Friar Broccoli wrote:

OK, I read your post carefully, and agree with you in every case
where you challenged me on a point of fact (with one or two
marginal exceptions).  Therefore, to keep this reply reasonably
clean I have snipped everything up to your concluding paragraphs
which I will write away from.  Sorry about the length.


John Harshman wrote:

Friar Broccoli wrote:

If the pattern were of a general increase in brain size in
animals in response to an environment of increasing
complexity, we would expect such increases to be broadly
distributed over most or all groups. Instead we get occasional
bumps in a few groups. (And I see no sign that arthropods are
more clever now than in the Cambrian.) We certainly see no
trend, even in the groups that have received these bumps,
toward human-level intelligence.

 To convince you that intelligence is an inevitable (or at least
 a not improbable) consequence of evolution I believe I must
 demonstrate that it has a function in a specific type of niche.
 I must also show that that niche will pretty much always come
 into existence during an evolutionary process and that the
 function is sufficiently important that evolution will always
 drive some creatures toward it.  What follows is my first
 attempt:

I don't think you have managed to show any such thing. Your "type of niche" is broad enough to include both smart and stupid animals, which no indication that the smart ones were inevitable, much less human intelligence. Nor do I think you have shown that the particular "niche", active animals, is inevitable either.

 To lay the groundwork I will first return to your assertion
 that all groups should be increasing in intelligence in
 response to an increasingly complex environment:

That wasn't my assertion. It was yours. You said that increases in intelligence were inevitable if the environment became more complex. I disagreed. You list below is a refutation of your original claim.

 Basically, this isn't necessary, there are lots of other ways
 of responding to complexity like:

 - hide in the mud or dirt (worms, bivalves, many insects)
 - breed so fast that you reproduce faster than you're eaten
 - organize yourself so that if most of you is eaten, you keep
   growing anyway (plants)
 - remain so small that it's hard to find you
 - grow too big for anything to eat you
 - develop a hard to penetrate shell
 - become completely inedible (starfish)
 - develop a better digestive system so you can eat a larger
   variety of things that are trying to be inedible
 - developing a really good immune system
 - become poisonous
 - develop immunity to poisons, which among other things will
   often allow you to eat poisonous plants and animals
 - become really fast

 and so on.

 Some combination of other strategies like the ones I have
 mentioned above will frequently make much intelligence
 unnecessary and often completely useless.  With the possible
 exception of an immune system all the above appear easier to
 achieve than intelligence, so most life forms are going to
 "choose" one or more of the alternatives in most situations.

 I hope that deals with your {everything should be getting
 smarter} objection.  If so, it remains for me to show what
 specific environments strongly favor the development of
 intelligence.

 I think that can be seen just by considering what most
 intelligent animals have in common:

 1 they are often wandering around in the open.
 2 they eat a wide variety of foods.
 3 they are themselves good to eat.
 4 they are almost continuously active.

 There are quite a few complicating exceptions and probably
 other and better ways of organizing this list, but I am hoping
 you get the idea.

 The first two characteristics are both desirable and the third
 is (I believe) a biproduct of needs created by the first two.
 The forth I don't see as important and results (I believe) from
 the need to feed the demanding brain.

 The main driver is the benefit of being able to eat a wide
 range of foods, or at least switch quickly when conditions
 change.  (I note that Steven L mentioned this in one of his
 posts too.)  Being able to switch food sources is obviously a
 very good hedge against short term shortages as well as long
 term extinction.

 As a practical matter an animal can rarely use multiple food
 sources unless it is highly mobile, and in many cases you must
 be among the most mobile animals about or somebody else is
 going to eat your lunch before you, or often lunch is going to
 run away.

 The high mobility requirement usually implies that you are good
 to eat because you need things like flexible muscles to achieve
 that mobility.

 Thus, I contend, that there are clear practical reasons why
 good-to-eat mobile creatures will arise in any ecology
 containing multicellular organisms.

 But obviously, if you are mobile and good eats, some of those
 other mobile creatures are going to start eating you.  So you
 will need to augment your senses with software filters (brains)
 that allow you to reliably detect and then take the actions
 needed to evade your predators.  Which will require your
 predators to ... and you know better than I how an arms race
 works.

 OK, now I think I've got myself to the point where I have shown
 that the development of intelligence is at least an expected
 (and nearly inevitable) consequence of evolution following
 multicellularity.

No, I don't think you have. There is no particular reason why we would expect multicellular animals to have evolved. Multicellularity is rare enough, but mobile, multicellular animals are a single event. I will agree that given such animals, a nervous system is a likely eventual development, and a brain is also likely. But I see no reason why intelligence follows, unless you have set the bar very low.

And in fact a great many groups of animals have brains. But note that you don't need that good a brain, in human terms, do do what you discuss. Your description would fit the world just fine if the smartest animals around were all arthropods, and there were no vertebrates. It would seem that even reptilian-level intelligence is hardly inevitable.

Even if we limit ourself to land vertebrates, note that highly active, behaviorally flexible sorts have existed since the Permian at least. And yet human-level intelligence has arisen only once in 300 million years. That hardly seems a likely event, much less an inevitable one.

 I think it is also useful to try to account for the bumpiness
 (the starts and stops in the arms race leading to increasing
 intelligence).  For me the cause of this bumpiness seems
 obvious:

 At root, brains are really expensive to maintain.  We
 continuously use about 20% of our rest energy to maintain our
 brains.  Obviously, no animal can afford an energy sink this
 large unless it is continuously at least paying for itself, in
 resource acquisition and/or predator avoidance.  Thus static
 brain size (or even shrinkage if efficiencies are "discovered")
 will be the norm over most evolutionary time.  This will be
 reinforced by the use of metabolically cheaper solutions like
 quills, improved hearing etc.

 Thus, major improvements (especially those involving brain size
 increases) will only occur when a species has a big problem (or
 perhaps a big opportunity) that cannot be handled in another
 cheaper (and findable) way.  Human evolution during the last 3
 million years seems to present a pretty obvious case in point:

 There we were, out on the savannas with the protective forests
 dying all around us.  Slow moving lunch for large predators,
 and much of what we'd like to eat could run faster than us and
 so on.  The quickest fix was an increase in brain size to
 facilitate more effective cooperative behaviour together with
 tool making and so on.  No metabolically cheaper evolutionary
 solution (apart from extinction) appears to have been
 available, because humans were so much slower and poorly
 defended than everything else in that environment.

That's certainly a customary and facile explanation for why we have such big brains. I have no idea if it's true, or how one would go about testing the hypothesis. Nor do I think it makes much sense to suppose that we are slow and poorly defended. Best if you realize that the reasons we got to our current state are speculative.

 Once the predators and food supplies were no longer an
 overwhelming problem, we got progressively better at
 exterminating each other, which helped keep the brain growth
 process moving along.  (I know about the sexual selection model
 but am skeptical of it.)

 Although I can think of no way of proving (or even
 demonstrating) it, I suspect that brain size increase didn't
 quickly top out in humans, because each incremental increase
 was more than compensated for in terms of the additional
 resources that could be accessed to fuel the larger brain.
 (The big brainers could always out-compete their smaller
 cousins)

 So basically, I am pitching the idea that abrupt increases in
 brain size are caused by periodic environmental
 changes/problems that cannot readily be solved by other means.

That seems a meaningless statement. Increases in brain size are caused by environments in which increased brain size is advantageous. You don't seem to have characterized such environments in any significant way.

It's very hard to generalize from a single example, which is
what all these probability calculations have to do. And
clearly the diffusion model is wrong in detail. We have two
main departures: pre-adaptation and incumbency.

Some innovations are impossible except in a background of
particular, previous innovations. So we can't talk about
intelligence until we have a multicellular animal with a
nervous system. Human-level intelligence must arise through a
series of adaptations of varying probability. We couldn't
possibly have expected it to happen until the evolution of
bilaterians. Which happened only once, and so may be
considered unlikely by the only guide we have. After that
crucial event, it took another half billion years or more to
get us; again, doesn't seem a likely thing.

 You have previously argued that multicellularity (another
 pre-adaptation) may have been a one-off, so I'd like to briefly
 address that before turning to bilateralism.

No I haven't. There are several instances. What I have argued is that it seems to require something that eukaryotes, or perhaps a restricted group within eukaryotes have, though I don't know what that is. Anyway, somewhere, eukaryotes evolved something that made multicellularity more likely, though still not all that likely.

 There are lots of ways of looking at this question, but all of
 them boil down to competition forcing innovation.  We are
 pretty sure that eukaryotes themselves arose in large part as
 an adaptation to parasitism where a host worked out a
 cooperative arrangement with some pre-organelle parasites,

No, we are sure of no such thing. A simple alternative is a failure to digest some of one's food, which turns out to be advantageous.

 and
 over a long enough period of time weird combinations like that
 are going to lead to innovations which will eventually lead to
 things like multicellularity and sexual replication and so
 forth.

I see no reason why organelles would lead to any of that.

 Multicellularity looks particularly likely to arise because of
 bacterial mats, which although they are not single organisms,
 their members do frequently cooperate together in ways that mimic
 multi-cellular organisms in order to monopolize some resource
 like sunlight.  Thus, there is pressure on (or an opportunity
 for) some other organism to become truly multi-cellular in
 order to compete more effectively with the mats.

This seems odd to credit bacterial mats for the rise of multicellularity, since mats are billions of years old.

 Thus I think that something eukaryote like would have arisen
 eventually and multicellularity thereafter, although I must
 admit that I don't really understand what special properties
 eukaryotes have that make them unique candidates for
 multicellularity.

Nor do I. I merely know that no prokaryote has achieved much in that department, so I infer that there is something.

 Bilateralism I see completely differently.  Just as no sea
 animal can crawl out onto land to begin evolving air breathing
 lungs to function on land, without being eaten by a raccoon,
 and likewise no other species can evolve our level of
 intelligence (unless we want it to); bilaterals also have a
 huge advantage in incumbency.

 To save space and effort, I won't make the argument here, but
 it seems obvious to me that evolution was extremely rapid
 following the introduction of bilateralism, giving rise almost
 immediately to Arthropoda, Vertebrata, and Mollusca thereby
 taking immediate possession of all available niches.  Thus it
 happened only once because it was such a good idea.

That may be obvious to you, but it's not obvious to me.

 Also since
 we see a failed effort at trilateralism in the fossil record
 (the name of a famous university is struggling to rise into my
 consciousness), other attempts at bilateralism would surely
 have occurred as variants of that.

I have no idea what you mean there, unless it's Tribrachidium.

Incumbency would argue in the opposite direction. Perhaps the
presence of a group with a particular innovation fills up that
slot and prevents any other group from achieving it. This
certainly happens sometimes. Maybe the otters are just raring
to start chipping stone tools, but we keep them from it. This
too seems unlikely, since we are only a recent development,
and otters have had plenty of time to try it before we showed
up. Hey, we've only been in the Americas for 15,000 years or
so. Where are the American intelligent species? So incumbency,
in this case, doesn't seem to be a credible factor.

 I don't agree with this logic at all.  Secondarily, the Asian
 super continent (including Africa and India) is a big place with
 many competing species that can drive innovation in one
 another.  In smaller locations with less complex environments
 and lower variation, like islands; useful innovations (like
 flight) are more likely to be lost than generated.

Really? It would seem to me that if there were indeed reduced competition, useful innovations would be less likely to be lost. The reason flight is lost on islands is of course because it isn't advantageous; however, flight was not lost in North or South America.

 North and South America (which must be treated separately since
 they only joined up recently) and Australia may add variation
 but won't produce the most effective competitors.  Thus, I
 think the Asian super continent is the place where a major new
 innovation (like advanced intelligence) was most likely to occur
 first.

This is Old World chauvinism, nothing more. There's been plenty of innovation in the Americas, some of which has managed to be exported to the Old World. I should also point out that you are cheating by making Africa part of Asia, as it hasn't been connected until fairly recently, though longer than North and South America.

 More importantly, the preconditions for the development of
 advanced intelligence have been in place for maybe 150 million
 years.

Why that time period, especially? I would have gone with twice that long.

 So the 3 million years we took to make the move
 represents 2% of that time slot.  Thus at any particular
 moment, any particular candidate group (say Otters) had no more
 than a 1/50 shot at the top position.

I don't see the logic behind that.

 Since we wiped out lots
 of other species (probably including at least two other hominan
 subspecies) in our climb, it is clear that we aren't gonna let
 no dolphins muscle in on our territory now that we got it.

There is no good evidence that we wiped out any other hominids, though it's certainly possible.

 If the Otters had beaten us to the punch, we would now be at
 least as screwed as the dolphins now are.


 Thus in summary:

 Users of Intelligence occupy specific types of niches featuring
 mobility and flexible eating habits. To effectively occupy this
 type of niche an animal pretty much must have intelligence.

This is too vague to be useful. What constitutes intelligence? There are plenty of animals with flexible eating habits that are less intelligent than other animals with less flexible habits. The Oligocene increase in mammal average intelligence was fueled by an arms race between carnivores and herbivores, both of which got smarter. It wasn't flexibility in eating habits so much as in behavior affecting that particular conflict.

 Creatures occupying this type of niche frequently get into arms
 races with each other, which sometimes involves intelligence
 itself, which will then increase unless a less costly
 alternative adaptation is available.

No, I don't think so. I know of only one such arms race, the one featuring Oligocene mammals.

 Since these arms races are periodic increases in brain capacity
 follows a bumpy path.

 On the deeper issue of: is intelligence a likely outcome of any
 arbitrary evolutionary process beginning from single celled
 creatures:

 Well obviously it cannot evolve in an environment where
 complexity and thus competition are strongly constrained, as for
 example, in the subsoil of Mars.

 However, in any environment where multi-dimensional arms races
 can get going, innovations for prerequisites like multicellularity
 and bilateralism are close to inevitable.

That just doesn't follow. Now in fact there does seem to have been an arms race that we know of as the Cambrian explosion, but it would also seem to have been waged between predators and prey largely in the medium of armor. It may have involved brains too, but there doesn't seem to be a good way to tell.

 I hope I haven't made you sick, but I can never guess with you.

You haven't made me sick, but I don't think you have made your point. It's all just special pleading, in my opinion.