Semantic Compositions

SC's brain, mapped

Oh, great.

Courtesy of the BBC -- the ever-so-trustworthy folks who previously brought you the telepathic parrot -- a story on a study published in the journal Neuropsychology, demonstrating how damage to various brain regions correlates with an inability to grasp sarcasm. Isn't that special?

The ability to grasp sarcasm was tested by reading people stories like so (there were 8 kinds; this is the only one presented in the full paper):

A Sarcastic Version Item
Joe came to work, and instead of beginning to work, he sat down to rest.
His boss noticed his behavior and said, “Joe, don’t work too hard!”

A Neutral Version Item
Joe came to work and immediately began to work. His boss noticed his
behavior and said, “Joe, don’t work too hard!”

They would then be asked two questions:

1. A factual question (assessing story comprehension): Did Joe work
hard?
2. An attitude question (assessing comprehension of the true meaning
of the speaker): Did the manager believe Joe was working hard?

The paper indicates that the sarcastic stories were read only with a "sarcastic intonation", and the other stories with a "neutral intonation". That raises red flags for SC, who is reasonably sure that people exist who say sarcastic things all the time without always giving blatantly exaggerated phonetic cues ([yoo-hoo, over here! -- ed.]).  In fairness, the questions would seem to screen out people who simply can't figure out what's going on -- sarcasm errors were only counted when subjects answered question #1 correctly -- but still, there's more to sarcasm than a snide and flippant intonation. Or a simple mismatch of facts and utterances ([you'd know, wouldn't you? -- ed.]). But who said anything about being fair?

The authors are aware of this problem, and also conducted tests on their subjects' ability to recognize when someone has commited a verbal faux pas, and what a speaker's emotional affect (happy, sad, angry, etc.) is. The working hypothesis here is that people who have a hard time recognizing sarcasm probably have a hard time constructing theories of other people's minds (the researchers use the acronym ToM), which is held by some researchers to be the basis of autism. The figures obtained by these tests suggest that it's at least part of the story:

The relationships between sarcasm and social cognition (ToM, affect recognition) were examined with a correlation analysis. To examine the pattern of correlations between sarcasm scores and the performance in the facial expression and affective prosody tasks, we averaged the two variables into one variable that reflected performance in affective processing. For the entire sample, comprehension of sarcasm correlated moderately but significantly with the Faux Pas scores (r =.247, p = .039) and affective processing (r = .307, p = .011). No significant correlations were found when the different emotions (i.e., anger, happiness) were analyzed separately.

Your host hesitates to embrace the underlying suggestion of the paper, namely the idea that people who don't get sarcasm must somehow be brain-damaged. Not all clueless people actually have brain lesions (or at least, no study exists to prove that they do). An autism activist quoted at the end of the piece seems concerned that people will read about this and think of autistic people as brain-damaged -- which seems like an odd complaint given the symptoms of autism, although perhaps their concern is merely to reject further stigma. But the paper seems pretty sound on its basic finding, which is that there are different brain regions  involved in recognizing propositional and emotional content, then integrating the two.

If you could care less, and therefore presumably want to read more,  the Neuropsychology article can be found here. If you couldn't care less, stop here. If you're undecided about whether or not those two phrases mean the same thing, read Mark Liberman's thoughts on the matter. Or don't.

Lime mondegreen

Your host has written about mondegreens before, but today he wants to talk about a slight variant -- an attempt to get people to think they're hearing something different.

A recent Coca-Cola commercial features a guy in a lab coat staring at a can of Coke and a lime. He's got something of an Alfred E. Newman look on his face while trying to figure out what to do with them. These visuals accompany a soundtrack of a reggae song, which sounds to your host's ears like it goes "Put the lime in the coconut, and drink them both together". But on the screen, subtitles claim that the words are "put the lime in the Coke, you nut". So at first SC thought it was a misheard lyric just like the original "Lady Mondegreen" (actually, "laid him on the green").

But after seeing the commercial a few times last night, your host decided that it was absolutely impossible that the words he was hearing could be phonologically parsed as "in the Coke, you nut". No matter how many times he listened, he could only hear "in the coconut". (Read the next two words in the style of Richard Dawson) Google says? It's called Coconut, it's written by a singer named Harry Nilsson, and the line is, in fact, "Put the lime in the coconut". This fan site dedicated to Mr. Nilsson suggests that the lyrics were changed for the commercial. On the other hand, it also indicates that the musical performance is actually by Harry Nilsson, who died in 1994, and could not have rerecorded the song with altered lyrics. Adweek's citation of the commercial also supports the idea that the music is the original, and that only the subtitle introduces the "you nut" material. So it would seem that Coca-Cola's marketing gurus are trying to pull a reverse mondegreen, making you think that you hear something in the song that you probably don't.

Coke wins the Pepsi challenge

Here's a story that probably was front-page news in Atlanta a few months ago, but which SC only found out about from next month's Stereophile (as often is the case with such things, it arrived halfway through the "previous" month). The audio hook is completely self-serving and unobvious; more interesting is the actual science, in particular for what it says about the ability of certain words to take on a very real power over your thinking.

The article by Samuel McClure and colleagues, which originally appeared in the October 14 issue of Neuron, is titled "Neural Correlates of Behavioral Preference for Culturally Familiar Drinks". Despite the imposing name, it's basically the Pepsi Challenge carried out (in part) inside an fMRI machine.

The methodology is quite interesting, if only because of the difficulties involved with getting people to drink inside an fMRI tube without drowning them. This was accomplished like so (although we're getting ahead of ourselves):

Individual squirts of Coke and Pepsi (0.8 mL each) were delivered to subjects through cooled plastic tubes held in the subjects’ mouths with plastic mouthpieces. The volume of soda delivered on each squirt was sufficient to allow the subjects to fully taste the soda but were small enough to allow them to easily swallow while lying in the scanner. A computer-controlled syringe pump (Harvard Apparatus, Holliston, MA) allowed for precise delivery of the colas.

Of course, it's not merely enough to know what's going on inside people's brains when they drink a soda; you have to know whether or not they have an opinion about what they're drinking, and whether or not they can actually tell the difference. The former tells you if they think drinking the soda is a pleasurable experience, the latter if they actually can tell, or if the observed reaction is due to something other than the chemical composition of the soda.

Unsurprisingly for anyone who thinks that all colas are just "brown, carbonated sugar water" (McClure et al.'s description), there was no statistically significant correlation between stated and "behavioral" preferences -- in other words, when they actually did carry out the Pepsi Challenge, they found that people who said they liked Coke better couldn't actually tell between unlabeled drinks. It's not really true that this is the case for everyone; McClure et al. don't mention whether or not they threw out subjects who could actually discriminate above chance, or if they even had any in their sample. SC knows such people exist, though -- he is one. The point here is to measure the suggestive power of branding, and so it's probably most helpful to stick to that part of the population (a rather large majority) that really can't tell the difference.

An interesting variation of the Pepsi Challenge (not carried out inside the fMRI machine) involved having subjects drink from one labeled cup and one unlabeled cup. When the label was "Coke", subjects showed a strong preference for the labeled cup, but when the label was "Pepsi", the distribution of preferences was notably more random. They don't mention whether or not people who expressed preferences for one label or the other actually gravitated towards their stated choices when available; it would be rather strange if they didn't, unless they suspected that the labels might be deceptive. In fact, they were not; in the "semianonymous" trials, both cups contained the same soda.

But the really interesting result came from the fMRI tests -- when subjects were given Coke through the tube/syringe contraption mentioned above, they showed meaningful brain activation in 6 areas, when told it was Coke. No such reaction occurred with Pepsi. No deception was involved in this experiment -- it was a straightforward test of labeled vs. unlabled Coke, and labeled vs. unlabeled Pepsi.

Observing that significant cultural meaning is attached to both Coke and Pepsi, McClure et al. conclude that it's this specific information which acts to bias people's preferences, since it seems able to override preferences in the absence of any such information. From a linguistic standpoint, it's hard to argue with that conclusion -- clearly, the label "Coke" has power over people's minds in ways that "Pepsi" does not -- but it's not clear that it's the whole story, either. Admittedly, Coke is associated with a long history of cultural significance, and much memorable marketing, but they hardly have a monopoly in this regard. The word "enjoy" might trigger strong Coke associations, but SC would bet any reader who remembers the '80s and '90s can easily recall which soda is associated with the lines "The choice of a new generation" or "You got the right one, baby (uh-huh!)". It simply doesn't make sense to SC that the Coke label should be the beneficiary of cultural knowledge while Pepsi is not -- they've both been out there advertising heavily for a long time. Your host isn't familiar with psychological literature on advertising, but the necessary follow-up question is: "To what extent do the various slogans and jingles actually imprint themselves as associations with lexical items?". If Coke advertising historically has stuck with people, while Pepsi has not, this would be strong confirmation of McClure et al.'s conclusion, which otherwise remains rather speculative.

As for the audio conclusion mentioned at the outset? Well, nobody has ever successfully demonstrated an ability to distinguish between well-designed amplifiers in double-blind tests (there are caveats here about distortion levels and output impedance, but they're not relevant for this purpose). Suffice it to say that the Stereophile crowd is well aware that their editorial claims don't generally hold up under double-blind -- or even single-blind -- scrutiny. So when they brought up this study, it was to suggest that brand awareness produces increased listening enjoyment. This means it's critical to know that you're sitting in front of a pair of $350,000 Wavac amplifiers instead of a $200 Pioneer receiver. Such a claim could never be tested, because an MRI machine provides a really awful acoustic environment. But it does provide an amazingly useful justification for the continued existence of magazines which exist to create exactly the sort of brand awareness without which that receiver isn't worth much less than the super-amp.

Happy something or other

SC is traveling today, and won't be able to post anything lengthy for a day or two, but since Hanukkah is in full swing, might as well retell a favorite story of Dad SC's with a psycholinguistic point.

Readers may be aware that there are multiple widely accepted transcriptions for the name of this holiday. Among them are Hanukkah, Chanukkah, and single-k variants of both. Interestingly, Google suggests correcting Chanukkah to Chanukah, but not Hanukkah to Hanukah.

Keeping this in mind, consider the following exchange between Dad SC and a lawyer colleague some 20 years ago (when perhaps awareness of the holiday was lower than it is now):

Dad SC: Merry Christmas!

Colleague: And a happy Chaka Khan to you, too!

People worship all sorts of strange things, but SC rather doubts that his father's colleague really thought that Dad SC celebrated the existence of a then-popular R&B singer (for a much better reference, see VH-1's artist page). It seems more likely that Chaka Khan was actually a higher-frequency item in his vocabulary at the time, and sounded or looked similar enough to the word he was looking for, that it was simply used instead when the original word couldn't be found in his mental lexicon. This is a very natural outcome under some models of lexical retrieval/derivation; for one which would make the case for both sound and orthography, have a look at this paper by Mark Seidenberg and Laura Gonnerman.

Rats and Cobras

Courtesy of the Drudge Report, a story purporting to demonstrate media bias by showing video of George Bush in front of just part of a sign containing the word "families". Seen in a still photo, it looks like it's been framed to say "I lie". The story is fairly ridiculous, for the same reasons that taking a few frames out of a commercial with the word "bureaucrats" in the background didn't really demonstrate that Al Gore was being accused of consorting with rats.

One good debunking of subliminal messages comes from the urban legend gurus at Snopes.com. While it deals with the falsity of the experiments that purported to demonstrate subliminal effects, it doesn't actually explain why these messages are unlikely to work. SC's certainly got some things to say on the subject.

While the Snopes article claims that the original experiments displayed the messages for 1/3000th of a second, this is almost certainly not correct. Film is projected at 24 frames per second. It wasn't always like that; early films were run at 16 frames per second, and as much as 48 frames per second could be simulated with appropriate manipulation of the shutter. At no time in the history of film could an image have been displayed for as little as 1/3000th of a second -- and it's still not possible with modern video, which ups the frame rate to 29.97 frames per second in the U.S. (25 fps in Europe, which uses a different standard). A more plausible technical claim is advanced in the 1985 G.I. Joe episode, "Flint's Vacation", in which the second-baddest dude in pop culture history is inserted into every 20th frame, which is at least technically credible. As is clear from watching the episode, but not so clear from the script, Cobra Commander's voice is similarly spliced into the audio of the "news", and it's only by watching extended clips that the residents of Pleasant Cove receive their instructions.

Of course, the real problem with these claims isn't really that their proponents are unsophisticated about video projection. For the most part, word recognition just doesn't happen on the scales being discussed. One good summary of literature (through '94) on spoken word recognition is Gareth Gaskell's thesis, which includes measurements of 500-700 milliseconds for recognizing single words (although it's important to note that this is total reaction time; the actual time to hear and recognize the word is shorter, but it's still not on the order of the 30 milliseconds that a single video frame gets). Other papers mention throwing out measurements of less than 200 milliseconds for response times. SC could go dig up a whole bibliography of measurements from the course in psycholinguistics he took as a graduate student, but the point is merely to demonstrate that word recognition takes place on a much longer time scale than video frame refreshes. Assuming an audio track synchronized to the video, it is more or less impossible for a signal to be reconstructed in the brain with snippets that are far too short being played back about 0.8 seconds apart.

Of course, there's no claim of subliminal audio in the political stories here. But the numbers work out similarly for visual word recognition, and the same principle applies -- seeing something for one or two frames isn't going to be enough to recognize a word. Even in the cases where these things are on screen for a full second -- and "rats" wasn't, although SC can't be sure how long "ilie" was -- the stimulus is competing with a far more perceptually salient signal, which is to say the actual news presentation or the surface content of the commercial. It simply defies everything we know about word recognition to take seriously the idea that subliminal messages are meaningfully perverting an otherwise pristine political process ([that's OK, we've got plenty of other ways to pervert it -- ed.]).

(Edited on 10/6/04 at 1:59 a.m. to include note about response vs. recognition; 2:25 a.m. to correct math error.)

Mechanical ineptitude

Sometimes, your host does things which cause him to be so embarrassed that he doesn't know what to do short of posting about it and letting his readers join in the fun. This is one of those moments.

Today happened to be the birthday of Maternal Grandmother SC (henceforth, MGSC), and so to celebrate, Mom SC brought her up to a regional mall not far from Chez SC to spend the day, concluding with dinner with family. A sensible plan for a pleasant day.

While a gentleman never asks the age of a lady, nor discloses it if he happens to know anyway, suffice it to say that walking is no longer foremost among MGSC's talents. In order to facilitate a day of traipsing around malls, Mom SC therefore decided to surprise MGSC with a device she mistakenly believed to be a wheelchair. In fact, it is something called a "rollator", or a wheeled walker. A product similar to the one actually bought can be seen here; SC is unable to provide a more reliable picture for the reader's edification because the tag identified the walker as a product of "Dr. K.", a company which apparently has no Internet presence, nor is commonly sold online. However, the picture is similar enough to give the reader an excellent idea of what followed.

Although the rollator has four wheels, only two of them are actually able to turn for the purpose of changing directions. They happen to be the two wheels located under what looks like, from the perspective of someone who thinks it's a wheelchair, a backrest for the seat. Mom SC reported that MGSC had great difficulty using the wheelchair; would SC mind taking a look?

After inspecting the device, your host "discovered" that in order to successfully change directions, it helped to back up while turning in the opposite direction of the one you want to go forward in. Much like driving in reverse, really. A few minutes of practice sufficed to give SC reasonably impressive skill in "driving" with the rollator. MGSC found the whole thing rather counterintuitive, though, and is likely to return it.

When your host returned home, he discussed it with Mrs. SC (who is presently out of town, and who otherwise would have prevented this from happening). After SC explained to her the difficulty of a wheelchair that had to go backwards to turn, she started to suspect that the problem wasn't the chair. Why, after all, weren't the steering wheels mounted in front? It all came together when your host explained that mounting the chair in a way that would put those wheels in front would require lifting the rider's legs notably higher than most elderly people can easily tolerate.

That's when Mrs. SC realized that your host wasn't describing a wheelchair, he was describing a walker! Under no circumstances should anyone try to ride one of these devices; instead, the "backrest" is actually a handlebar, and the seat is meant as a convenience for use when the user gets tired. Why, then, is it equipped with handbrakes? Not to stop someone from riding too quickly, but to stop it when they want to walk around the side and sit down. Needless to say, Mrs. SC had quite a good laugh over SC's total inability to figure out the proper use of the rollator.

There is a linguistic lesson in this, or at least a psychological one. As with a Rorshach test, given the outline of a novel object, your host imposed on it the label that most closely corresponded with similar objects in his previous experience. Once the rollator had been characterized as a "wheelchair", he similarly assigned names to the parts which followed naturally from a belief that the object was a chair, and not a walker. Thus, the handlebar became a chair back, and the handles above the brakes became armrests. Only, the back wasn't a back, and the armrests weren't armrests. But SC had such strong associations between the form of what he was seeing and other wheelchairs, that he simply could not see a walker until someone who knew better told him so.

Here comes the calvary!

Readers feeling deprived of their daily dose of SC on Thursday may rest assured that your host spent much of the day in synagogue for Rosh Hashanah. Actually, that's technically untrue. He spent the day in a church, albeit for the aforementioned holiday.

This is not all that uncommon an occurrence. Orange County, CA, is not exactly home to an enormous Jewish population. It's big enough to have a decent number of synagogues, but many of them are constructed according to the principle expressed in the following joke:

Q: Two Jews are stuck on a desert island. How many synagogues do they build?

A: Three. One that each one goes to, and another one that nobody goes to.

[From The Big Book of Jewish Humor]

As a bonus humorous aside, Yahoo!'s listings for temples/synagogues in the Orange County city of Costa Mesa erroneously include an upscale jeweler, Temple St. Clair, located in a big regional mall. Oops.

Back to the church, though. Because many synagogues are constructed to handle the day-to-day demands of a few hundred people at most, they often simply don't have the physical resources needed to accomodate the large crowds that show up for the High Holidays. For this reason, it's not uncommon for services on these days to be held at other venues. In the case of the synagogue to which SC belongs, that's a nearby church called Calvary which has generously shared its facilities for a number of years now.

The name Calvary is attached to many churches; it's hard to say how many just by Googling the name, but your host has seen it in enough contexts to be aware of that fact. It's not a name with any theological significance to Jews, though, so I felt obliged to look it up in preparing this post. A Catholic encyclopedia that your host has consulted on previous occasions indicatesthat it refers to a place where Jesus may have been crucified; the discussion is nuanced and covers plenty of material that SC won't even pretend to be qualified to comment on. Suffice it to say that the word is of wholly distinct etymology from "cavalry", a term referring to either mounted or mechanized infantry, of the sort commanded by General Custer.

The salience of this point comes from the utter and total inability of SC's fellow congregants to say "Calvary" today, substituting "cavalry" at every available opportunity. This struck SC as another good opportunity to talk about exchange errors in speech.

It's a bit tough to be sure of what kind of error we're talking about here, as there are two varieties of disfluency that could be involved. On the one hand, it could be a word-internal exchange of phonological material, with two sounds slipping positions. Or it could be a lexical exchange where people with both words in their vocabulary simply are having trouble recalling the right one (especially since neither of them are likely to be high-frequency items among this crowd). Examples of these sorts of errors, not to mention a useful example of the experimental protocols for studying them, can be found in this paper by Rachel Walker and a couple of her students. A nice background presentation on speech errors also can be found through this (Google-cached) handout by Bruce Tesar (SC only found this on the web by searching for more examples, although sometime he'll have to talk about Prof. Tesar's book on learnability, which he absolutely loved). While we're at it with course materials found through Google, check out the discussion of Freudian slips in this set of lecture notes by Mark Liberman and Bill Poser's colleague Gene Buckley. Prof. Buckley appears to have the sense of humor that ought to be on display in Language Log

It's hard to imagine an experiment that would conclusively demonstrate whether the Calvary/cavalry error is lexical or phonological. One way to eliminate the lexical hypothesis might be to have groups of Jewish and Christian test subjects read aloud sentences containing the words, and see if there's any significant difference in the error rates for the two words. Of course, such a study would be filled with all sorts of background problems, most notably controlling for the actual world knowledge of the Jewish test subjects. And if the error rates were similar, you wouldn't have any new knowledge about where in the speech production process things were going wrong. The services for the second day of Rosh Hashanah will be held in the synagogue (attendance figures for day 2 are usually lower), so that means he's got until next Friday night (read: Yom Kippur) to come up with an experimental protocol to use on his fellow congregants.

But you can't fool Mom

Courtesy of Arts and Letters Daily, an interesting article from the Economist on lie detection.

The "standard", and SC uses the term loosely, has long been the polygraph, which measures "breathing rate, pulse, blood pressure and perspiration", all of which presumably go up when you're lying. Unless they don't. The results are certainly quite variable by person -- all of those variables increase for your host every time Mrs. SC grills him on the disappearance of chocolates from the pantry, and he doesn't even have to lie about it. For that reason, it's not exactly the gold standard in admissible evidence.

No better, and adequately deconstructed by Mark Liberman is "voice analysis", a technique purporting to detect emotional stress in individual speech (which in turn is supposed to be indicative of lying).

Since these systems haven't proved very useful, current research is focusing on detecting lies by measuring brain states. The Economist article discusses three such methods, none of which strike SC as all that convincing. Essentially, the goal of the various research methods is to demonstrate increased brain activity in regions that correlate with lying, or lack of activity in places that would be expected if the necessary mental representations were being activated. That latter approach has apparently been successfully demonstrated by showing that a brainwave called P300 doesn't trigger when a person is asked about something they don't know any of the details of.

The problem that your host sees with this approach to lie detection is threefold: 1) brain regions don't activate nearly so consistently as one might hope when measured across different people, 2) assuming that brain states correlate with stress isn't much better than assuming that blood pressure or pulse correlates with it, and 3) it's not at all clear that clarity of recollection (strength of mental representation?) has much to do with beliefs about truth. I might be perfectly certain that I've visited the Louvre, based on a dream plus the facts (not true, but we'll stipulate them for the moment) that I've studied the paintings there, memorized a map of the museum, and have seen enough photographs of the interior to be able to describe the layout of the rooms in detail. Nevertheless, I've never actually been to Paris (except for the airport, on the way to Spain). Someone asking me questions about the Mona Lisa might conclude I was telling the truth if I said I'd been there, based on my ability to answer questions in detail, which presumably would correlate with brain activity showing that I've got Louvre representations up the wazoo ([or the prefrontal cortex -- ed.]).

This sort of objection isn't all that different from philosophical "brain-in-a-vat" arguments; essentially, your host contends that we can't be sure that brain states derive from the stories which happen to be their most plausible explanations. We might think that the odds are fairly low that the brain states occurred as a result of some other cause, but SC would love to be the lawyer who gets to ask a jury if they want to make that bet in sending someone to the chair.

Lie detection based on physiological variables strikes your host as largely a dead end. Even supposing that some factor could be shown to correlate pretty well with intent to lie, there'd still be a separate question of whether or not the statements are actually true, not just whether or not the teller believes them to be true. If I tell Mrs. SC "yes, I ate the chocolates, all the chocolates", but she knew that there hadn't actually been any in the pantry to begin with, then she might decide that I'm not so much a liar as confused. And then it's pretty much impossible to decide how much credibility to assign to anything else I've said. Of course, I could be lying about that.

Watch this

Another amusing anecdote from SC's trip actually predates his adventures in Fair Park. On the way to Dallas, your host flew through Las Vegas ([where he also dropped $15 on the way back, low-roller that he is -- ed.]); during the course of his time in the airport, he was introduced to the wares of a company called OurVersions. Like many companies in the business of hawking trinkets to distracted and bored tourists, OurVersions specializes in the gimmick of offering what are purported to be reliable imitations of products selling for more from famous brand names. However, they offer a unique twist on this game which relies on a psycholinguistic trick.

When selling imitations of perfumes, the company is on safe ground saying things like "Inspired by Carolina Herrera" or "Inspired by Poison", because the line "inspired by" gets them off the hook for any sort of trademark violation. The alternate names provided are designed to sound like the names of the well-known products in question, but there's clearly enough differentiation to survive any sort of legal challenge.

Fake watches, however, are a different story. No small part of the prestige of a watch comes from the fact that the name is often (not to say always) plainly visible somewhere on the face or band. The Rolex crown commands a considerable premium, as does the Ulysse Nardin anchor, the Omega Greek letter, or the TAG Heuer shield. Reproduction of these names or their accompanying logos would be, to put it mildly, illegal. (Here's an interesting guide to spotting fake Rolexes; apparently, fakes have their own hobbyist culture, with elaborate grading schemes and things that only a connoisseur would know to look for.)

In any event, the OurVersions folks aren't looking to break any laws, but they still need a way to make their customers think they're putting a fast one over on more expensive watch brands (and their retailers). So they advise their customers to compare their watches to "famous name designers" like "Gianvianca", "Nicollé", "Courtier", "Artifacts", "M. Röthchild", "Ricci" and a few others SC didn't bother to write down.

The joke is, not one of these is the name of any company which could reasonably be construed as manufacturing a famous watch brand. However, "Courtier" sounds like "Cartier", "Artifacts" might be reminiscent of "Fossil", and "Ricci" could lure the unsuspecting into mistaking it for either "Gucci" or cosmetics name "Nina Ricci" (which is licensed out for real designer watches).

So what OurVersions' display racks are relying on is the possibility that, given a relatively inattentive observer, spelling or phonetic cues will fool them into thinking that a fake designer name is actually a real one. Psycholinguists (not "psycho linguists") have something to say about this; in a model like this one by Mark Seidenberg and Laura Gonnerman (see also Prof. Seidenberg's extensive publications list, for plenty of further elaboration), the fact that both spelling and pronunciation cues are presented by these fake names ought to be enough to at least stimulate the thought of real designer names. This isn't to say that all psycholinguists would agree that this trick should work; the influential (albeit rather outdated) Cohort and TRACE models would likely predict that the differences are great enough that almost nobody would bite it, especially in cases where there's no sound similarity (like "Artifacts" vs. "Fossil"). Readers who may not have much familiarity with psycholinguistics, but recognize the Steven Pinker brand name in popular psychology might wish to get ahold of this position paper that he wrote for Science; suffice it to say that Pinker probably also would dismiss the possibility that the OurVersions people are onto something. (For the record, his examples all deal with verb morphology, but the model he proposes, and still defends today, denies that bundles of phonetic/orthographic/meaning cues could retrieve words only vaguely related to what's heard.)

SC couldn't help but feel that he had missed an opportunity when he saw this one. In order to drive site traffic, your host should have come up with a name that advertises how he might be compared to Famous Brand Name language bloggers. So maybe this site will reopen shop someday as Lingualog or Conversation Chapeaux.

Can't keep names straight?

That would put you in good company (at least in SC's mind), specifically with San Diego Padres radio broadcasters Ted Leitner and Jerry Coleman. Now, Coleman is something of a legend among baseball fans for his relentlessly malapropism-laden broadcasting (also for being a decorated war hero who flew in both WWII and Korea, and giving up what might have been a Hall of Fame career with the Yankees to do so). This evening, both gentlemen made calls suggestive of a little difficulty with lexical access.

One situation: Padres third baseman Sean Burroughs is at bat; shortstop Khalil Greene is on second base. Burroughs hits a double, sending Greene home, and Leitner calls it: "Sean Greene will score easily on the stand-up double by Burroughs." Now, there is a player named Shawn Green, a starting outfielder for the L.A. Dodgers (boo! hiss!). One might write this off as a random speech error, but not much earlier, with Giants pitcher Dustin Hermanson on the mound, facing Padres catcher Ramon Hernandez, Coleman made the following call, "Here's the pitch...and it's a grounder to third; the toss over to first, and Ramon Hermanson is out to end the inning".

In both cases, the speech was quite fluid and natural; while Ted Leitner corrected himself shortly after the mistake above, there was no sign that he noticed anything was amiss while uttering the original sentence. Jerry Coleman, on the other hand, doesn't even bother correcting himself on such mistakes anymore.

I'm not sure what name goes to speech errors like this (or if there even is a name for it aside from "word exchange"), although they're commonly studied in psycholinguistics. They're not spoonerisms, because it's the exchange of whole words, not segments or syllables. Your host things that this is not quite an eggcorn, either, because in at least some cases, the production is itself a valid lexical item (i.e., Sean Burroughs + Khalil Greene = Shawn Green, which is a name a baseball broadcaster would know and use in that context). There's no Ramon Hermanson in the majors, but that production isn't the result of a mishearing, like eggcorns are. If there isn't already a name for this, maybe it's time someone relieves the world of that sad oversight.

Magic number theory

Mark Liberman writes today about the "Resnikoff-Dolby 30:1 Rule", which basically says that 30:1 is the ratio that we like our compression at: 1 guide card to every 30 index cards in a card catalog, 30 index entries for each entry in a table of contents, etc. The pattern repeats in enough places that Prof. Liberman says:

I'm prepared to believe that Resnikoff and Dolby are on to something. The main thing that makes me skeptical is precisely that I haven't heard of this idea before -- and that's a sad sort of argument.

SC has never heard of this idea before, but there's another one just like it that he's surprised Prof. Liberman didn't think of. Your host refers, of course, to the greatest paper ever in the history of psychology, George Miller's "The Magical Number Seven, Plus or Minus Two".

Prof. Miller adduces plenty of examples from experiments on color recognition, pure tone recognition, face recognition, degree of saltiness, and other areas to suggest that we could juggle about 7 distinct categories at once. This occurs across so many apparently unrelated phenomena that it seems to be indicative of something fundamental. But as Prof. Miller ended the paper:

And finally, what about the magical number seven? What about the seven wonders of the world, the seven seas, the seven deadly sins, the seven daughters of Atlas in the Pleiades, the seven ages of man, the seven levels of hell, the seven primary colors, the seven notes of the musical scale, and the seven days of the week? What about the seven-point rating scale, the seven categories for absolute judgment, the seven objects in the span of attention, and the seven digits in the span of immediate memory? For the present I propose to withhold judgment. Perhaps there is something deep and profound behind all these sevens, something just calling out for us to discover it. But I suspect that it is only a pernicious, Pythagorean coincidence.

As Freud might have said if he had really been a psychologist, sometimes a coincidence is just a coincidence.

Hearing in color

Interesting story today from ABC News on a phenomenon that has interested SC for years, synesthesia.

As SC has always understood it, the claim of synesthetes is that perceptions in one sense trigger perceptions in another sense, even though a non-synesthete wouldn't notice any stimulus of the second sense. So a synesthete listening to a saxophone might perceive the sound as literally being blue in a way which is, at best, only metaphorically true for other people.

But that's not quite what seems to be going on in the experiments described in this article. In an effort to show that words had distinct colors, researchers showed a subject a list of 100 monosyllabic words and asked him to name the colors of each one. Repeating the task a month later yielded enough similarity that the effect was judged significant, and therefore psychologically real.

However, all of the text was printed in black. SC wonders what would happen if the words were printed in different colors. That is, if the word "fish" was printed in red, would the synesthete report that the color associated with it was whatever he had reported before? Or would he say "red" (assuming that's not the one that he had already associated with it)? And if he said "red", would this demonstrate that there was an association between the meaning of the word and the color, the shape of the letters and the color, or would this just trash our understanding of synesthesia altogether? The researchers involved think that the perception of color for words is a glitch in visual processing; if the words were in a specific color other than black, and that color was what people reported seeing, SC thinks this would require assigning the glitch to some other place in the brain.

Classification is a tricky thing

Radagast is, to put it mildly, not a fan of The Truth Laid Bear's "blog ecology". In its place, he proposes a couple of schemes for categorizing blogs which are grounded in biological science (full disclosure: SC had a sneak peek at an early draft). In fairness to TTLB, it's not clear that the rankings are meant quite as seriously as Radagast's response, but the resulting discussion brings up an interesting linguistic point.

Radagast is bothered by the categories that TTLB created because they don't match his conception of the world. However, it's not so much that the general approach to classification (something based in biology, his specialty) bothers him, but the understanding of what concepts are important to biological classification. George Lakoff addressed this issue (among many others) in a book which heavily influenced SC as an undergrad, "Women, Fire, and Dangerous Things". The central theme (that's a terrible pun, if you've read the book) is an argument that humans are prone to perceiving things in ways which are friendly to their categorization schemes, and more importantly, that these categorization schemes are organized radially. Rather than the neatly hierarchical class-subclass relationships of classical logic and computer science, Lakoff claims that we organize things into prototypical examples of categories, and then grade things as being more or less like them. So a sparrow is a good example of a bird, because to our minds, it possesses many of the qualities that make something "bird-like" (feathers, wings, it actually flies, has short legs in proportion to the rest of its body, etc.). An ostrich, on the other hand, is a terrible example of a bird (the legs are enormous, it doesn't fly). It's not that an ostrich isn't a bird, it's just not the first thing that pops into mind when we think "bird". Lakoff even provides references to studies demonstrating that exact point.

Categorization schemes are rather dependent on your cultural upbringing. Lakoff contrasts our understanding of counting with people (SC forgets from where) who split their number markers into one, two, and many. But some categories are grounded in biological and psychological facts that are independent of culture -- for example, languages with small vocabularies for describing colors will tend to pick color words that maximize the differences between the meanings of the names. Black and white are always the first named color. If there's a third named color, it's always red. As the vocabulary expands, the categories start to overlap more.

Bringing this back to the issue of blog taxonomies, Radagast and the Bear have rather different conceptual schemes for organizing the world. Deciding which of the approaches to favor comes down to a choice of which of the schemes more intuitively fits our own models of the world (which is why the "ladder" model is so appealing, even though it may not be scientifically accurate). Or we might choose to try to readjust our conceptual schemes to accomodate the biological view.

Your host is ultimately inclined to take Lakoff's work metaphorically rather than literally, a point which even George Lakoff would probably smile about (given his intense focus on metaphors). While there are a lot of sound intuitions, and some solid experimental results, behind Lakoff's conceptual schemes, they don't offer much for those seeking to develop a formal theory of syntax or semantics (which Lakoff was a lot more concerned with in the '70s and early '80s). Logicians who have engaged Lakoff's approach feel that a lot of his arguments against traditional categorization and logic reflect an insufficient appreciation of just how well the classical machinery of logic can be levered to reflect his results. Lakoff is also a committed philosophical materialist and reductionist, and attempts to use this model to argue for those views, which your host neither shares nor feels are necessary to accept the points about how we categorize things. Finally, Lakoff cannot resist the temptation to inject his politics into his science (which Language Log has noted in two posts; start here). This last point is sensitive, and SC will refrain from specific comment on it, except to say that he thinks that the work which Lakoff has produced in applying this approach to politics is more reflective of Lakoff's own conceptual scheme than it is a useful exposition of the views of either American political party (or of the way in which they create and disseminate said views). Having included all these caveats, SC feels that Lakoff's ideas on the organization of the mind are important ones to deal with, and provide a very useful framework for understanding how people can look at the same situation and draw such different conclusions.