Sunday, October 31, 2010

The Business of Intelligence

A report from Science Daily highlights the problems of modern intelligence research:
ScienceDaily (2010-10-30) -- General intelligence is not enough. Practical intelligence can mean the difference between entrepreneurial success or failure. Psychologists have identified multiple kinds of intelligence, but a new study has found one -- practical intelligence -- to be an indicator of likely entrepreneurial success.
Charles Spearman created the concept of the "g factor", a general intelligence that underlies the styles of intelligence measured in iq tests. I'll stay away from the whole issue of iq tests, I have two professional friends with a behaviorist orientation and they seem to hate IQ concept with a vengeance. I think I know what they are getting at but I still maintain that IQ as a metric is fine but has a theoretical construct or as data for the same it is useless. The history of the debate about IQ is in itself instructive that when dealing with the question of intelligence we are still very much in the dark. At the end of this post I will reference a research item that demonstrates we may have absolutely no idea about the true basis of intelligence.




The first issue we need to think about is the word - intelligence. The dictionary definitions are so broad as to suggest the word and associated concept is analytically useless. The very word itself presupposes a g factor. Hmmm, echoes of Whorf Sapir, but then Spearman created his concept through a mathematical analysis of various results in differing iq test sub domains. So is the Whorf Sapir hypothesis relevant with respect to the training one receives in mathematics?

"Intelligence" is not  a mathematically nor analytically derived concept, it is a folksy idea that is kicked about with jolly abandon. Could it be that a mathematical strategy arose from an assumption embedded within the concept of intelligence? A self fulfilling prophecy? Is that what Wittgenstein meant when he wrote that mathematics is just tautology? Can he show me the way out of this fly bottle? Crack!

Intelligence is a metric without a fixed quantity. That is oxymoronic. The metric within iq tests is derived from statistical modelling of a given population. The Flynn Effect, a long observed phenomenon where across nations iq has been rising, requires constant adjustment of iq tests. IQ tests are good predictors of future academic results, probably the best predictors. There has always been a strong correlation there. In recent years though an anomaly has become apparent.

While the Flynn Effect appears to be flattening, academic standards have clearly been falling. The correlation cannot hold, the linkage is broken but that was probably inevitable. I have previously explored the reasons for this breakage in an earlier post. Basically it goes like this: a few generations ago only the best and brightest went onto to tertiary education. Now even the just above average and dimly glowing are going to tertiary education. The standards have to fall because the total pool of graduates is of a lower overall quality. I'd like to see studies with a decile breakdown of student performance. What I would be looking for is whether or not the decline in standards is across the board or confined. I also expect that the top 2-3 decile of graduates would be equal and more probably superior to days gone by. Surely such studies have been done? Can anyone help me here?

So there is all this chit chat about some general decline in intelligence and this amongst what is unequivocally the most educated generation of human beings. That is irrational Captain, we need to quickly find another star system. What we are perceiving is the perception of a "dumbing down" that is the result of increasing education. All these people complaining about the dumbing down, did they study set theory at school? Who has been dumbed down? Uhura, fix that viewing screen.

The above argument only holds with respect to academic performance. There are other observations that do suggest a dumbing down is occurring but again those observations can be interpreted in different ways. I don't buy into that, I take the more optimistic view that upcoming generations have realised that earlier generations wasted a great deal of energy on intellectual type things that were not worth the bother. Too many fly bottles and too few Wittgensteins. That is not something I wish to investigate here as I have already wandered far from my original intention. Lazy Sunday afternoon, got no mind to worry ... .

What the above does illustrate is that when it comes to the question, "What is intelligence?" it appears that many have made up their minds before even thinking about what it is they are using to make up their minds. The beauty of philosophy is that it makes you return to the first truly important questions.

In this news release they are investigating the idea that being successful in business requires more than "book smarts". [At this point I read the whole news article instead of the intro]. Thus the author wisely proclaims ...

There are many kinds of intelligence, including emotional, social and creative. Practical intelligence is just one; but a critical one for entrepreneurial success, said Baum.
Can I blame Howard Gardner for this? We've gone from Spearman's g factor to a multitude of intelligences. Sulu! We're heading into an asteriod belt. Take evasive action! Is intelligence even a thing? When we say someone is intelligent we are observing that they demonstrate certain skills in performing certain actions. In this context "intelligence" signifies the efficacy of their performance but why presume that each demonstration of that efficacy is a function of some common thing across all those differing domains? Another ghost in the machine?* I wish I hadn't asked that question. That's the terror of philosophy. It's Sunday afternoon, something too difficult to think about it.

But about that research item ... 

This is one research piece. There are other examples of apparently extraordinary intelligence in organisms with no nervous system. This is all rather new to me, but in light of some ideas expressed by the mathematicians Stewart and Kaufmann, and some other research items addresses molecular activity through fundamentally different perspectives than those commonly entertained today, but still being consistent and acceptable explorations, I am led to believe that our understanding of intelligence may be so rudimentary as to be fundamentally misleading and utter nonsense.

*Many thanks to the British philosopher Gilbert Ryle, The Ghost in the Machine(1949), which I read when I should have been reading Sports Weekly. It is easier to kill a man than the ghosts within him.

9 comments:

Frank LeFever said...

I suppose I am one of the "behaviorist" friends John alludes to, but as a neuroscientist I violate the principles of the purest behaviorist (B. F. Skinner, thesis mentor of my thesis mentor) because I do peek into the "black box" -- often with John's collaboration.

I sometimes cite Muriel Lezak's presidential address to the International Neuropsychological Society, title something like "Time to Bury the IQ Concept".

As with most (all?) unanswerable "questions", the problem lies in failure to agree on (or even to OFFER) a definition of what "it" is that we are discussing.

[As an undergrad philosophy major, I was cured of philosophy when I read that Wittgenstein said the history of philosophy was mostly full of arguments about questions which were not framed in a way that allowed an answer.]

Decades ago, in an effort to be scientific, psychologists seized on a note by Percy Bridges (physicist) on "How to Make Our Ideas Clear" and adopted "operational definitions": i.e. defining things in terms of what measuring operations were required to determine whether it existed or did not, occurred as an experimental result, was an independent variable, or whatever.

The deep flaw is that they then tended to confuse pronouns: they typically forget that "an" operational definition is not "the" operational definition.

It's all a matter of how one person or another chooses to define intelligence. There is no such "given" as some "thing" called "Intelligence" just waiting for someone to describe it "correctly".

Anonymous said...

well - i have a little reading to do.

and frank lefever - "...in an effort to be scientific, psychologists seized on ..." - superb!

i fervently hope that in time much of what constitutes the 'science' of psychology will be debunked.

Peter Patton said...

Hi John

I have never been interested enough to look beyond what I read on blogs, but couldn't the Flynn Effect be not much more than that the sample we have of the standard deviations on each side of the 21st century (or late 20th century) population is so much, much larger - in both pure number and as a proportion of the the population - than they had back whenever they first settled on their standard of 100?

That is, we have simply worked out that their original anchors for 100, were a bit off?

Then there are the issues related to mass migration of populations across the globe. How many Asians, Indians, Africans, or even Greeks were included among Weschler's subjects in 1949? or Stanford-Binet's in 1930, or Raven in 1940?

Further, there was probably a bias towards the sub-100s in mass/war deaths over the 20th century, and that's before we even dare start whispering about the abortion stuff.

I also would not be surprised if their original samples never had equally representative samples from both side of the "great divide". It seems quite probable, they were a lot more accurate about the 130s than the 70s say.

Given how immediately IQ tests became ubiquitous after the 1940s, it would be expected that the very narrow fields of cognition that these tests cover, would insinuate themselves very snuggly into the broader curriculum, in a way that did not happen before 1940.

For example, remember that Harvard Entrance exam from 1896? No block-designs to rotate, no number sequences, and no Analogies!

Anyway, surely, they would have done some work seeing how far to the right, the Flynn Effect has crept. I wouldn't be surprised if the Flynn Effect turns out to be little more than the new 100 is around the level of 90 from the mid-century, but 110 has changed much less, and 130 not at all.

I also have a hypothesis that a reverse Flynn is taking place 3 sd to the right, as the education system has slowly, but decisively removed the more 'g' loaded topics from the broader school/university curriculum - for completely cultural political reasons - and even away from standardized test themselves. For example, the GRE has just abolished the use of Analogies.

Without Analogies, how can you really tap into the 130+ area anyway?

And if you can't identify folks, beyond even 2 sd to the right, what good will be the GRE be for graduate schools?

John said...

That is, we have simply worked out that their original anchors for 100, were a bit off?

Could be, "n" can affect results. But it is beholden upon you to establish evidence for that.

I also have a hypothesis that a reverse Flynn is taking place 3 sd to the right

I sincerely hope you are wrong about that but the evidence is against me. I would love to see a decile(with a smaller breakdown at the tails) breakdown of iq scores. These average values aren't telling us enough.

We discussed this issue long ago on L. Pradeo, where I put up the suggestion that because of the democratization of education the standards had to be lowered. This also impacts on averages. There are lots of confounders here. For eg: we are now saving lots of babes that otherwise would have died. ADHD has been rising and to some extent we have been able to bring these people into mainstream education but that might be lowering the averages. Autism rates are going through the roof. Developmental disorders are more common and that could be a direct result of our ability to save so many babes. Premature birth could figure very prominently here.

Peter Patton said...

I am an unhappy camper, tonight, with sadly, my greater faith - than yours - in the usefulness of IQ to tell us important things, only confirmed. I have wasted several hours of my life reading this "Social Psychologist" Jonathan Haidt dude - linked to both positively by SL and CT.

Just reading him made me want to punch him. The guy would have an IQ of about 120. His Moral Psychology and the Misunderstanding of Religion fails at every level:

(i) knowledge/learning in subject matter/s;

(ii) cognitive finesse - IQ

(iii) awareness of the world around you, the very topic you are writing about.

He is trying to explain to his fellow 'progessive', 'liberal-left' American Democrats types, why they have been underestimating their right-wing GOP evil enemies.

Here's a quote. Check it out for yourself, and tell me IQ is meaningless. ;)

It might seem obvious to you that contractual societies are good, modern, creative and free, whereas beehive societies reek of feudalism, fascism, and patriarchy. And, as a secular liberal I agree that contractual societies such as those of Western Europe offer the best hope for living peacefully together in our increasingly diverse modern nations (although it remains to be seen if Europe can solve its current diversity problems).

http://www.edge.org/3rd_culture/haidt07/haidt07_index.html

John said...

Moral Psychology - dumb as. Psychology should be about the study of behavior. Morals are behavior. If we want to understand moral behavior we will have to do it without using morals as causative factors.

That religious people are more moral etc etc etc means nothing to me. Mark Twain has more insight:

Don't part with your illusions. When they are gone you may still exist but you have ceased to live.

The Edge is popular science, which most people take to be science. Two different things. The Edge is for the popularisation of science. That means it is for the dumbing down of science. Consider all the books on QM and relativity. What those books are doing is beguiling the public into believing about phenomena where the theories breakdown. Everyone is discussing what the theories can't explain, which is kinda stupid. It is pointless to use theories to explain things which are not explained by the theory. All theories are bounded and must be used only within the relevant boundaries. Psychologists miss this point all the time.

IQ is useful for demonstrating cognitive ability. I have read that it holds pretty good within 2SD but beyond that tends to fail. Hence the findings of Liam Hudson, that beyond 130 other factors become important in scientific success. I have read about creative people and the general view is that the one most common characteristic in them is perseverance. It is not enough to be smart because solving problems is not just about being smart. There is chance and circumstance involved. Serendipity is something we prefer to not think about because we like to think we can logically solve every problem. But at the cutting edge there is no logical path, there is trial and error, there is relying on a massive amount of learning and experience which is nothing else avoids us going down the dead ends explored by others.

If you look at lists of successful people they often have iqs >130. What you don't see is all the other people with iqs >130 who are not famous. I worked with two women who had iqs >150 but you wouldn't know it working with them ... .. So if we want some insight into into what makes a smart person successful we would be better of studying smart people who are not successful.

John said...

When I was studying neuroimmunology I certainly didn't bother with The Edge, though I might entertain Scientific Blogging. Mostly I read the primary literature, then reviews, but for the most part I had no interest in textbooks and when it came to news releases I was looking for the primary article to read not the babbling nonsense that permeates so many science news releases. My favourite journal was the Journal of Biological Chemistry. No theories, mostly raw data but often unreliable data. At the cutting edge the error rate is always high. Here is the dirty secret in biomedicine that few dare mention: we don't have theories that can withstand the assault of primary data for too long. In neuroscience it is even worse.

Psychology is filled with crap. What amazes me is why so many graduates leave college and stop reading the primary literature. Biggest mistake a smart person can make. If you want to advance your understanding go back to the source, not the ramblings of websites dedicated to making huge hit counts or for that matter big scores on citation indexes. Blogs are worse.

Peter Patton said...

John, just browsing through your archives, I liked your post on "Epigenetics and PTSD". I've done some thinking, reading, and chatting about 'transgenerational transmission of trauma' and am thoroughly convinced of its empirical existence. But I am fascinated by the transmission mechanisms.

While I'm much further along the track on the psycho-social mechanisms, I have always had the strongest feelings in my waters, that the transmission is more than social, and even more than psychological.

I lack the science training - particularly chemistry - to get into fisticuffs over DNA, but I have thought a lot - and laterally, given my relative ignorance of experimental sciences - about 'cellular memory'. Now, that I seen you link PTSD with 'epigenetics', I shall go and furiously beaver away at getting up to speed on the molecular biology necessary to contribute sensibly.

But - finally, the reason I even posted - have you considered that the maze-mastering slime you linked to above might be telling us to look at intelligence and cellular memory at the same time?

John said...

Peter,

For 25 years I have been arguing that if we wish to help countries that have been ravaged by constant internal strife, where childhood trauma is the rule not the exception, our efforts must be mostly aimed at creating better environments for the current generation of children because our chances of altering adult behavior are small. This touches on the cellular memory idea. That is, the behaviors we see in those countries, and amongst the disadvantaged in our societies, are not arising just from sociological and psychological issues but are are arising from much deeper and, too often, intractable roots. It is a despairing issue because there is no obvious solution. The "poverty cycle" has a physical substrate. So yes, the issue of "cellular memory" here presents us with unique challenges that are only just being realised. The answer will not come for the Left or Right, both of which entertain atavistic notions about human behavior. We need a New Kind of Psychology.

Sadly most refuse to recognise that adult human behavior is not as malleable as we would like to believe. The problems here also point to profound philosophical issues with respect to understanding human behavior. That article you referenced highlights that in psychology many still refuse to recognise that when thinking about brains and behavior we must move beyond folk psychology and its concepts. For example, at one point the author states "brains try to .... ". Brains don't *try* to do anything. Brains respond. Now look at his definition of morality:

Moral systems are interlocking sets of values, practices, institutions, and evolved psychological mechanisms that work together to suppress or regulate selfishness and make social life possible.

Circular, waste of time. He is not an exception, psychology is littered with these types of mistakes. Skinner tried in vain to warn people about the dangers of invoking everyday ideas to explain behavior. The ghosts of dualism seem impossible to exorcise.

This presents us with a much larger problem. Should we even bother to read the likes of Mill and other philosophers, most of whom have tried to understand human behavior and society by reference to everyday concepts? The big exception here is epistemology - getting our concepts right.

You may now gain some insight into why I have always veered away from trying to understand human behavior. We don't, or at least I don't, have the right tools to approach this subject. So I prefer molecules. Easier. Ironically Darwin long ago recognised the depth of the problem:

Darwin's M Book

"Origin of man now proved. - Metaphysics must flourish. - He who understands baboon will do more towards metaphysics than Locke."