The Unz Review - Mobile
A Collection of Interesting, Important, and Controversial Perspectives Largely Excluded from the American Mainstream Media
Email This Page to Someone

 Remember My Information



=>
 TeasersiSteve Blog
/
Flynn Effect

Bookmark Toggle AllToCAdd to LibraryRemove from Library • BShow CommentNext New CommentNext New Reply
🔊 Listen RSS

Monsieur Hulot ponders office cubicles of the future

I finally got around to watching a couple of movies by the great French comedian / director Jacques Tati, 1959′s Mon Oncle and 1967′s Play Time.

Home sweet home for M. Hulot

Tati, a successor to Chaplin and Keaton, made post-silent comedies without much plot or dialogue but with a lot of sound effects and visual gags.

Tati liked the eccentric, traditional Paris and aimed his satire at the coming wave of rationalized, Americanized modernism.

These films star his standard character, a tall, amiable pipe-smoking duffer named Monsieur Hulot who grapples, not particularly effectually, with the modern world.

Monsieur Hulot is a man well suited for life in the picturesque and extremely French working class neighborhoods of Paris.

His nine-year-old nephew, who lives in an expensive all-white modernist mansion where he is bored silly, is entranced by his opportunities to visit his uncle in his downscale French world of motorbikes, horse-drawn carts, and dubiously added-on apartment buildings.

Here’s a photo from 1966 of my father and I trying to recreate the famous poster shot from Mon Oncle on my dad’s new Honda 90:

For decades, this photo and the uncharacteristically carefree expression on my worrywart dad’s face were vaguely associated in my mind with the words “French comedian.” So I’m very happy to have finally watched the movie (even though, it turns out, we were going the wrong direction).

But Monsieur Hulot is more than a little lost when he must venture into modern International Style districts.

In fact, Play Time is set in a steel and glass skyscraper district of Paris that didn’t exist yet.

Tati went broke building a set that looked like Sixth Avenue in New York City.

One running joke are the travel posters inviting you to visit destinations such as London, Hawaii, Mexico, and Stockholm by showing the same International Style skyscraper in each:

Watching Mon Oncle and Play Time got me thinking about the Flynn Effect of rising raw scores on IQ tests.

The Flynn Effect is one of the more unexpected and interesting social science discoveries of the later 20th Century.

The Flynn Effect has been strongest on IQ subtests that are least affected by local cultures and that most resemble having to deal with electronic machine logic.

The futuristic Raven’s Matrices tend to be less culturally biased than other IQ tests but has had a very large Flynn Effect of about 3 points per decade, or a standard deviation per half-century.

I don’t think it’s a coincidence that the first American IQ test, Louis Terman’s Stanford-Binet, emerged in what’s now Silicon Valley and that Louis’s son Fred Terman, Stanford’s Dean of Engineering, has perhaps the best claim to be the father of Silicon Valley. My general hunch is that one cause of the Flynn Effect is that early IQ test designers had a pretty good sense of the direction in which the world would be moving: away from tacit, locally idiosyncratic patterns of behavior and toward machine logic ways of thinking, a shift in which Silicon Valley has led the world.

Much of the joke of the Monsieur Hulot movies is that Monsieur Hulot is a pre-Modernist man. He is accustomed to the Paris of the first half of the 20th Century, which in Tati’s view is about as good as human life gets. But, despite, being a curious fellow open to new things, Monsieur Hulot can’t seem to become accustomed to the global culture of the second half of the 20th Century.

In this scene from Play Time, Monsieur Hulot has ventured to a new skyscraper to take care of some business. A 75-year-old messenger boy tries to notify the man Hulot has come to see via a new message machine that’s like a Raven’s Coloured Progressive Matrices IQ test:

In Mon Oncle (1959), Monsieur Hulot investigates his wealthy sister’s ultra-modern kitchen (my apologies for the sound on this clip being slightly out of sync — Tati movies need the sound effects to be perfectly synced with the action):

Tati was a rich kid who was a gentleman rugby player before he got into comedy. He started out in show biz doing impressions of rugby players, so I imagine that’s what he’s doing bouncing the plastic coffee pot.

And here’s a link to a classic scene in Mon Oncle in which Monsieur Hulot’s upwardly-mobile in-laws try out their new automated garage door with the electric eye trigger that their dachshund can’t open for them when they get trapped in the garage because he feels guilty, although he doesn’t know why, and thus isn’t wagging his tail.

This is the saddest scene ever.

And then they call their maid to set off the electric eye, but she’s afraid of electricity.

 
• Tags: Flynn Effect, IQ, Movies 
🔊 Listen RSS

The Flynn Effect of rising raw scores on IQ tests is one of the most interesting phenomena in all the human sciences. It was first noticed in the 1940s, but for a long time little attention was paid to the fact that IQ test publishers had to renorm their tests periodically because people kept doing better on them. This pattern began to be explored by political philosopher James Flynn from around 1979 onward, and the phrase “Flynn Effect” was coined in his honor in 1994′s The Bell Curve.

One interesting aspect of the Flynn Effect is that it tends to be larger on the less culturally biased tests, such as the outer space-looking Raven’s Progressive Matrices:

Historically, much effort was put into the obvious challenge of developing IQ tests that are stable across space, from culture to culture. In contrast, nobody until Flynn paid all that much attention to the question of IQ tests being stable across time.

For example, the alien-looking Raven’s Matrices IQ test that was introduced in the 1930s in the hope of being more culture-free than previous IQ tests has seen a huge Flynn Effect of around 3 points per decade, or a standard deviation (15 points) in a half century. A score on the Raven’s that would put you at the 50th percentile a half century ago would only put you at the 16th percentile today.

The more human-seeming Wechsler Intelligence Scale for Children (WISC) saw a still-substantial Flynn Effect of about two points per decade, but that’s less than the Raven’s.

Screenshot 2015-10-17 00.15.05

Importantly, the size of the Flynn Effect from 1947-2002 differed sharply amongst the subtests on the WISC as shown above, from only 2 points over the 55 years on the “Information” and “Arithmetic” subtests to 22 points on “Picture Arrangement” and 24 points on “Similarities.” (In the table above, the Flynn Effect column is taken from my 2007 review in VDARE of Flynn’s book What Is Intelligence? )

The kind of cognitive facilities that come up in normal conversation, such as vocabulary, arithmetic and general knowledge, have only seen small Flynn Effects, which is why the Flynn Effect isn’t easily noticeable in much of daily life (although I’ll point out below where it can be seen).

Recently, James Thompson’s Psychological Comments had a table of the “cultural load” of each WISC subtest from a 2013 paper:

Kees-Jan Kan, Jelte M. Wicherts, Conor V. Dolan, and Han L. J. van der Maas. “On the Nature and Nurture of Intelligence and Specific Cognitive Abilities: The More Heritable, the More Culture Dependent.” Psychological Science 24(12) 2420–2428

… Cultural load was operationalized as the average proportion of items that were adjusted in each subtest of the WAIS-III when the scale was adapted for use in 13 countries.

I presume that means adjustments in questions beyond simple translation. IQ test publishers validate new editions of their tests in each country in which they intend to sell them, and that lets them notice proposed questions that don’t work well due to local idiosyncrasies. (In contrast, the PISA international school achievement tests have a “we’ll fix it in post-production” philosophy of dropping poorly designed questions after the PISA test is given. But in either case, it’s important to figure out at some point which questions just don’t work the same across space and which ones work well around the world with just simple translations.)

Wicherts et al have noticed that heritability is strongest on the most culture loaded subtests, which is very important. But I want to focus today upon the potential implications of their data (the Cultural Load column in my table above) for better understanding of the Flynn Effect.

My table above combines the two sets of figures for Weschler substests. (Note the oranges to tangerines comparison of WISC [Flynn Effect] to WAIS [Cultural Load] — there are a ton of technical issues here, such as the Digit Span subtest being missing from Flynn’s data, but I’m just going to blunder onward.)

Eyeballing my table, it looks like there’s a moderate negative correlation between the size of the Flynn Effect and the size of the Cultural Load. The correlation is -0.44.

This overall pattern shouldn’t be surprising because it’s in line with the general difference between the Raven’s and the Wechsler’s: the more a Wechsler subtest is like the Raven’s, the higher the Flynn Effect. Conversely, the more culture-dependent a Wechsler subtest is, the lower the Flynn Effect.

For example, “vocabulary” is the most culturally sensitive Wechsler subtest, not surprisingly, and it’s got a quite small Flynn Effect. Interestingly, vocabulary’s also a really good subtest of overall intelligence. For instance, the ongoing General Social Survey includes a 10 word vocabulary test and that has proven to be a surprisingly decent proxy for IQ.

If we leave out the “Similarities” outlier, the correlation is -0.74.

My best theory for what’s going on with the Flynn Effect besides obvious ones like better nutrition is that the world has seen a major cultural / environmental shift that has been going on in most cultures around the world at a fairly steady pace that makes young people better at certain subtests, typically on Performance IQ subtests, but doesn’t do them much good on Verbal IQ subtests except for “Similarities.”

As I wrote in 2007 about “Similarities:”

Finally, the fastest rising subtest on the WISC, Similarities, rewards abstract scientific thinking, what Flynn calls viewing the world through “scientific spectacles.”

A child gets a maximum score for replying that dogs and rabbits are “mammals.” A kid in 1947 who had never seen a nature documentary on TV would likely have said “They have four legs” or something else more concrete than the Linnaean category “mammals.”

In 1947 a child in the hollers of Kentucky would probably know more concrete things about dogs and rabbits than an urban child today. But IQ tests have tended to anticipate the direction in which global culture has evolved, away from the concrete and toward the abstract and two-dimensional, toward what can be represented on a piece of paper or a screen.

Whatever this change is, it’s reminiscent of Moore’s Law in its endurance and steady pace. As you know, around 1968 Gordon Moore of Intel, the famous Silicon Valley silicon chip firm descended from Shockley Semiconductor, pointed out that Intel had been able to double the number of transistors on a standard size piece of silicon every year or two throughout the 1960s, and he believed that the industry would be able to keep up this pace for some time into the future. This more or less proved true for at least four decades, with world changing consequences, such as the coining of the term “Silicon Valley” in 1971 and the rise of Silicon Valley to immense economic importance.

I don’t know if Moore’s Law is still in effect (the laptop I bought in 2015 is only trivially faster than the one I bought in 2012, the first time in my personal computer owning career, which goes back to 1984, that a new computer wasn’t tangibly faster). Similarly, I don’t know if the Flynn Effect is still operating everywhere. (I haven’t really been following the data in this decade.)

But Moore’s Law has been kind of like the Flynn Effect in that it has been relatively incremental, decade after decade, rather than erratic, and the effects have been felt globally even though its heartland has been Silicon Valley, kind of like how IQ testing’s heartland has been Silicon Valley ever since Lewis Terman released America’s first IQ test, the Stanford-Binet, a century ago.

Moreover, Moore’s Law (in the sense of higher performance in general) has had multiple causes. For example, when clock speeds on CPU chips topped out, the chip companies were able to regroup and keep Moore’s Law progressing for a number of years further by doing other things. Similarly, it’s likely that better nutrition both contributed to the Flynn Effect (the U.S. added micronutrient supplementation of both iodine and iron to staples between WWI and WWII) in the past, but improved nutrition has been less of a contributor to the Flynn Effect in some countries in recent years as nutrition has gotten about as good as it’s going to get. But other more mysterious factors apparently stepped in to keep the Flynn Effect going a while longer.

So, Moore’s Law is an informative analogy for the Flynn Effect.

But I would go further and suggest, somewhat hand-wavingly, that one of the driving forces of the Flynn Effect has been Moore’s Law, or, to be both more precise and more vague, some kind of superset of a direction to technological change of which Moore’s Law is a subset.

One of the big changes in daily life over recent centuries has been the growth of what I might call humans having to deal with “machine logic.” People today deal far more often each day than in the past with semi-intelligent machines who can only be dealt with in a certain way according to their own logic. You deal with the ATM rather than with a bank teller, with a gasoline pump rather than with a pump jockey, with elevator buttons rather than with elevator operators. You can’t wave your hands around with these machines until they figure out what you want done. You have to follow a precise logical series of steps.

(This trend may not continue forever. For example, searching the Internet using Google today requires users to use less logic than searching the Internet using Alta Vista in 1998 required. The term “Boolean operators” was useful to understand to get more out of Alta Vista, while Google is so smart today that you don’t have to be as smart.)

This trend toward people having to interface more each decade with machine logic hasn’t just been happening since the silicon chip was invented. Before the silicon chip was the transistor, perfected by William Shockley, and before that the vacuum tube, which Lee de Forest made significant progress upon in Palo Alto around the time Lewis Terman of Stanford was adopting Binet’s pioneering IQ test for the American market.

Granted, I’m waving my hands around in making this argument in the hopes that you’ll grasp what I’m trying to get across. I don’t have this reduced to a precise series of steps that a machine intelligence could understand, but I do think I’m onto something: that the high Flynn Effect, low Culture Load IQ subtests are a kind of like mastering dealing with information technologies, and kids these days get more practice in that than we did and we got more practice than our parents did.

In contrast, kids these days likely have less practice dealing with complex 3-d entities, such as repairing automobile engines. Instead, they are used to dealing with 2-d paper and, ever increasingly, 2-d screens. But IQ tests tend to shy away from much in the way of 3-d testing, other than some blocks subtests on the WISC and other children’s IQ tests, largely for reasons of economy. Asking and answering questions in a 2-d format, whether on paper or on a computer screen, is cheap.

But because 2-d is cheap, the real world has also moved in the 2-d direction that IQ tests anticipated.

 

One thing that seems pretty likely is that in each person’s life, he has a window where it’s easy and fun to learn to communicate logically with a new set of systems, and over time that window closes. For example, when I was in the marketing research industry, I jumped all over the coming of the personal computer in 1984 and the Internet in 1996.

More senior executives at the information company where I worked back then tended to find the new personal information technologies difficult to master. They were used to issuing orders to intelligent human beings, such as their secretaries, who wouldn’t take everything quite so literally. The founders of the company where I worked were superbly intelligent at dealing with human psychology, but they found arbitrary machine logic daunting.

But similar information technology developments in this century have not struck me as fun at all to learn about. On Twitter, for example, I’m basically clueless about whether I’m replying to one person or to thousands. Today, I feel like the Vice Chairman of my employer back in 1984 when he gave me his $9,000 IBM PC XT with the coveted 10-meg hard disk because he was too old to learn to type.

Generation after generation, children grow up in an environment ever denser with the kind of systems logic that the more Flynn Effected-Wechsler subtests ask about. Growing up, kids these days get more practice with the kind of thinking tested on the Raven’s and on some of the Wechsler subtexts. And they legitimately are better at it.

The Flynn Effect is a side effect of the developers of the IQ test being on “the right side of history.” We’re used to hearing progressives denounce IQ tests as obsolete pseudoscience on the wrong side of history, but, in reality, IQ testing in the United States has some amusing organic ties to the triumph of Silicon Valley. Louis Terman’s son Fred Terman (1900-1982), a professor of electrical engineering at Stanford, was the perhaps the single most important figure in the rise of Silicon Valley. The mentor of Hewlett and Packard, he largely invented the model of Stanford grad students like Larry Page and Sergey Brin starting up high tech firms like Google.

You are supposed to believe that the Termans were all wrong, but it sure looks like we’re living in the world the Terman family anticipated.

 
• Category: Science • Tags: Flynn Effect, IQ, Moore's Law, Robots, Silicon Valley 
🔊 Listen RSS

My 2012 MacBook Air recently gave up the ghost and so I replaced it with a 2015 MacBook Air. In the past, a new laptop was a fun thing because it was so obviously better than the previous one. But this time, even though I paid for doubling the RAM from 4 gig to 8 gig, eh … This is not say that there are no improvements, but they tend to be harder to notice ones, like if in the past I could go 4 to 10 days between rebooting, now I can go to 6 to 15. Or something. Which is great, but less than galvanizing. Basically, the laptop computer now works really, really well for what I use it for, so my complaints are extremely marginal.

Still, Moore’s Law is getting old and less vigorous. From the NYT:

In recent years, however, the acceleration predicted by Moore’s Law has slipped. Chip speeds stopped increasing almost a decade ago, the time between new generations is stretching out, and the cost of individual transistors has plateaued.

Technologists now believe that new generations of chips will come more slowly, perhaps every two and a half to three years. And by the middle of the next decade, they fear, there could be a reckoning, when the laws of physics dictate that transistors, by then composed of just a handful of molecules, will not function reliably. Then Moore’s Law will come to an end, unless a new technological breakthrough occurs.

To put the condition of Moore’s Law in anthropomorphic terms, “It’s graying, it’s aging,” said Henry Samueli, chief technology officer for Broadcom, a maker of communications chips. “It’s not dead, but you’re going to have to sign Moore’s Law up for AARP.”

In 1995, Dr. Moore revised the doubling rate to two-year intervals. Still, he remains impressed by the longevity of his forecast: “The original prediction was to look at 10 years, which I thought was a stretch,” he said recently at a San Francisco event held to commemorate the 50th anniversary of Moore’s Law.

But the ominous question is what will happen if that magic combination of improving speeds, collapsing electricity demand and lower prices cannot be sustained.

Moore’s Law probably offers some useful perspectives on IQ testing and the Flynn Effect. First, just as Dr. Moore was amazed by how much longer his forecast lasted than he had expected, it’s amazing that IQ tests are now in their second century and still more or less working. An incredible amount has changed from the early 20th Century — how many stand-up comedy routines from back then are funny today? — but the basics of cognitive performance testing have been pretty stable at least from the 1930s onward.

I don’t think that’s guaranteed to endure another century. For example, the Tiger Mothering / Test Prep phenomenon suggests that massive cultural effort can be used to if not beat, then at least bend high stakes tests.

The Flynn Effect of steadily rising raw tests scores on many (but not quite all) types of IQ tests is reminiscent of Moore’s Law in that it doesn’t seem to have any single cause. Moore’s Law represents a fairly steady trend, but one achieved over a half century due less to any single cause but to a wide variety of small advances. Similarly, improved nutrition may have played a big role in the Flynn Effect in the past, but less so in more recent years.

Moreover, I think a lot of the cultural change in the world in recent decades in what children learn is important to be able to deal with drives Flynn Effects, and is in turn driven by Moore’s Law. Today’s children, for example, find it natural to interface logically with machine intelligences through 2-dimensional interfaces.

That’s a cultural construct, to some extent. If you time-transported Peak Shakespeare from opening night of “Hamlet” to a modern street and told him to figure out how to pay the computerized parking meter with a credit card, the Bard probably wouldn’t be able to deal, judging by the number of middle-aged people I see staring worriedly at the new parking meters and poking them uncertainly. Shakespeare’s vast resources of eloquence, erudition, human cunning, and psychological acuity would be useless in dealing with the stupid parking meter.

A theory I have is that early IQ researchers, such as Stanford’s Louis Terman a century ago, had an intuition about which way the world would evolve, and both for reasons of foresight (a reasonable case can be made that Terman’s son Fred Terman was the Father of Silicon Valley; the other usual candidate for the title is Fred’s pal William Shockley — notice a pattern?) and reasons of convenience (dealing with 2-dimensional pieces of paper or 2-dimensional screens is cheaper than dealing with 3-dimensional objects), they came up with tests that gave a foretaste of what a lot of work is like today: you deal, through 2-dimensional interfaces, with machines that only accept logic.

 
• Tags: Flynn Effect, IQ 
No Items Found
Steve Sailer
About Steve Sailer

Steve Sailer is a journalist, movie critic for Taki's Magazine, VDARE.com columnist, and founder of the Human Biodiversity discussion group for top scientists and public intellectuals.


PastClassics
The “war hero” candidate buried information about POWs left behind in Vietnam.
What Was John McCain's True Wartime Record in Vietnam?
The evidence is clear — but often ignored
Are elite university admissions based on meritocracy and diversity as claimed?
A simple remedy for income stagnation