disagree vehemently). SC thinks they're a good starting point.
So how might she have violated Grice's maxims? The issue of quantity seems to vindicate her, as Grice's conditions for that could be construed by a malicious speaker as "say no more than the minimum necessary to establish your position". Since she said something which was true under a very strained interpretation, SC will give her a reluctant pass here, even though he thinks she's a weasel. The issue of quality is a toss-up; on the surface, she made a clearly false statement, but we're analyzing the "in each order" interpretation here. She could truthfully claim that she doesn't believe the "in each order" statement to be false, and SC would have to agree (again, while swearing furiously). The evidence is right there in her cart.
As for relevance, well, she was definitely addressing the situation; namely, the question of whether she had too many groceries to be checking out in the express lane. So she's clean. But then we get to manner. Did she avoid obscurity? Absolutely not. Did she avoid ambiguity? Nope -- she only gets a pass on the first two maxims by assuming that she was very ambiguous. Was she brief? Well, yes. Last, but not least, orderly? Hmmm...this would-be parliamentarian rules her..."Out of line!".
So she strikes out on both truth-conditional and pragmatic grounds, and we can now conclude that she's a liar. However, due to the fact that SC has now been obssessing about this for at least 15 hours, he'll conclude with an apocryphal story from which you can draw your own conclusions about his mental health.
Two monks who have sworn never to touch a woman are walking by a river. They come across a woman who is desperate to cross, but cannot swim. One of the monks promptly picks her up on his shoulders, and carries her across. The other monk says nothing. Hours later, though, he can no longer restrain himself, and says, "How could you touch that woman? You've broken your vows!". The monk who carried the woman replies, "I put her down by the side of the river. You've been carrying her ever since."
Like most people, your host spends a lot of time worrying about a fellow named Mort. Mort Gage. Mr. Gage discovered SC's checking account some time ago, and removes a large part of it with disturbing regularity. Judging by the number of people who appear to be similarly afflicted, Mr. Gage must be a very wealthy man by now, and so SC wishes he'd decide he's satisfied and just go away.
Since, however, the shadowy Mr. Gage has not indicated plans to do so anytime soon, SC finds that he spends a lot of time attempting to compensate for the ravages of Mr. Gage's attentions. As a consequence of this, your host is well acquainted with the writings of various members of the financial and business press, including the Wall Street Journal and TheStreet.com.
This morning, writing for TheStreet.com in regard to Comcast's abortive attempt to buy Disney, hedge fund manager Doug Kass commented that:
Unlike my view, with Microsoft playing an integral role, Reif reminds us that CMCSK can call upon investments in Time Warner, Liberty Media, Time Warner Cable and multiple cable partnerships in order to raise the cash component.
Over a number of years, SC has noticed that people working in finance will frequently refer to companies by their stock symbols. This is not an unreasonable shorthand, so long as it's unambiguous. But in Comcast's case, CMCSK is not identical with stock representing ownership of the entire company! In fact, there are two different stocks held by different owners of Comcast, as well as a third stock representing additional assets owned by Comcast as a parent company. And there are other companies where the ownership is even more fine-grained, like General Motors, which has no less than 12 distinct stocks representing ownership of different parts of the company, or different classes of shareholder rights for the same part of the company.
So when Mr. Kass says that CMCSK is pursuing ownership of Disney, a reader with SC's general disposition might conclude that in fact the decision is not unanimous among all owners of Comcast, and that holders of CMCSA disagree. Of course, such a reading is not the intent of Mr. Kass, nor of anyone else SC can recall writing this way. If the sentence actually was meant to be read as referring exclusively to only some portion of the company, it would become practically impossible to read financial writing -- does part of Comcast intend to acquire all of Disney? Or does part of Comcast intend to acquire one (or more) of the 7 publicly traded parts of Disney? For any company of significant size, the number of potential interpretations of the sentence would become proportional to the number of stocks representing interest in the company.
Of course, given the substantial amounts to be made in merger-and-acquisition banking, to say nothing of the legal fees, some enterprising banker or lawyer might come across this discussion of agonizing hairsplitting and say, "What a great idea!". Rather than disabuse them of the notion, SC will simply point out that valid contact information is available on this website in order to discuss SC's cut of the resulting deals.
Kai von Fintel has thoughtfully provided a link to an interesting and worthwhile paper for those interested in the polarity of yes/no questions. It's a tad dense, so SC won't read it all for a day or two, but for those who aren't going to be reading the paper at all, your host would like to share the sorts of examples, found in this paper, that make doing semantics fun:
Scenario: S hates both Pat and Jane. The prospect of an excursion
without them pleases S. S does not have any previous belief about
whether either of them is coming or not.
A: Pat is not coming.
S: Great! Is Jane not coming (either)? That would be the best!!!
S’: # Great! Isn’t Jane coming (either)? That would be the best!!!
Scenario: Michael has been upset at Sue since yesterday’s meeting.
The speaker is wondering how this could have been avoided. The
speaker has no belief about what Sue should or should not have done.
A: Michael has not been happy with Sue since yesterday’s meeting.
S: Should she not have talked to him (at the meeting) / (yet)?
S’: # Should she not have talked to him already?
If the reader is beginning to suspect that semanticists are actually closet romance novelists -- YOU'RE RIGHT!
More seriously, this is how semantics usually pro |