On human intelligence… a perspective from computer science

Whenever I read social scientists, there is often, implicit in the background, the concept of “intelligence” as a well defined quantity. I have some amount of intelligence. Maybe you have a bit more.

But what does computer science have to say about any of this? That is, what is this “intelligence” we are talking about?

  • Except for storage capabilities (memory) and speed, all hardware is equivalent. Though I don’t know much about biology, I doubt that any brain runs at twice the speed of another brain. We are left with memory as a constraint. However, what computer science tells us is that you can always extend your memory with external support (in this case, use a pen and paper, or just Google) and all it might do is slow you down. Thus, it would seem that the main difference between individuals ought to be speed: some people may learn faster than others, or execute some tasks faster.
  • There can be huge differences in software, but software can be upgraded.

If one human being can do something intellectually, then most other human beings can also do it, albeit more slowly. They may need more tools, and they may require more time… but that is all. Yet, given that we have more sophisticated tools with each year that passes, I would expect that the difference in “human intelligence”, whatever it is, ought to diminish in important with time.

I expect that what sets people apart is not this ill-defined intelligence, but rather pure grit. If you want to do something, but you apparently lack the “intelligence” to do it, then it may simply be a matter of finding or building the right tools.

In some sense, this is what software is all about: extending our intelligence.

The idea that one could measure my “intelligence” by locking me up in a room and having me write a test with pen and paper is ludicrous. As I wrote this blog post, I used a dozen sophisticated pieces of software, including Wikipedia and Google Scholar. Where does “my” intelligence ends and where do the tools start? There is no line. What you perceive as the result of “my” intelligence is the result of an aggregate of hardware and software where the genetics of my brain are just one of many ingredients.

Someone with a gigantic memory in 1900 could certainly be a much better scholar. However, in the day and age of Google Scholar, memorizing thousands of quotes and references is much less of an advantage. You may work faster if you don’t need to constantly look up information online. However, as Google and its competitors get better, the scholar with a gigantic memory eventually loses out to someone who relies more on computers… the same way Chess players lost to computers.

To be a scholar in year 1500, you need to have a superb brain and superb education. In 2014, almost anyone can become a scholar… engage in deep intellectual debates, write compelling essays…

Classifying kids as talented or not talented is a gigantic failure of imagination. Software technology has the potential to turn the kid who couldn’t figure out fractions into someone who can do advanced science.

Daniel Lemire, "On human intelligence… a perspective from computer science," in Daniel Lemire's blog, December 24, 2013.

Published by

Daniel Lemire

A computer science professor at the University of Quebec (TELUQ).

31 thoughts on “On human intelligence… a perspective from computer science”

  1. You seem to downplay the fact that having access to those equalizing software is dependent on amount of money one has.

  2. @rejuvyesh

    I do not. But on a worldwide scale, we are getting more and more equal.

    Back 30 years ago, there was a huge difference between me and the average Chinese. This difference is far less significant today (though I am still much richer).

  3. Quite an interesting topic, Daniel. And I do agree at a broader level with most of these thoughts, and specifically that of questioning the usefulness of measuring one’s intelligence in an artificial mental and physical enclosure.

    Man, as a social being, has been benefiting immensely by codifying and sharing the experiential knowledge of generations and passing them on to his offspring. Presence of the computers, cheap secondary memory and tools like Google et al. have indeed been acting as catalysts in this regard. Because of this, every child born today need not start from ground zero. There is a platform to build on top of.

    Intelligence still lies in questioning the status quo, disrupting an accepted belief through systematic and scientific proofs, making things that much more simpler for the upcoming generations so that they solve different problems altogether. It also is affected by the courage and self belief to stand up for such a revelation one so firmly believes in. And this would always be the differentiator. But surely, we would not want to pit people against each other and measure them and their intelligence.

    Also, another thought that always crosses my mind is that of this apparent “help” the external memory is providing us. As we evolve, we may, at some point, become heavily dependent on them alone and become some kind of “zombies”. And who is producing all this data in the external memory? Not a select few “intelligent” folks. This is the data generated by everyone – at their current level of intelligence. Won’t this production – consumption cycle limit the universe of overall knowledge available at some point? Wouldn’t we be better served to develop a habit to constantly look outside of these set notions? That is where, intelligence still has a big part to play in my opinion.

    With the works and quotes of Daniel Kahneman and Kalidasa alike, this certainly is worth delving deeper.

  4. This post only makes sense if were to you consider “intelligence” to be nothing much more than the sum of the facts at your disposal. But, facts (“knowledge”) is just the fuel for intelligence, not its substance. An intelligent person is able to reason about facts, and use them to develop an understanding of something. An idiot with a lot of facts at his disposal is still an idiot.

    A kid who can’t be taught to understand fractions cannot, in fact, be turned into scientist by arming him with advanced software technology.

    A mystifying post.

  5. Though the hardware is decidedly *not* the same. Genetics differ. Epigenetics differ. The brain rewires.

    Given how life experience reshapes the brain, distinguishing hard and soft is… a bit squishy.

    Some psych folk suggest a collective social mind. So “mind” always had an external aspect.

    How quickly our mental spider can find and follow threads in the external web, makes a difference. Might even magnify the difference, at times.

  6. At a very low level, yes, no brain runs an order of magnitude faster than another. But it is obvious to me that there are people who are “multiple times smarter” than I am. The difference is not in the hardware clock, but rather in the efficiency of the data structures and appropriateness of the representations their neurons organize themselves in. You worry about data structure efficiency for a living, after all.

    I am far from certain that there is always a way to extend one’s cognitive capacity with better tooling. I’m finding it very hard myself. There is a lot of biochemistry that we may not be able to do much about, at least not yet. Emotion and motivation – grit – does not map to von Neumann’s model.

    The “all hardware is equivalent” assertion is also very misleading. A slightly different hardware model can make a huge asymptotic difference in algorithmic complexity. That is not just a quantitative difference, it is one of quality.

  7. This is only true for *general purpose* computer hardware. While you could argue that any Turing complete machine can simulate another one, that is not possible without a huge overhead. The brain has special purpose vision, audio, language, etc. circuitry. You can’t just consider these to be exchangeable.

    Extending your memory with external support works only if you know exactly how that memory is to be used. But much pattern matching that happens in the human brain is essentially a black box. We don’t know how it uses the ‘memory’. If you need to be able to hold 10 things in your mind simultaneously to see some pattern or to understand some concept, and your brain can only fit 9, then no matter how much paper you use, you will not be able to see it.

  8. intelligence can become a utility just like fire or electricity. In few centuries, computers will be compared with discovery of fire.

    a human is nothing but a biological computer. he can be an idiot, or an artist. But it doesn’t define her. idiots change all the time. They just need proper “biological compute context” (society, upbringing, education etc).

    imagine every human having access to any intelligence he wants, probably freely provided by the government. He doesn’t have to be intelligent to use intelligence. “intelligence” will be molded into forms that any idiot can use. Just like fire or electricity.

    so yeah, human capabilities wont matter much in the future. All that matters is what kind of tools (not tool in traditional sense, but any forms that enhances any humans often without his knowledge or expertise) he has access to.

  9. All your reasonning neads to be established as true is to postulate an informational theory of intelligence. But in reality, this is not an universal kind, nor a true these, from a biological standpoint nor from a psychological one. The concept of «information» in biology, sociology, anthropology and psychology, is generally not (except in some areas in those sciences where scientists have borrowed somewhat freeloaddingly directly from information theory) and cannot be equivalent to «information» in information theory. You just cannot measure bits of information inside one’s brain, because there aren’t. What you will find are neurons, axons, white matter (which is fat), hormonal exchanges, some electrical current, and many mitochondries transforming nutriments into other things. But no bits of information. True, you could modelize each and every process with the means offered by information theory; but that would be just that: theoretical models, nothing more than some sorts of analytical devices.
    So: what is human intelligence? Certainly not something comparable to memory, which is usually conceived as a faculty operatively connected to learning; not as storage per se. Human intelligence cannot be reduced to one specific trait of human’s brain functioning. The most generally accepted view of intelligence, in cognitive sciences and philosophy at least, is holistic, taking into account pragmaticists concerns (relationship between an organism and its social and ecological environnements), the backgroung knowledged transmitted by parents, the probability set of social interations, some genetics as well, and some other few things. Which is not to say that this view tends to forget what it wanted to tell in the first place; it only says that scientists have been somehow mislead by the search of the graal that would have explained in only one elegant equation the fundamental reasons of human behavior. Such an explanation is part of the informational mythology.

  10. Well, I know that there are a lot of people who program, type, write, and read 2-3 times faster compared to me. I have also come across people who are 2-3 times less productive than I am. So, I suspect, there is at least an order of magnitude difference in intellectual productivity. I do not know if it is related to brain speed or not, though.

    Regarding the memory: one still needs a good one. Having a huge memory is, probably, not essential (if you can look up things efficiently), but you need to remember WHAT, WHERE, and HOW to search. This still requires a better-than-average memory.

    Same applies to computer programming: if you have a huge system, you need to remember the basic outline. And, as many professional programmers know, the better you know the system, the less effort you spend in maintaining and updating. It is again can be easily an order of magnitude difference.

    Last, but not least, reasoning capabilities (deductive, inductive, and abductive) are more important than memory. Here, even though there is some progress already, even in deductive part, Google or other automated tools are far less helpful

  11. I agree with Tom above: you’re confusing intelligence with knowledge. As experiments like http://pss.sagepub.com/content/early/2013/11/22/0956797613502983.abstract show, there *is* a difference between children and chimps in their ability to understand abstract rules. This (the ability to understand abstract rules) is what I consider to be intelligence.

    As a personal example, I had school mates who were completely unable to understand the equality of (a + b) ^ 2 = a ^ 2 + 2 * a * b + b ^ 2; furthermore, after managing to *memorize* it after writing it dozens of times, they were unable to expand (x + y) ^ 2.

  12. @Leonid

    So, I suspect, there is at least an order of magnitude difference in intellectual productivity. I do not know if it is related to brain speed or not, though.

    Given time, and a good dictionary, I could probably decipher a Russian text. You could probably do it 100x faster.

    Should we conclude that you have a superior brain?

    Regarding the memory: one still needs a good one.

    Let me give you an example where technology alleviate memory challenges.

    Java has a huge API. I know some of it, but most of the method names escape me. I have been programming in Java since 1998, but I still don’t know perfectly the grammar.

    Still, with modern IDEs, I don’t need to remember any of this. Method names will be suggested to me. If I type something that is not allowed, the IDE will tell me right away, as I am typing.

    The net effect is that I am probably just as productive as someone who has memorized the entire API.

  13. @Marcel

    there *is* a difference between children and chimps in their ability to understand abstract rules

    Our brains are at least 3 times larger than the brains of chimps.

    My claim has never been that hardware (brain) does not matter… rather, my point is that differences between human beings are not a matter of orders of magnitude. No human being has a brain 10x larger than another human being. Or even 2x larger.

  14. @Marc @Tom

    I don’t confuse intelligence and data (facts/knowledge).

    Let me give you another example. Back 30 years ago, there was a huge cognitive difference between someone like me and cab driver as far as getting around in some large city like Montreal. Today, with a smartphone, I have instant access to traffic data, I have route recommendations, GPS… This technology levels off the playing field.

    You may decide that whatever the smartphone is giving me is just “facts and knowledge”. And at some level, it is, but this distinction is irrelevant to the point I am making.

  15. @14 – but you are. Your examples are all about differences in knowledge, that’s why memory can be substituted by technology. Think about my example with the students unable to understand the expansion of (a + b) ^ 2 – this has nothing to do with memory. Those students were ultimately able to *memorize* the formula, but they were unable to *understand* it, hence their inability to apply it to the (x + y) ^ 2 case.

    Knowledge / memory can be helped by automation (I have programmed in Delphi for 15 years and I never managed to memorize the syntax of the Format function); the capacity for abstract thinking cannot.

    Maybe I misunderstand your point, but I submit that your claim that “anyone… can write compelling essays” is incorrect. Most people cannot.

  16. @Marcel

    I wouldn’t make too much of a fuss about someone’s inability to learn algebra.

    Elsewhere, on this blog, I recount the story of how, when I was a young kid, I was put in a class for kids with learning disabilities. I got this label because I couldn’t learn my phone number, or to count up to 10, or tie my shoe laces. I was labelled as a “slow learner”.

    I went on to get a PhD in mathematics at 27 and I was a full professor before I reached 40.

    I don’t think any of my colleagues, or people who have worked with me, would qualify me as a “slow learner”. Some people online call me an idiot, but it is more because they disagree with me than because they tested my intelligence and found it lacking.

    For the record, I still don’t know my phone numbers. I can count up to 10 now, but I never learned my multiplication tables. Though I can tie my shoe laces now, my knots don’t hold.

    So, why won’t I memorize phone numbers? Because I feel that the effort is not worth it.

    I suspect that a lot of kids who won’t learn algebra feel the same. It is simply uninteresting to them. Some of them might have more limited intellectual abilities as well… but their abilities are only relatively limited… they still can learn algebra, it would just take longer… and the higher the bar, the more motivation they would need.

    Of course, some people have brain defects and they genuinely cannot do abstract reasoning. But that’s another story.

  17. @Daniel I am talking about people who received training in reading and programming. And who are professional programmers. Probably, I won’t be able to read 3x as fast compared to my current speed, no matter how hard I try. Some people can. Same thing with programming.

    If you learn Russian well, you may even be able to read faster than I, despite I am a native speaker and you are not. This is not very likely, but it is certainly possible, because I am not a speedy reader. 🙂

  18. Ok, I still think you’re mixing up intelligence (ability to understand a basic formula in algebra) with knowledge (couldn’t memorize my phone number, couldn’t count up to 10), but I seem to be unable to make you understand the difference 🙂 Even weirder, you do seem to see that there is one (“they genuinely cannot do abstract reasoning” – so there is something that takes more than just time and motivation), so I’m somewhat stumped.

    I also disagree with the idea that “differences between human beings are not a matter of orders of magnitude”. I not only am at least an order of magnitude more productive than other programmers I know – I am many orders of magnitude more productive than non-programmers. I just don’t see a Mozart replaced by 100 well-intentioned Psys – http://www.youtube.com/watch?v=9bZkp7q19f0

  19. @Marcel

    I don’t consider the case of people with brain defects. I am considering healthy human beings.

    Obviously, I can open your skull with a hammer. You won’t be very smart once I am done. But this is irrelevant to my arguments.

  20. @Leonid

    I am talking about people who received training in reading and programming. And who are professional programmers. Probably, I won’t be able to read 3x as fast compared to my current speed, no matter how hard I try. Some people can. Same thing with programming.

    You definitively can learn to read many times faster. Basically, it involves using different heuristics. It is true that some people just stumble on a faster way to read, but there are trade-offs. People who read fast don’t pay attention to fonts or punctuation…

    Some people are definitively better programmers (10x more effective), but they are usually far better “trained”… and by “trained” I don’t mean “schooled”.

    I think that the massive difference in productivity for programmers has to do with attitude. Some people program to “get the job done”… others constantly hone their skills. Over a few years, the programmer who has constantly tried to improve his skills will be many times better.

    My experience is that most programmers don’t even know about StackOverflow. And, certainly, only a tiny minority participates in such online exchanges. Most programmers just put code together. If it works, it works. They would never bother to even use Google to see if there is a better way to solve their problem. They may use a brute-force algorithm and never even try to do better.

    The 10x programmers take pride in their work and try to always get better. I don’t think you will find that their brain genetics are substantially different. (Of course, their brains *will* be different. I am talking about genetics here.)

  21. @Daniel ok, so do you see this as a binary solution? You either are intelligent or not? I think it’s more of a continuum, with people at the higher end of the scale able to understand more complex abstract problems than those at the lower end.

    Hmm… I think you might agree with the above statement but still believe that the people on the lower end can compensate for the difference with effort. I am unable to think of a way to prove that wrong (though I still believe it to be wrong, as I said).

  22. @Marcel

    I wrote: “Though I don’t know much about biology, I doubt that any brain runs at twice the speed of another brain.” So, yes, there is a continuum, but it is just very unlikely that your neurons consistently fire twice as fast as mines.

    It is more likely, however, that you could have a memory much more effective than mine. But I elaborate on this issue quite a bit.

    Of course, then you need to *learn* and how much you learn will determine how smart you are… a better memory might help you learn faster, of course… but that is all.

  23. @Daniel, I think you might be wrong about reading. There are techniques to improve reading, but they are mostly at the expense of understanding. Some people are naturally much faster readers than others .

  24. @Leonid

    I checked out is Jackson and McClelland (Processing determinants of reading speed). On the various tests they have used (reaction time and so on), they do find differences between healthy test subjects, but they are relatively small (e.g., 10%).

    Of course, some people have reading disabilities. Strangely enough, many of them manage to become highly successful.

    There are also individuals (e.g., some rare autistic people) who have vastly superior reading abilities but it usually comes with a trade-off.

  25. @Leonid

    We are all descended from a few human beings. Despite this high similarity, in any one specific task, you will find people that are much better than average. Chess, mathematics, reading…

    But that is not my point, my point is that technology and education can level the playing field.

    My point is that expertise is more about training and education than about having a superior mind.

    And that means that your desire to be good at what you are doing matters probably a lot more than innate abilities.


    You have legs, I presume. So you can run. Maybe you think that you can run faster than someone without legs… But can you run faster than this man? http://en.wikipedia.org/wiki/Oscar_Pistorius

    I doubt it.

    A guy like Pistorius, with a disability, is able to run faster than you will ever run because he uses technology to compensate.

    If you can get a leg-less man to run super fast, why can’t you get someone with so-so innate intellectual abilities and turn him into a scientist?

  26. @Daniel, technology does help this is for sure. There are limits, however. Wikipedia, Google Scholar, on-line forums do help me, but, mostly because I am already at a rather advanced stage. It’s like Coursera: it helps but mostly people with good education.

  27. @Leo

    There are limits, however. Wikipedia, Google Scholar, on-line forums do help me, but, mostly because I am already at a rather advanced stage. It’s like Coursera: it helps but mostly people with good education.

    I think that Wikipedia and Coursera will appear quite primitive to us in 20 years. I am actually surprised we don’t have yet brain implants giving us direct access to the equivalent of Wikipedia at all times. You can be quite sure that it is coming.

    But it is true that just because we have the technology, we cannot be sure that people will be able to use it. Though leg-less guys can run fast… very few people have access to such technology.

  28. @Daniel, aren’t you implying that if you chop off my legs, install prosthetic ones, I will be running like hell? 🙂

    I don’t want any of those chips in my head, not at the current modest effectiveness of the search!

    There does exist at least a two-fold difference between the best and the average:
    So, do their brain tick twice as fast an average one? I have heard that fast reading is not exactly an acquired skill, these people are mostly naturally-fast readers. So, some improvement was achieved through training.

  29. @Leonid

    (1) I am saying that had you be born without legs, people would have predicted that you could never run. And they would have been wrong. With proper training, you could be running faster than all of them today.

    I am not saying that some people don’t run faster due to better genetics. I am saying that technology can make this different less acute.

    (2) Carter’s research is interesting and aptly summarized by the article you link to. He found out that nobody can read faster than 600 words per minute and still really understand the content. Though the article cites an old research, Carter’s recent papers still reiterate the same upper bound. It is about equivalent to 2 pages a minute.

    The point of his research, and of this article, is to debunk the myth of the super reader. It says so very clearly: at best, the best reader is twice as fast as the slowest reader. This is a worst case upper bound. And it is in a competition setting, not sitting down for the evening with a good book… in which case the article says we all read at the same speed.

    Carter constantly downplays this myth of the super reader in his research. What he says is that the people who pass for super readers are simply people who have learned to skim efficiently when needed.

    In any case, this is all beside the point. What if someone reads twice as fast as you do? Yes, some people will do lots of things faster than you ever could.

    A far more interesting question is: how could we use technology to read faster? Because, let us face it, 600 words per minute is not going to be enough.

Leave a Reply

Your email address will not be published.

You may subscribe to this blog by email.