Semantic Compositions

Homonyms: bad for business

There are plenty of conspiracy theories to be found about the warranty policies of consumer electronics manufacturers. Suffice it to say that anybody who wants to only give you a 90-day warranty knows perfectly well that the mean time between failures of at least one critical part is between 4-6 months, and folks who offer you a year are making similar bets, albeit with slightly longer lead times. Of course, if failure rates are high enough, even if products routinely outlast their warranty periods, it's bad for brand equity and bad for business. Suffice it to say that SC can never be enticed to purchase another Sony portable device, having gone through 3 CD players that all would have done the Impossible Mission Force proud ("This CD player will self-destruct in 5 seconds after the expiration of the warranty...").

Apple Computer is now suffering through accusations of somewhat less than stellar build quality, per this Chicago Tribune article from earlier in the week. From reading the article, it sounds like the major problem is with the miniature hard drives in higher-capacity units, which are fairly fragile parts under the best of circumstances. SC's Shuffle has been suffering through some 9 months of abusive treatment in airplanes, on trains, and in the gym, and while the one-year warranty hasn't quite run out, it doesn't strike him as likely that the device will fail a la Discman. Of course, the Shuffle also doesn't have a hard drive.

Now, Apple will sell you a maximum warranty of two years, which would seem to imply that they expect failures to occur around year 3. So it would seem to have been less than horrendous for the same Chicago Tribune article to quote a corporate suit like so:

An Apple spokeswoman, Natalie Kerris, said iPods have a failure rate of less than 5 percent, which she said is "fairly low" compared with other consumer electronics. "The vast majority of our customers are extremely happy with their iPods," she said, adding that an iPod is designed to last four years.

By the standards of other portables, this wouldn't be terrible at all. Dell should try this hard with much more expensive devices. However, nobody in their right minds in consumer electronics would ever acknowledge that they make any kind of statistical estimates about product longevity, or that they plan for something less than heirloom durability in their manufacturing. Their competitors would immediately jump all over them in advertising, ignoring the fact that they make the exact same calculations.

So in that light, it is not at all surprising that the same Apple employee quoted in the news article spoke to an iPod enthusiast site's editor today, and insisted she meant "for years", not "four years". This strikes SC as wildly unlikely on linguistic grounds; even if there is absolutely no contrast in vowel quality or length between Natalie Kerris' "for" and "four", wouldn't she have different intonation patterns between these two? As in:

"We expect iPods to last FOUR YEARS" (quantifier interpretation)

"We expect iPods to last for YEARS" (preposition interpretation)

SC thinks this is a case of a Freudian slip that needed to be corrected to remain correct to the PR truth that no product is ever expected to fail, ever. Good thing for Apple that this is one speech error with just the right amount of plausible deniability.

On the difference between infrequency and strangeness

About 10 days ago, Radagast put up a link to a site focusing on disastrous baby names, a subject which SC approaches with some trepidation, one man's trashy taste being another man's treasure. However, a comment there caught SC's eye for reasons rather tangential to the subject matter at hand. Mixed in with the collected discussions of naming boys River Sunshine and girls Kali (never let the poor girl see Indiana Jones and the Temple of Doom) is this remark:

I have an unusual name (Abigail) and I've never had any problems.

The site's owner mocks this comment with "Um. Yeah.", since the name is a fairly well-known one, if not always frequent (although it's apparently taken off wildly in popularity in the last 5 years; check it out in the Baby Name Wizard). But there's a subtler point as well, which is that there is a difference between a word being frequent and being outright alien.

Zipf's law tells us that there is an inverse relationship between a word's frequency and its rank in a table of word frequencies within a language. But it is silent on the question of whether words down at, say, the 20,000 or 30,000th spot sound like typical examples of the language.

To try to ilustrate this point, SC went in search of word-frequency lists, which is never an easy task (darn English, stop changing long enough for someone to get a good, reliable measure!). One helpful resource came from a book that your host wasn't previously familiar with, but should have been, titled "Word Frequencies in Written and Spoken English". Using the British National Corpus, they provide a listing of about the top 5,000 words by frequency (they don't give that number, that's just SC ballparking it by counting the number of times he had to hit "page down" to cover half the list). No sign of Abigail anywhere, but "Gaulle", "Nelson", and "Macdonald" can all be found right near the bottom. So can "colitis", which just proves that you shouldn't pick a name by frequency alone.

A somewhat more dated (1996, as opposed to 2001) collection compiled by Adam Kilgarriff (also based off the British National Corpus, but apparently an earlier version) provides an even greater opportunity for examining names as having an unusual nature (seeing as it covers some 921,000 distinct words). Abigail shows up 78 times per million words; so do "acrimony", "acronym", and "adorned", all reasonably well-known words. So much for infrequency meaning unfamiliarity. But how about other names? Say, "Rex"? It shows up as a noun phrase 805 times per million words (some of which may be Latin citations, not occurrences of a name), and SC would reckon it's more unusual than "Abigail". "Bruno" and "Brutus" check in at about 243 and 114 citations/million apiece, and SC would regard both of those as sounding more unusual than "Abigail" as well.

But what, exactly, is the baseline against which these are measured? A name can only be considered common or uncommon relative to other names, not really to "colitis". "Smith" and "Jones", those most common of English surnames, are only good for 7565 and 4447 citations per million themselves. "John" checks in at a comparatively imposing 30,111, but it should be kept in mind that this covers all citations as a noun, which may include a good number that don't refer to a person. "Bob" manages some 3,679, and that's got to be fairly high up as names go, as well. So if Abigail occurs maybe 2% as often as "Bob", that's hardly a sign of an uncommon name. And yet, it occurs even less often than Aethelred (131 cites/million), which is unquestionably a more unusual name to modern English speakers (although it was almost certainly the other way around 1,000 years ago; things change).

All of this is not to say that poor Abigail was entirely wrong in her intuition. Relative to other names, hers could certainly be more common, and she's obviously picked up on that. But we all know a good deal many more words than we use every day, and the fact that we might go some time between using them hardly makes them strange.

Japanese in Hawaii

Although SC has many other pictures to share from Hawaii, the post he has been looking forward to writing most is about the relationship of the Japanese language to the islands. This is because, aside from being the most linguistically interesting thing about the islands (to SC), it provides some genuine insight into the culture of present-day Hawaii. In what follows, if you have trouble reading anything in the pictures, just click on them to go to Flickr, where you may have a look at larger versions.

Hawaii is almost as close to Japan as it is to the American mainland, so it's not at all a surprise that the islands are filled with Japanese tourists. This was especially true in Honolulu, and around Oahu more generally; we didn't observe this to be so much the case on Maui. Due to the inevitable need of businesses to cater to their audiences, this means that signage is frequently bilingual as well -- in drugstores, on street-facing windows, and especially on menus. Again, this was particularly prevalent in Honolulu, where almost no menu could be found without a Japanese version -- if it wasn't printed fully bilingually. Here are two pictures that illustrate the case, both details of menus from Jackie Chan's (yes, that Jackie Chan) restaurant at the Ala Moana Center in Honolulu.

The above version is all in English. The version below is the same menu, but printed in Japanese with both English and Japanese versions of the names of each item, as well as an all-Japanese description. The details are from different sections of the menu because it was all taken in one shot -- SC wasn't feeling very optimistic about getting this done right through glass, and didn't feel like doing multiple takes. The menus are posted one above the other.

Now, while these menus are identical in content, it is possible to glean some sociological facts from other menus. For example, the restaurateurs have figured out that Japanese tourists are likely to eat lunch out, but less likely to have dinner out as well. Perhaps this has to do with taking cruises, or being herded back onto tour buses around dinnertime, but regardless, this is a distinctly different pattern of behavior than is expected for English speakers. How did your host discover this? Witness a pair of menu details from an upscale Italian restaurant in the same mall:

The above picture is from the restaurant's lunch menu. Note that it's fully bilingual, providing equal information in both English and Japanese -- perhaps even a bit more in the latter. Now have a look at a section of the same restaurant's dinner menu.

Note the absence of Japanese entirely. This isn't part of any sort of conspiracy or discrimination; it simply reflects the different eating habits of the tourists from one country versus other tourists, or locals.

This solicitousness extends to areas other than language. On a visit to the Dole Plantation, one can eat a not-half-bad dessert called Dole Whip, a non-dairy pineapple-flavored treat somewhat like soft-serve ice cream. Would SC eat it again? Sure, although he'd just as soon go for some of their much-more-enjoyable pineapple-cinnamon bread. Why non-dairy, though? Hint: the Japanese diet features a lot less milk, and a lot more soy substitute for same. Could they make a better product by using milk? Almost certainly. But then half their visitors would be unlikely to eat it. If you're interested in trying Dole Whip for yourself, you can order the powdered form here, and recreate it in a home ice cream maker -- expensive restaurant mixing equipment isn't required. You can also try the stuff at Disneyland's Tiki Room, although for some reason it escaped this longtime annual pass-holder's attention until now.

It's not hard to imagine, given the dependence of the islands on Japanese tourism, that if they were to secede from the United States for any reason (the controversial Akaka Bill being one possible trigger), they would undergo closer economic integration with Japan, possibly as a satellite territory. Particularly in Honolulu, the yen functions as a widely accepted second currency, with many stores willing to accept them directly. SC has lived in Southern California for 27 of his 29 years, and cannot recall seeing anything similar with the Mexican peso (although he might be aware of such examples if he spent more time in border towns like San Ysidro). This is hardly to suggest that any such event is likely in the future, but rather to point out that being able to deal comfortably with Japanese tourists on their own terms is key to running a successful business in Hawaii, and that this fact manifests itself in a variety of linguistic and sociological forms.

But sometimes, attempts to accomodate the Japanese-speaking crowd don't turn out as expected. One popular attraction at the Dole Plantation is a train ride, on a 1/2-standard-gauge train called the Pineapple Express. The train has three cars, with the forward two playing an English narration, and the third in the rear playing a Japanese translation of the tour. After being seated in the middle car, SC overheard the following exchange between the ticket-taker and a Japanese-looking woman who was last in line:

Ticket-taker: We have a Japanese tour. (Gesturing toward rear.) Japanese, back car.
Tourist (in flawless American English): That's alright, I'll come back later.

SC's tropical trip

SC is back from Hawaii, and while your host is kicking himself for forgetting to set some posts to "publish on" from "draft" (hence the lack of automatically posted material while gone), he's got plenty to share from his time in the islands. Aside from being a gorgeous tropical paradise, there are all sorts of language-related observations to make about the islands, but we'll start off with the ones that don't need pictures ([because he's been too busy showing them off to his family over the weekend to have scanned them yet -- ed.]):

First off, SC was surprised to learn that pineapples are about as Hawaiian as Spam. This analogy is not as inept as it sounds -- while there is absolutely nothing native to Hawaii about Spam ([and nothing natural about it anywhere -- ed.]), it has been rather enthusiastically adopted as an alternative to edible meat based on the comparative ease of importing it over unpreserved meats. The pineapple is similarly foreign; the very word for it in Hawaiian, halakahiki, translates into English as "foreign fruit". While the fruit was introduced to the islands in the 1600s, the Dole Plantation's own account of pineapple history (see previous link) suggests it wasn't much of a factor until the 1880s.

There are several ways to view the relationship between Hawaiians and pineapples. One might say that this demonstrates a certain inauthenticity to all of the pineapple imagery associated with their tourism business. Alternatively, one might see the relationship as much like that between Italians and tomatoes -- while the tomato is hardly native to Italy, it has become so utterly intertwined with their culture that it would be churlish to make a show of questioning its status. SC's sense is that there is still something of a shtick for the tourists about it -- this song, played on the train at the Dole Plantation, encourages that view -- but there's no denying that pineapples and Hawaii have been good for each other.

Speaking of that linked song, it clued your host in to a longstanding mispronunciation on his part. The name of the instrument known as the uku'lele does not begin with a diphthong. The appropriate IPA transcription (using a question mark for a glottal stop) is /u.ku?.le.le/. Hawaiian does possess the /iu/ diphthong, but if it was meant to be pronounced that way, it would be written that way.

A common contraction for macadamia nuts, which SC had not previously seen (although Google shows it to be common to just about all macadamia growers), is to refer to them as "macnuts", or somewhat more frequently, "mac nuts" (the space occurring about 3x as much, according to Google). It makes sense -- macadamia is a long word to fit on a sign -- but your host can't help but feel every time he hears that word that "Macnut" should be the name of a Hawaiian tragedy.

Last up for now is a name for a common Hawaiian meal, known as "plate lunch". Contrary to the paradigm established in phrases like "steak dinner" or "turkey dinner", a plate lunch does not center around the eating of a plate. Instead, it refers to something that fits rather neatly onto a standardized styrofoam plate with multiple compartments for separate items: two scoops of rice, one scoop of macaroni salad, and some sort of meat, often cooked in teriyaki sauce or heavily fried. SC can't get that notion that he's supposed to eat an actual plate out of his head, but perhaps this is only because that version of a "plate lunch" sounds so very much like what he'll have at his company cafeteria today.