Consciousness has many aspects, from experience to wakefulness to self-awareness. One aspect is imagination: our minds can conjure up multiple hypothetical futures to help us decide which choices we should make. Where did that ability come from? Today's guest, Malcolm MacIver, pinpoints an important transition in the evolution of consciousness to when fish first climbed on to land, and could suddenly see much farther, which in turn made it advantageous to plan further in advance. If this idea is true, it might help us understand some of the abilities and limitations of our cognitive capacities, with potentially important ramifications for our future as a species.
Support Mindscape on Patreon or Paypal.
Malcolm MacIver received his Ph.D. in neuroscience in 2001 from the University of Illinois and the Beckman Institute of Advanced Science and Technology. (This was after an unconventional childhood where he dropped out of school at age 9 and later talked his way into a community college program.) He is currently a professor of Mechanical Engineering, Biomedical Engineering, and Neurobiology at Northwestern University. In 2009 he was awarded the Presidential Early Career Award for Science and Engineering.
Click to Show Episode Transcript
0:00:00 Sean Carroll: Hello everyone and welcome to The Mindscape Podcast. I'm your host, Sean Carroll. Those of you who have read my book The Big Picture will remember that I talk a lot about consciousness, not so much how consciousness works, but the fact that it doesn't need any mystical spooky stuff to go ahead and eventually explain it. We don't understand the explanation yet, but we have every reason to think that we'll get there in perfectly, physical, natural terms. So one of the things you have to establish, if that's an argument you're gonna make, even though we don't yet have the explanation, is that it's possible to imagine a purely naturalistic explanation. And part of that is understanding how consciousness could have arisen. And of course, consciousness is an incredibly complicated, multifaceted thing. So it's not one simple answer. There are stages along the way.
0:00:48 SC: So I talked in The Big Picture about an example given by today's guest Malcolm MacIver, who's a professor at Northwestern University, about one of the steps that conscious creatures might have gone on. And in particular, when fish climbed onto land. Now, obviously, there's a lot of physiological changes when you go from being an aquatic animal to a land-based animal, but MacIver claims that there's also intellectual changes. There's changes in how you think, and it's based on the idea that how we think is strongly influenced by how we sense, which is in turn related, if you want to... The podcast I did with Lisa Aziz-Zadeh on embodied cognition, where we talked about how we think is influenced by our bodies in general. Here on today's podcast, we're gonna be thinking about how thinking is influenced by how we sense the world. So whether it's through our eyes or our ears, or in the case of certain fish, they have electrical impulses that give them a handle on the world around them.
0:01:48 SC: So MacIver argues that the transition of climbing onto land gave not only a different way of thinking about the world and looking at the world, but a different mode of imagination. So not only could you see much further once you've climbed onto land, but that fact about your sensorium led directly to an evolutionary change in how you could compete for food and other resources. Namely, you could plan ahead. Suddenly there is an evolutionary pressure to develop a capacity for imagination which is crucial on the road to consciousness. So in today's podcast, we dive into this idea. We talk about how we sense in general, how we think, and then the implications of this understanding for how we should think, how we can better improve our actual cognitive capacities both as individuals and as a species, as societies, how we can plan for the future. We are stuck with the brains that evolution gave us. Can we do better than that? So it's a wonderful conversation mixing ideas from science and philosophy with real implications for how we live in the world.
0:02:56 SC: Let me also mention, which I like to do occasionally, that there's a webpage for The Mindscape Podcast. I have a feeling that a lot of listeners never go to the webpage. That's perfectly fine. But if you do, you'll find not only show notes and links to important things but entire transcripts of the episodes. Those transcripts are of course paid for by our supporters on Patreon and PayPal. Also on the webpage you can find links to become a supporter yourself. Throw some cash at The Mindscape Podcast. So the webpage is preposterousuniverse.com, which is my website in general, slash podcast. We'd love to see you there. There's also little discussions, you can leave comments in all the episodes. And with that, let's go.
[music]
0:03:55 SC: Malcolm MacIver, welcome to The Mindscape Podcast.
0:03:57 Malcom MacIver: Great to be here.
0:03:58 SC: So we have a lot of ground to cover, from fish swimming around to planning for the future of humanity and... But you're a somewhat complicated guy, and you're a wonderful role model for a successful interdisciplinary career being as you are a professor in engineering departments and yet working on all sorts of things. So maybe in contrast to some other people we might have on the show, why don't we talk a little bit about how you got here. What was your academic path that led you to where you are now?
0:04:27 MM: Oh, okay. It's a very tortured path. I guess I began... First of all, I didn't go to a formal school. I was sort of self-taught. Eventually got myself into community college in Canada and used that to get into university. And there essentially I was curious about everything, and I immediately figured out if I did a philosophy degree, I could sort of get to do everything.
0:04:53 SC: Like everything. Yeah.
0:04:54 MM: Yeah. But then two years into my degree, I started getting quite worried that I might not be able to find a job. So I added a major in Computer Science to my degree.
0:05:03 SC: Oh, very smart on both counts. That's good. [chuckle]
0:05:05 MM: And something very interesting happened. I started taking classes in Philosophy and Computer Science which were talking about the same things like Turing machines, and philosophy of mind, and the nature of computation. And that got me very excited because suddenly the thing that I thought I was doing for a bread and a meal ticket was actually intellectually also satisfying for me. And so that sort of led to this whole line in CogSci and AI that I pursued for a while. I started a two-year... I did a Masters in Philosophy of Science, did a two-year stint of a four-year program, PhD program in CogSci and philosophy of mind at Indiana University which is actually where I was classmates with Dave Chalmers and Anthony Tromero and some other people who've gone on to do really well. But two years into that I decided I was getting kinda frustrated just thinking about the nature of the mind and thinking about the nature of computation and wanting to do something with my hands as it were and decided to re-apply to grad school and get into neuroscience. And that's eventually led to this program at Urbana, University of Illinois in Urbana, where I did essentially neuro-ethology, but with a strong computational bent to it.
0:06:21 SC: You have to tell me what neuro-ethology is.
0:06:22 MM: Okay. So it's the study of the neural basis of natural behaviors. It has a very long and ancient, maybe not ancient tradition, but old tradition of essentially the marriage of ethology and neurophysiology.
0:06:40 SC: I don't even know what ethology is.
0:06:41 MM: The study of animal behavior. Tinbergen...
0:06:42 SC: Animal behavior? Okay.
0:06:43 MM: Looking at birds rolling eggs, Konrad Lorenz, these are famous ethologists.
0:06:49 SC: I know those names, I didn't know the whole Greek word. Ethology.
0:06:52 MM: Yeah, yeah. So...
0:06:55 SC: E, just the letter E not AE or anything crazy like that?
0:06:57 MM: Ethology. Just ethology. So, those two fields got together in neuro-ethology and the pursuit of neuro-ethology is one of looking at things like predator-prey behaviors, any kind of behaviors you'd see in the natural world, so rather than bringing an animal into the lab and do what is effective for creating scientific measurements, repeatability, you take what you get in nature, or you try to recreate it in the lab. And so that's... There's a whole tradition there. And that's what I pursued, except it was a lab run by a former high-energy physicist who was at Caltech actually did a lot of work in quirks, Martin Nelson is his name.
0:07:49 SC: So many people are former energy physicists, that's hilarious. [laughter]
0:07:52 MM: So as a result, it was a really interesting mix of both animal behavior and computational modeling and really highly quantitative neurophysiology, so it was a really great training for me.
0:08:06 SC: Okay, and none of this is engineering?
0:08:08 MM: None of this is engineering. So, where engineering came in is at... A little, the last few years of my PhD, I was getting more into devising... Actually, we devised one of the very first, perhaps the first, robot that used the same principles as a weak electric fish, to sense objects.
0:08:28 SC: Weak electric fish are an actual kind of fish.
0:08:31 MM: Yeah, actual kind of fish. They don't get plugged in, they are real animals that essentially emit a small electric field and detect objects. Sort of like a bat uses sonar to detect objects. So that's what my PhD was on, it was on characterizing their sensory system and modeling natural behavior, pre-capture behavior in electric fish, as they hunted their favorite prey which happened in this particular species to be water fleas. And so... But near the end of my PhD, I got into building robots that duplicated some aspect of their capability. In this particular case, it was how they use electric fields to detect objects and identify objects.
0:09:15 SC: The kind of the deep, the underlying theme here is trying to figure out how thinking or cognition relates to the body and the biology? Is there a theme?
0:09:28 MM: Well, the theme of, I guess, neuro-ethology is the most effective angle into understanding neural circuitry, which is a lot more conserved across animals than we previously understood, but... So you can study a fish, you can study even insects and get a lot of insight into neural organization in general. So the idea is to use behaviors, which were highly selected for as levers into the circuitry that supports those behaviors. And so that's really what electric fish, as a field, is sort of oriented around, and the great thing about electric fish, which is why my advisor and many other people with quantitative or physics or electrical engineering backgrounds have gotten into them, is that whereas with most sensory systems, for example, for a whole faction it'd be very hard to give a step impulse. So you can't really go from no odor to full odor to no odor. There's just all this complex mixing going on. With electric fields, when you're working on electric fish, it's trivial to take their field, add a sinusoidal modulation or a square wave modulation, and then look at what their brain is doing as a result. So you can apply all of these techniques from electrical engineering and systems analysis theory to electric fish, with very little difficulty.
0:10:56 SC: How big... So these are real fish. How big is an electric fish? An adult?
0:11:00 MM: The ones that are typically studied in the lab. The apteronotids are 15-30 centimeters long.
0:11:06 SC: Okay. And do you... Is there a vibrant black market in electric fish. Do you you buy them at the fish market?
0:11:12 MM: Well, what's funny is that they're actually popular in the tropical fish market, and a lot of people get them not even knowing that they are electric fish. They have a bizarre body form because they don't... Most fish swim by flapping their tails. These guys have undulating membranes at the bottom of their bodies. We call that the ventral fin. And they undulate that sort of like a curtain to move forward and backward, and frequently you can't tell what direction they're swimming, and they swim equally agilely in both directions.
0:11:41 SC: But it's not electric eel. That's something different.
0:11:43 MM: No. Electric eels are related. They're actually in a close phylogenetic relationship to the ones I'm telling you about, but they have evolved a much stronger electric organ which can put out 600 volts at an app to stun their prey, and then they can have a leisurely time going around nibbling up the things that are stunned.
0:12:05 SC: So, the eel is using electricity as a weapon.
0:12:08 MM: As a weapon.
0:12:09 SC: The electric fish is using it as its sense organ.
0:12:09 MM: They also have... Yeah, and they also have the low voltage type to do sensing with.
0:12:14 SC: Okay.
0:12:14 MM: But the weakly electric fish only have the low voltage.
0:12:18 SC: So, you get this electric fish, you bring them to your lab, and you put them in a tank, and you zap them and see what happens?
0:12:24 MM: Yeah. You, depending on the experiment, I did a lot of behavior with them. We also did primary afferent neurophysiology, where you're actually modulating their electric field, you take their field, you add some sort of change to it that simulates then swimming by a rock or a plant, and you feed that back into the tank.
0:12:42 SC: Okay. And so you're tricking them. It's a brain in a vat. It's a fish in a vat.
0:12:45 MM: Yeah. It totally is, which is, it's really remarkable. You can do virtual reality with very little difficulty with an electric fish because they are sensing electric signals which you can modulate with perfect control all over the body.
0:13:00 SC: And is that their primary input? Or do they also have sight and taste?
0:13:04 MM: They have eyes. They're often blurry and cataract by the time we get them. They don't seem to use them very much. The reports are the rivers that they hunt in are Turbot. So, even in the day time, you can't see more than a few centimeters through them.
0:13:20 SC: Right.
0:13:21 MM: So it looks like they're mostly used for, is it safe to come out yet? Is it dim enough to come out and hunt at night? They hunt at night. Or it's a time to go hide in the weeds?
0:13:31 SC: Right. 'Cause their prey are just little bugs.
0:13:33 MM: Exactly.
0:13:33 SC: And they wanna stay away from the...
0:13:34 MM: Although there's a whole buzzing electric civilization of 200 species, and many of them will hunt other...
0:13:39 SC: What? No one told me this.
0:13:41 MM: They will hunt other electric fish. In fact, many of the fish we get in the lab have several inches of their tail missing because they've been nipped and they have... Their tail is where their electric field is strongest, so it's almost like a lure. And they happen to have evolved a way to regenerate their spinal cord in that area, which has been a subject of much interest because vertebrates, regeneration of the spinal cord is, obviously, an important ability.
0:14:10 SC: But there's an electric arms race among several different species?
0:14:12 MM: Yes, yes. It totally is an arms race with many different...
0:14:15 SC: Underwater, almost invisible.
0:14:16 MM: Underwater, yes. The whole pulses went from a uniphasic to biphasic to all these different modes of signaling for sexual speciation. It's an amazing story.
0:14:32 SC: Do we know how recently this developed in evolutionary biology?
0:14:35 MM: I think it was, if memory serves, it was a around 100 million years ago that they branched off of the non-electric osteoglossomorphs, I think it was is the... I should know this better being now...
0:14:49 SC: We don't care. We're not gonna judge you, yes. [laughter] Okay. So, how did you hornswoggle Northwestern university into thinking that this is engineering somehow?
0:14:57 MM: Well, so I came to actually here at Caltech, where we're recording, and I worked with Joel Burdick who's a great roboticist here, and we did really fun work now on making a robotic replica of the locomotion system, which turns out to have all kinds of fascinating properties.
0:15:18 SC: The locomotion system of the fish?
0:15:20 MM: Of the fish. So, this rib and fin locomotion system, which gives them bi-directional motion, forward backward motion with amazing agility. And so we built a robot here, several robots here, that replicated parts of, well, the entire rib and fin, essentially. We had a scale model. I remember this. The first one we did was with wheels and servo motors, and the last one we did was quite long. It was about a meter long. We had it in a flow tank in the basement of the mechanical engineering building and did a variety of fluid studies of it.
0:16:02 SC: It seems that having electric fields as your primary sense work is a weird thing in the animal kingdom.
0:16:09 MM: Yes.
0:16:10 SC: And also having this locomotion is a weird thing amongst the underwater dwelling things. Is there a relationship between these two rare effects?
0:16:16 MM: Yes. There is a very tight relationship. And so the way it goes is these are animals which will... Typically, when they're hunting, you'll see them hunt backwards, you'll see them hunt forwards, there's a great advantage to hunting backwards, to moving backwards while hunting. In that by the time the prey or object of interest has gotten to their mouth, they've scanned it with their electro-sensory system. So because it's a really near-field sense, you kinda need to drag it across the receptors of the body before you can identify. So if they're swimming backwards, that's really easy. They've scanned it by the time it gets to their head. If they're swimming forwards, what they do is they do this sort of... They'll rotate or something to get closer as they're swimming forward, then they'll do a rapid corkscrew back, that's sort of back to get the object if it's something that they want to eat.
0:17:13 SC: So the relationship is that this electric sense is very useful, but it's short range.
0:17:19 MM: It's short range, and if they were to try to... If they only had a big tail fin, and they had to go forward and circle around, they would've lost the prey, because they can... So I spent much of my PhD quantifying exactly how far they can sense their objects of interest out to. And for a 15 centimeter fish, it's something on the order of 3.5 centimeters, for a millimeter size prey. So they're hunting really... These particular species that we're talking about, which is Apteronotus Albifrons, are hunting really... Or the Black Ghost, much sexier name, Black Ghost Knife Fish, hunt these millimeter-sized water fleas. And so they can only sense those 3.5 centimeters or so away. And so if they were to swim forward and circle back, they would've totally lost their...
0:18:07 SC: Still lost. Game over.
0:18:08 MM: Yeah.
0:18:08 SC: Now, isn't this obvious evidence in favor of intelligent design? The fact that the fish have built this locomotive mode and the electric fields complete swimming mechanism?
0:18:15 MM: Completely... Yes. Absolutely.
0:18:17 SC: Or do we have the evolutionary understanding of which came first?
0:18:20 MM: That's a good question. I don't know if we know from the fossil record which came first. I suspect because undulatory swimming is actually not that uncommon, it wouldn't surprise me if that might've come first, in fact there are these non-electric fish, like Xenomystus Nigri, which has a ribbon fin, and they use passive electric sense but don't actively generate it. But, again, passive electro-sense is a super short-range sense. And so having a ribbon fin kind of goes with that. It may not be active electro-sense dependent, but it might've co-evolved with either active or passive, probably passive first, I think that's the ancestral condition.
0:19:06 SC: Okay, so you worked on the locomotion aspect of the electric fish, as a post-doc here at Caltech, and that does sound a little bit more engineering-y to me, right, how things move in weird circumstances.
0:19:17 MM: Right, right, right. And so what I was doing with Joel is trying to figure out what roboticists called the small time reachable set of the animal, which is given...
0:19:26 SC: What do human beings call that?
0:19:27 MM: Which is essentially where you can get to given any feasible control input. So for a car, you can go forward and backward, initially, and with some turning of your wheels, you can kind of get an hour glass shape, and then you can do parallel parking and such, but you can see how that shape evolves in your mind's eye as you get more time, from kind of an hour glass shape to more of something like a circle. And so I was curious what this would be for electric fish, since this is actually something that was actively being worked on by some kind of geometric differential geometry people who are interested in the problem.
0:20:08 SC: So some mathematicians pushing conductor fields.
0:20:10 MM: Exactly, exactly.
0:20:11 SC: Who are interested in these fish.
0:20:12 MM: Yeah, well they were originally doing it for carangiform, the tail-flapping type swimmers, and I wanted to do it for the ribbon-fin swimmers. And that got me into, oh well, this is a motor volume or as I now call it, or small-time reachable set that has this very interesting geometry at small time scales that changes at larger time scales, and it has this very interesting relationship to the sensory sensorium of the fish.
0:20:39 SC: I just want to pause here because it is kind of amazing to me the relationship between motion, maybe it's not that surprising when I put it in these words, the relationship between the motion of objects and organisms and differential geometry and higher mathematics. I remember vividly, there's this wonderful little book on differential geometry by Walter Burke, where he gives an explanation for why people are... Find it difficult to get an intuitive grasp of parallel parking. He says it's because it involves the non-commutation of vector fields.
0:21:12 MM: Yeah, V-brackets.
0:21:13 SC: Yeah, exactly. Because you have the operation of moving forward or backward, you have the operation of turning your wheels right or left, and they don't commute, it really matters which you do first. And if you understand that perfectly well, it helps you parallel park.
0:21:24 MM: Exactly. Yes.
0:21:25 SC: And so I can see why if you had different modes of moving, the differential geometries would become interested in that.
0:21:31 MM: Yeah, yeah.
0:21:32 SC: And so that got you interested in differential geometry or...
0:21:36 MM: It got me very interested in differential geometry, although I didn't pursue it too far. At about that time I got this... Multiple people emailed me this sort of job that was opening up at Northwestern, that was looking for a neurobiologist also trained in engineering and...
0:21:54 SC: I see. Apply.
0:21:55 MM: So I cut my... [chuckle]
0:21:56 MM: I cut my postdoc about a year short, just as I was getting into the more arcane and interesting aspects of the differential geometry approach to these animals so, but it did get me interested in how to understand the relationship between sensory spaces, the sensorial of the animal and their locomotion ability, which has led to many interesting avenues subsequently.
0:22:25 SC: Right, so once you're at Northwestern, you're still studying the electrical fish?
0:22:28 MM: Yeah.
0:22:28 SC: But you become... It sounds like from the story that you had some initial interest in thinking how the nervous system was working.
0:22:38 MM: Exactly.
0:22:38 SC: Then you became interested in the sort of more mechanical.
0:22:40 MM: Exactly.
0:22:40 SC: Locomotive aspect. And then, now you're a faculty member, you can do whatever you want.
0:22:44 MM: Yeah.
0:22:44 SC: And so you start thinking more about thinking in the nervous system.
0:22:47 MM: Right, Yeah, yeah. And so the way that went was roughly speaking I... So the last bit of my postdoc was this paper characterizing exactly how the sensory and motor volumes coordinate over different time scales. And at the end of this work, I decided it would be fun to look at how visually-guided animals underwater, how they coordinated their sensory spaces with their locomotor spaces. So in particular...
0:23:21 SC: So sorry, you'd figured out some relationship between this very special weird thing?
0:23:26 MM: Yeah.
0:23:26 SC: This beastie electrical fish.
0:23:28 MM: Yes. Yeah.
0:23:28 SC: They can sense a certain distance around it using electric fields, and you've related to this weird locomotive thing it has.
0:23:34 MM: Precisely.
0:23:35 SC: So you wanna know... Okay, what about the non-weird stuff, what is that relationship?
0:23:38 MM: Well, right. And so just to quickly summarize the electric fish's sensory volume looks like a cylinder around their body, okay, and their small time reachable set or their motor volume is essentially cylinder as well, and how they get that cylinder...
0:23:54 SC: So they probably not...
0:23:56 MM: Out of their actuator is a really interesting problem, which we solved, but I won't go into the details. But then the question is for an animal that looks forward, and so now imagine in your mind's eye, sort of a pie-shaped wedge of space in front of them, and then they have a tail flapping system that only pushes them forward into that space, into that wedge of space. And they can yaw or turn left and right, roughly proportional, somewhat related to their, the angle of that pie, so pie slice. And so I was interested in characterizing how those two things coordinated in a visually-guided animal, and so we did the calculations and some estimates and turned out that there's... I expected, I guess, to see something, we walk around with this giant sensory volume ourselves, visual sensory volume, and my intuition was that that visually-guided fish underwater would have this giant pie wedge thing that they would be moving around looking for their prey, and it turned out to be really pathetically small, and so small that I thought there were there was an error in the calculation.
0:25:00 SC: Sorry, which one was small, the sensorium?
0:25:02 MM: The length of the pie. So how far out...
0:25:04 SC: Which pie? The sensorium pie or the locomotion pie?
0:25:07 MM: The sensorium pie.
0:25:08 SC: Okay.
0:25:08 MM: So these visually-guided fish hunting for small bugs don't see them very far ahead, and it was such a small distance ahead I thought there was a problem with the calculations, but eventually...
0:25:19 SC: So they can move very quickly.
0:25:20 MM: Yes.
0:25:20 SC: They can get far, but they can't see far.
0:25:22 MM: Yes. Correct. And I thought this can't be right because in fact, what matters for all of these animals is not the static volume but the sweat volume through time so as you... Because they're searching for empty space or relatively empty space for food for live prey, and it turns out the sweat volume for the electric fish was almost identical to this pie wedge shaped sweat volume that the visually-guided fish have.
0:25:53 SC: The shape was different, but the volume was the same.
0:25:54 MM: Shape was different, but volume was almost the same. Okay, now electric fish have to pay for every joule of the energy that they're investing in electric field, whereas visually-guided fish are parasitizing the sun's energy. So there's a difference there in the bioenergetics. But nonetheless...
0:26:09 SC: So it can't be a coincidence that these two very different organisms using two very different mechanisms have evolved the same sensory volume.
0:26:18 MM: It was bizarrely similar in size to me. But so eventually I nailed down that it's simply due to the attenuation length of light and water, so it turns out that water is a terrible medium for seeing with light.
0:26:32 SC: For eyeballs, yeah.
0:26:34 MM: And I just didn't... I guess people who do scuba diving and work in submarines can tell you this from their experience that suddenly things appear out of nowhere, but you know, I didn't have that experience but so, so the numbers showed really clearly, and you lose a huge amount of information on the order of meters, and that's imperfectly like laboratory clear water. It's not a function of turbidity where if you add any particles in the water, then your range really goes down to nothing. So that led me to... I was at that time getting interested in what happened to vertebrates at the water-land transition in the upper Devonian. Why did fish evolve limbs and all the other things they needed to evolve to come up on the land, you know what was the basis of that?
0:27:26 MM: And it occurred to me that vision must have changed greatly after the water-land transition, because suddenly you're going from a medium which is like a huge thick blanket on your sensorium to essentially infinity. You can see a distance further than is in most cases behaviorally useful. [chuckle]
0:27:47 SC: We can see the moon, right? It's not that hard.
0:27:48 MM: Yeah, you can see the moon, yeah.
0:27:49 SC: It's obviously a shock.
0:27:49 MM: You're not going to hunt for prey on the moon. [chuckle]
0:27:52 SC: Right, so what are you gonna do with that knowledge?
0:27:56 MM: So what I was curious about was the eyes of the ancestral aquatic vertebrates were pretty small, and I expected that they got as big as would be useful in this highly attenuating medium, and you can calculate it out, and we've done that. And they are about as big as they should be.
0:28:15 SC: The eyes of a fish?
0:28:16 MM: Yeah. Because getting them bigger, essentially your visual range is proportional to your eye size.
0:28:21 SC: Aren't the eyes of a whale really big?
0:28:24 MM: I don't know the size of the eyes of a whale. They usually use their echolocation systems for range 'cause the attenuation length of sound in water is similar to the attenuation length of light and air. And so whales and dolphins are a really interesting example of vertebrates that got used to giant visual sensoria doing the one thing that you could do in water to regain that kind of range.
0:28:50 SC: It's very interesting, yeah.
0:28:51 MM: Which is use the high frequency sound production system that land animals evolved and do it for imaging.
0:29:00 SC: Since I know nothing about this, are there underwater animals that use like a location that are not former land animals?
0:29:08 MM: Not to my knowledge. There may be one or two species of these teleosts that use clicks. I'm fuzzy on remembering the species, but I think there might be a couple other examples. Now teleosts are very recently evolved animals, so there might be that case.
0:29:29 SC: But anyway, the ancestors of dolphins and whales who got used to this luxurious broad sensorium of seeing the whole world and they somehow cleverly figured out how to reproduce it underwater.
0:29:39 MM: Exactly. Yeah. And so now, you asked me a long time ago about how does this connect to thinking. And I'm very interested in the relationship between how far away you can sense objects, and how you modulate your behavior. So the intuition there is, just think of yourself driving rapidly, too rapidly down a foggy road, and you see something suddenly. You don't have a lot of executive planning that's gonna go into deciding what to do next. You're gonna swerve, you're gonna break. But in any case, you're gonna do something simple. And so my thought was that, it would be central to the evolution of more complex cognition to be able to sense things far enough out where it makes sense to actually have a plan. And the whole reason I was driven to do a whole bunch of analysis of what happened to vision once we had gone onto land was an interest in why it is that vertebrate brains got a lot more complex, substantially more complex once we came up on land.
0:30:45 SC: So the idea is that, under water you might imagine from our very parochial air, land-based point of view, that thinking ahead of time is just a very useful skill to have, but it also has costs, right?
0:31:00 MM: Yes.
0:31:01 SC: Energy and volume in our brains and so forth. Then you're saying that for a fish, it just doesn't help. There's no point to planning ahead too much.
0:31:08 MM: You can do it, and they do do it to some extent. It is just that it's much, much more useful once you come up on land, and energetically, it's a gift. As soon as you bend that cornea that was inappropriate after you came out of the water, you need to bend it a little bit for the different refractive index. But as soon as you fix that which can be done very rapidly, evolutionarily speaking, you now can see hundreds of body lengths away, and suddenly you go from an animal, which was trapped in kind of a reactive bubble, sort of always driving in the fog if you will, to an animal which may not have had the neural circuitry initially for anything other than reactive control to an animal, which with the right mutations might be able to stall the issuing of motor commands to the motor system and movement towards the prey to think about, "Well, if I do option A, the prey will run away, but if I do option B, the prey won't run away. So maybe I'll do option B."
0:32:09 SC: But wait a minute. There's a fraught set of words that you're using here, right? [chuckle] It's not only planning ahead, it's also using some sort of imagination.
0:32:17 MM: Absolutely.
0:32:18 SC: Some sort of ability to simultaneously contemplate different hypothetical scenarios in your head.
0:32:25 MM: Yes. Yes.
0:32:25 SC: And so you're saying that the thing that prodded evolution to develop that capability was climbing up on to the air onto land.
0:32:32 MM: Correct, yes. And so we now know the address of that system in the brain, it's the hippocampus in coordination with the frontal lobes, that you can actually now monitor animals thinking ahead. Thinking about different paths when they are given a challenging situation where...
0:32:47 SC: By putting them in FMRI machines or something?
0:32:50 MM: Well, actually the studies I'm thinking of right now are ones on what's called "vicarious trial and error in rodents", and those studies, such as out of Redish's Lab at the University of Minnesota, were done with electrodes in the brain.
0:33:06 SC: Okay.
0:33:06 MM: Looking at plate cell activity in the hippocampus. Plate cells in the hippocampus essentially are activated according to where you are in space. And what they observed is in mazes, where there was not a reliable reward location where you'd switch it, so the animal had to think about what to do next, then at choice points you can actually see the plate cells racing down one arm of the maze or the other arm of the maze, before they act.
0:33:34 SC: Wow.
0:33:34 MM: So they're pausing there, and they're thinking about different futures.
0:33:37 SC: And were these parts of the brain there in the fish and adapted?
0:33:41 MM: No. So there is a part of the dorsal pallium. So this is an area of increasing interest to me 'cause I'm certain, I'm not one to say that suddenly we had it. There had to be precursors of course. And fish have plate cells or have a sense of home, I should say, they have a sense of where they are in space, they have spatial maps, they have cognitive maps in the brain, so certainly they have some of the primitive infrastructure for this. But to do planning to do imagination, there's no evidence... We don't know of a fish being able to do this. And that part of the brain, the hippocampus, and of course frontal areas in mammals got much, much larger with land animals.
0:34:25 SC: Right. I mean I do know this work that at least claims from the FMRI studies that the part of the brain that lights up when you ask someone in an FMRI machine to imagine a future circumstance, is the same part roughly speaking, as when you imagine... Ask them to conjure up a memory of a past circumstance. So it seems to be a bit of using existing circuitry for that kind of thing.
0:34:48 MM: Yeah. Yeah. Well, that sort of... Right. That discussion comes up in certain analyses of how our parochial thought about memory is not quite right. It's not like a random access store, it's more like a recipe, which once you re-create the experiences in your sensory cortices, you sort of regenerate the experience. That's where that comes up. But this is something a little bit different. This is... There's this... The hippocampus is a very, very special whole multi-nodal structure that binds information across different modalities and does relationships really well. And that is where this imagination... At least is one core part of the imagination circuit.
0:35:28 SC: Well, that's always the problem. We use words that were invented hundreds of years ago, like imagination and of course, they have different aspects going on in the brain.
0:35:36 MM: Yeah, exactly.
0:35:36 SC: But you pinpoint this particular one and you think you can... So there's a hypothesis.
0:35:40 MM: Yes.
0:35:41 SC: That it's related... That the use of this, to do this hypothesizing and future planning is related to this transition from sea to land.
0:35:49 MM: Precisely right. So we're now... We did this analysis of what happened to vision from after the emergence from water to land. We did this... We had a hypothesis about what happened to the eyes. It turns out our initial hypothesis is wrong. We thought the eyes would blow up in size after we came up on the land. And it was much more interesting because we saw a blow up in eye size about 10 million years before we came up on the land.
0:36:16 SC: In the Fossil record?
0:36:17 MM: In the fossil record. And it turns out at the very same time, skull morphologies disclose that animals had these flattened skull shapes with these orbits way on top of the skull table and for... They look all the world... For like crocodiles. What would they have been doing? Well, it turns out there's this huge bloom of invertebrate life on land at that point. 50 million years preceding the vertebrate water to land transition.
0:36:44 SC: So the bugs and the plants were already up there?
0:36:46 MM: Exactly. So what we can then say is, whereas in the past, we always thought of if you were to ask what's the most useful body part for bringing us up on to land, you would have said, "Well, limbs," and everybody thought about the limbs. And how did the digits evolve, and how do they separate from... Well, it turns out that the eye increase preceded that stuff, like the separation of digits and all that stuff, was preceded by a big increase in eye size. So, it looks like what might have happened, something more causally interesting, which is that eye size grew up and showed all this bounty that made it worth while, clambering up on the land for. And only after that point, do you start to see things like rib cages and all these other things that you need to live on land with. And digits, separated digits, all that came after the increase in eye size.
0:37:45 SC: So some fish were lucky enough to realize that if they poke their eyes up above the water, they could see much further?
0:37:47 MM: Yeah. Yeah.
0:37:47 SC: And they got a tiny little...
0:37:49 MM: Yes.
0:37:50 SC: Epsilon advantage.
0:37:51 MM: Yes.
0:37:51 SC: And therefore their descendants...
0:37:53 MM: Yes.
0:37:54 SC: Whose eyes were more capable.
0:37:56 MM: We have a theory about that too. There are these breathing holes that are now Eustachian tubes that were right behind the eyes, and we were going through a very low oxygen period of the earth's...
0:38:11 SC: In the earth's history.
0:38:12 MM: In the earth's history, right at that point. And so we think the animals were actually, [chuckle] coming up for a breath of air.
0:38:17 SC: Oh, okay.
0:38:18 MM: And just coincidentally, would catch us.
0:38:19 SC: They realized, "Oh my goodness." [chuckle]
0:38:20 MM: Yeah. "Oh my goodness, there's this bounty of unexploited food, undefended food." And we were so excited that we actually started getting the the fossil records of the insects at the time. And you can see evidence for chemical defense systems suddenly appearing in the fossil record about the time that the vertebrates came up on the land.
0:38:38 SC: Chemical defense systems from the bugs?
0:38:40 MM: So little ozopores and millipedes have these chemical defense systems which currently are active against vertebrates predators. And they just pop up in the the fossil organ.
0:38:49 SC: So they didn't need defense before the vertebrates?
0:38:50 MM: Part of that, I guess, millipedes didn't like eating other millipedes. I'm not really sure...
0:38:53 SC: Yeah. Okay.
0:38:53 MM: But all the other bugs that were there, apparently not a system that was needed before that part of time.
0:39:00 SC: Okay. So the fish had this evolutionary advantage. Some of them figured this out.
0:39:05 MM: Yeah. Yeah.
0:39:05 SC: And there was this new pathway that opened up that there was yet another evolutionary advantage, "Oh, I can plan for the future."
0:39:13 MM: Right. So our thought is that, and now we're doing all these computational simulations. And we're almost ready to submit a manuscript that basically shows that if you have a small range, small sensory range, and this is not exactly mind-blowing stuff. You would intuit it, but if you have a small range, planning is useless. [chuckle]
0:39:38 SC: Yep.
0:39:38 MM: If you have a bigger range, planning becomes progressively more useful.
0:39:41 SC: I know plenty of people who have a small range, and they're not very good at planning, so that makes perfect sense.
0:39:46 MM: What was fascinating is, after we did those initial simulations, these were all in open-world scenario, so no clutter, is that we got to this plateau, survivability...
0:39:54 SC: So this is your simulation?
0:39:56 MM: Yeah. Simulation, or predator, or prey.
0:39:57 SC: So you do like a computer model?
0:39:58 MM: Yeah. So we have a predator and a prey. And we're using mark-out decision processes to model the sort of, the game between the predator and prey. And the prey has a thick sensory range. The predator has an infinite range. Anyway, we got to a certain point where we said, "You know, this isn't very landlike. And we're getting survivability rates for our prey that are really quite low. And it seems like they ought to do a bit better. Why don't we try putting clutter in the space, because that's what land gives you is, lots of clutter. Lots of geometry."
0:40:31 SC: It's not fun to play hide-and-seek in an empty room.
0:40:33 MM: Exactly. Yeah. So we tried that, and we saw a fascinating pattern, which was, planning became much, much more useful. But only in a restricted range of clutter, which matches terrestrial, typical terrestrial ranges. So purely open space, not very much advantage for planning. Total clutter, where all you see is jungle and forest, also, not very useful for planning. But there's this sweet spot, mid-level entropy, if you actually calculate the entropy level of the terrestrial space of the clutter. At mid-level entropies, you get this beautiful increase in the advantage of planning.
0:41:17 SC: And that's because there's much more complex structure there in the space of possibilities. And so now, there's a payoff.
0:41:22 MM: Exactly. So now, what we see is that every time the prey moves, the value of that move, and the location of that move is highly contingent on the predator's location. Because if there's an obstacle in front of it, it's got one value. And if there's open territory where the predator can just lunge, then it's got a totally different value. And so, this diversity of values that appear once you have clutter is really what is exploited by planning. And...
0:41:54 SC: Do you have any idea whether or not the first fish to come upon to land, did so in moderately cluttered environments?
0:42:00 MM: Well, so there's lots of good data now on... That forest systems were going at that time. So...
0:42:08 SC: It's plausible, anyway.
0:42:10 MM: It's plausible. Yes, for sure. I think that that's a convincing case that once you're on land in the right circumstances, planning, in some vague sense, is useful. Is it obvious that evolution had the wherewithall to take advantage of this new opportunity?
0:42:30 SC: So it's clearly the case that it did.
0:42:33 MM: Right. So [chuckle] Well, you know.
0:42:35 SC: We knew it...
0:42:36 MM: And so do mammals, many other mammals. And so do, there's a very strong behavioral evidence that many birds, as well. The neurophysiology is lagging in that group of animals. So it's there. There's no question of it.
0:42:48 SC: We do plan. We do imagine the future.
0:42:50 MM: We do imagine futures.
0:42:51 SC: Some of us better than others, but yes.
0:42:52 MM: Exactly. And so part of this is, "Well, what's the archeology of that? How did it occur? Why did it occur?" And I'm interested in that question from multiple angles. One is that, I'd like to know why are we so bad at planning where there's tight, spacial, and temporal limits to our ability to plan. It seems very, very constrained. And escaping those constraints might be incredibly useful. Or transcending them might be incredibly useful in contexts such as looming threats that are transgenerational, like climate change, or possibly, AI might be another. But in any case, threats where action has to be regulated, behavior has to be regulated across time scales that are much longer than what we are natively comfortable with our planning systems.
0:43:43 SC: Would you go so far as to relate this emergence of the imagination, if you wanna be slightly more sensationalist about it to...
0:43:51 MM: Yeah. No, it's perfect word for it.
0:43:52 SC: To consciousness?
0:43:54 MM: Yeah. So I think that there's... From my standpoint that's the most productive way of thinking of consciousness currently.
0:44:04 SC: What is the... In terms of imagination?
0:44:07 MM: In terms of imagination. So the first animal that had to examine its mental furniture in order to derive a useful behavioral program, the first animal that had to imagine.
0:44:23 SC: So wait. You're linking there... We were previously talking about planning in terms of, I could hide under the tree, or I could climb on the rock etcetera.
0:44:31 MM: Correct. Yes.
0:44:32 SC: But now there's a bit of self-awareness. Is that a necessary part of thing?
0:44:37 MM: Well, it seems kind of ineluctably connected in that. It is in your brain that you are examining these futures.
0:44:45 SC: You're part of the future.
0:44:45 MM: And you're part of that future. Of course, it's not everything that people want from the word consciousness. But I think, scientifically speaking, at least for me what is crucial is having something that is actionable scientifically. [chuckle] And from that standpoint, the idea that at least a core component here of what we mean by consciousness is this space of imagination that opens up for purely adaptive, very easily calculated... Now we can show in the simulations, why it rises or why it would arise, that it's a perfectly natural response to a challenge that is brought by the interaction of long-range sensing, visual sensing, with this dynamic context where memory is not quite sufficient, where you need to take into account moment by moment positions of your threats and act accordingly. So having that ecological scenario drives animals to evolve this system where they examine the contents of their own thoughts.
0:46:02 SC: I think that if we conjured up, your former classmate David Chalmers was here. He would cast the hard problem of consciousness in terms of what it is like to feel something ineluctably first person's subjective experience. Is that what you're addressing here, or are you sort of slightly past that, into a bit more objectively measurable?
0:46:26 MM: This is not that aspect of consciousness.
0:46:31 SC: Okay. You're still in the realm what David would call the easy problem of consciousness.
0:46:34 MM: Yes.
0:46:35 SC: Which everyone agrees is hard. [chuckle]
0:46:37 MM: Yes, yes.
0:46:37 SC: Okay. Very good.
0:46:39 MM: And in fact, as you know, and Chalmers' probably spoke of this, as in many other context at least, that there's an aspect of that, which seems completely outside of the capability of science. Because science is supposed to be intersubjective, then if this is something that has to be subjective, then it can't. So therefore the ontology has to grow, etcetera. And so yeah, but that's way beyond what I'm talking about. I'm talking about something very practical, which is the space of imagination, where we would ourselves figure in that imagination. And so that's the first moment at which our cells became, well, an item of thought.
0:47:26 SC: Right, now self-awareness is clearly an important aspect of anyone's definition of consciousness, and so that's what you're getting at.
0:47:32 MM: Yeah.
0:47:32 SC: And so I have to ask the obvious question, I'm sure everyone asks, what about the octopus? Octopus never climbed up on the land, did it?
0:47:40 MM: I just recently wrote a proposal to look at these guys.
0:47:45 SC: So we should say, octopuses are really, really smart [chuckle] in some sense.
0:47:49 MM: So everybody who looks at them carefully has told me, "Don't go so fast with that." Because, of course, you hear things about them looking at people, opening jars and then going and doing the same. But I've heard from the same people who have done a lot of work on them, that they're able to do it once, and then they can't ever do it again.
0:48:12 SC: They don't remember how to do it.
0:48:13 MM: Well, or that it was actually they got lucky or something like this. I'm not denying that it looks like they do.
0:48:21 SC: You can say octopus. [laughter] Very prejudiced against octopuses.
0:48:24 MM: They're such a thorn in my side, as soft as they are. No, I think there's something very interesting going on there. And here's a couple of thoughts. One, octopus are this, they're nudibranch mollusks, meaning they're unshelled, meaning they're unarmored. And yet, they're this delicious protein that can't swim very fast. And they're hunted by the animals with the greatest sensory range in the water, which are dolphins and whales, which have turned their sound production systems into these amazing long-range sonar systems. So they have had to survive by their wits, and furthermore, they're hunted by, I just learned from Roger Handland, who's done a lot of work on them, that they're preyed on by diving birds, who are also extremely intelligent animals and can see very long distances from above, and they typically hang out in shallow areas, and these birds dive and hunt them.
0:49:24 SC: So it's hard out here for an octopus is what you're saying?
0:49:26 MM: So it's really hard living as an octopus 'cause they don't have great escape mechanisms. They're great food. So possibly, and this seems terribly ad hoc, so I immediately grant that this may just be a completely different way to evolve, cognition, than with the land vertebrates have done, but it is interesting that they are preyed upon by some of the smartest vertebrates, and it is interesting that they have these tremendous eyes, although I would argue because vision is so useless for the most part, at least as a long-range sense in water, that vision is giving them things like acuity for manipulation and that sort of thing.
0:50:05 SC: And they're not very good at... Well, I shouldn't say they're not very good at, but the design of the octopus brain, as it were, is utterly different, right, than a...
0:50:15 MM: It's totally different. Yeah.
0:50:16 SC: It's spread out all throughout their body.
0:50:18 MM: But that's not necessarily... So birds also have a very, very different design for their brain, but they also seem to have something like a hippocampus-like structure, and there's a part of the octopus brain that people have identified as being possibly hippocampal in nature. One thing that seems to be the case is that nature is very good at evolving the correct solution regardless of the starting point. So if you need a hippocampal-like structure, then if you started off as invertebrate-stock you'll come up with it, if that's what you need.
0:50:56 SC: Okay. So let's get back to... We did the octopus, got that out of the way, octopus-lovers out there, I get it.
0:51:02 MM: I'm glad that was a totally satisfying answer.
0:51:04 SC: I know that there... Look, there are people I know who are very smart, ethical people who won't eat octopus because it has a certain level of intelligence. These are very good questions, I'm not dogmatic in any direction about them, I don't have an idea of what the right thing to think is, I'm a very curious about this.
0:51:16 MM: I'm hoping to find out.
0:51:17 SC: Good, good. We'll have you back when you find out. That'll be important. But okay, so back to the vertebrates, where we live.
0:51:23 MM: Yes.
0:51:24 SC: So you have this hypothesis, and the hypothesis about the relationship between imagination and climbing up on the land, so there's some evidence in the fossil record that the evolution of eyes did what you might expect it to do under this scenario.
0:51:40 MM: Exactly.
0:51:41 SC: Is there evidence one way or the other for the evolution of brains doing what you'd expect them to do?
0:51:46 MM: Well, that's tricky because they don't fossilize. And so what we can do is you can do things like look at commonality between the structure that is hippocampal-like in birds and in mammals, and their last common ancestor lived about 50 million years after the water to land transition.
0:52:05 SC: Okay, so it's not great.
0:52:08 MM: So there's a big argument, was it parallel evolution? Was it convergent evolution? But, yeah, so it's really hard to know how far back that goes from the fossil record.
[background conversation]
0:52:41 SC: Right, because what we can't get actual fossil records, the best we can do is look at the different things that are alive today and try to roughly map them.
0:52:48 MM: Exactly. Yeah.
0:52:49 SC: It's very hard. It's hard 'cause I'm very... Just the origin of life and you can think about the simplest organisms that live now, and you might think, well, they're closer to what life like when it started, but they're still billions of years away.
0:53:01 MM: Well, not only that, so one of the places we can look is at the pallium, which is the structure that is thought to be ancestral to hippocampus dorsal pallium in particular and part of that, and we can look at it in teleost fish, and I have collegues of mine who've done this and have found evidence for things like spacial maps there, and so on and so forth. Tricky thing there is that our last common ancestor with those animals was hundreds of millions of years ago, and so they've been busy evolving potentially novel structures, more complicated and unique in their own ways. So you have to be careful about comparisons of extant creatures to get after reverse engineer or back calculate what happened long time ago. So not to say it won't be eventually possible, there's increasingly, there's this realm of sort of recovering soft body structure by high... Essentially scanning imaging fossils that were really well preserved, and people are starting to be able to see neural structures that way. And so maybe eventually we'll be able to tell more.
0:54:10 MM: Hippocampus is pretty deep in there, so at least for that structure I have to be a bit skeptical but...
0:54:12 SC: Yeah. But nevertheless, we're gonna fearlessly draw conclusions from this line of use made for human beings.
0:54:17 MM: Totally in favor, yes, yes, yes.
0:54:19 SC: And so, there is the thought that if this story is right, whether or not it's right, there's a fact that our much beloved and wanted ability to plan and think and be conscious is relying upon what's in our brains. And if it did come from... Evolution does not have teleology. It doesn't aim toward the future; it just responds to the moment. And does this story of how we got here have lessons for where we are now?
0:54:46 MM: Yes, and I think the answer to that is yes. And so, that's where the research is going right now, is clearly, we have some capacity to plan for the future, but it seems to be both spatially limited, as in we don't think about spaces very far away from us. I'm not thinking right now about a trail in Africa I might go on in a few hours. And we don't think about things very far way in time, either. We might stretch ourselves to think about our retirement, but not very often. [chuckle]
0:55:21 SC: We're not even very good at that, yeah.
0:55:22 MM: We're pretty bad at that. And then, what are we gonna do about problems that are multi-generational, such as climate change? And so, I'm very interested in what the biological mechanisms are for setting the length of our planning horizon. And there's some very interesting work going on right now on that, and I'm hoping to contribute to that. I recently got an award from the NSF to look at this in mice with Dan Dom back at Northwestern, where we're gonna, essentially, study planning in rodents, and giving them a whole bunch of challenging environments, including predator-prey context, and start to look at what sets the time base of the horizon. So is it, "Can we extend it to days or to weeks or to months? Or is it something that anything beyond a few minutes is purely through cultural technology, such as... "
0:56:24 SC: Well, okay, so the idea is that the fish, naturally, from evolution, plans seconds in advance. And once you climb onto land, now, at least, minutes, makes sense.
0:56:35 MM: Minutes, at least, minutes. More...
0:56:36 SC: The time that it takes between when you see the lion...
0:56:38 MM: Possibly tens to minutes.
0:56:40 SC: Right, okay. But clearly, human beings can do a little bit better. And maybe dogs and cats can't. Now that I'm thinking about it, I'm not even clear that my cats can do seconds in advance. But how can we do... So maybe the first question is, why can we even do that?
0:56:53 MM: Well, right. I think that we can do better than that is interesting. And the way in which we can do it, as in it's... I'll call it "effective valence", how much we care about the things we think about far in the future, or far in time, I think is very telling. And from that perspective, this... Peter Singer has this great analogy of... I'm sure you've heard it: Singer's Pond. This thought experiment of you're driving up... You've just bought a $5,000 suit, you're driving by a pond, you see a struggling child in the pond. What do you do? Well, of course, you get out of your car, and you run into the water, destroy your suit. Well, his question is, "Why don't you send the $5,000 to save a child whose life would clearly be saved by that $5,000 who lives in Africa?" And my answer to that would be, "Well, you have a very carefully titrated system between your care system, your planning horizon, and it all sort of meshes, but it meshes in a very local way."
0:58:07 SC: And that makes sense in terms of evolution, why do evolve...
0:58:10 MM: Precisely. And so, getting beyond that, I think, is a hard bargain, at least, in terms of getting it natively with circuitry, so we can think about cultural technology to do it. But it'd be nice to identify, at least, what's constraining us in terms of biological mechanism. So if we wanted to, in the future or in the sci-fi future to be able to go in with electrodes or some sort of neuro-prosthetic, which would extend our planning horizon, that we could do so. Right now, we don't know enough about it to know how we do it.
0:58:41 SC: Are there even other animal species that do manifestly plan weeks in advance?
0:58:46 MM: Depends on... So there are animals that cache, etcetera. So it depends on... So some of this can be sort of assimilated to instinctual behaviors that are genetically hard programmed. And so, it's clear that the animal has no explicit representation of this occurring, and they're just interfering with the routines, it's very easy to throw them off, and they just repeat. But so, I think squirrels and birds and other mammals have seemed to represent planning, at least, on the order of days, if not weeks, for caching purposes. Beyond that, I think it gets really fuzzy because of this issue of, "Could it be just genetically hardwired?"
0:59:34 SC: Right. But doesn't... So for humans, doesn't it make sense to think... I know this is just too simplistic, but I have this idea that there was a phase transition. When we became... The ability to express ourselves abstractly and in language and communally that gave us a way... And maybe this is what you mean by the cultural artifacts, but even just individually, I think that this gives us an ability to conceptualize the future that isn't just hardwired.
1:00:01 MM: Absolutely, it does, and... But my suspicion is that the gap between that kind of rationality, and or the local care system where our native planning horizons mesh with care, where motivation also has a seat, is kind of the thing that we need to really crack in order to motivate ourselves to regulate behavior, to affect time scales far outside of our natural range. So in other words, so a looming existential threat that might be multi-generational in scope. How do you motivate yourself now to do X, Y, Z? And obviously, we could have something like the climate change panels that are telling us, we need to do this now, and governments be if they're responsible will do that in imposing like carbon taxes, etcetera. But on a personal motivational level I guess that's where that gap can start to really affect us between the natural care system and the cognitive system that a bunch of people discussing this can say, "Look, we need to now reduce our carbon footprint by this amount. Or else."
1:01:29 SC: Right. I do think that there is this maybe a little parenthetical and off-track, but that's okay, that's why we're here. There's a question of how much you should care about things in the future or things that are far away.
1:01:41 MM: Exactly. Yes.
1:01:41 SC: I had Tyler Cowen on the podcast, and I called him, he wasn't happy that I called him this, but I called him the temporal version of Peter Singer.
1:01:48 MM: Yes, I listened to that.
1:01:50 SC: Because he was advocating that we have essentially a zero discount rate for caring about the future rate. He had certain policy recommendations that followed from that that you can disagree with. But it's interesting philosophy question, should we care about people far away and people in the future just as much as our near neighbors and our current selves?
1:02:08 MM: So there's the normative question of how we discount the future, and then there is the empirical, and empirically we are hyperbolic discounters. So we don't, essentially.
1:02:19 SC: So, explain what that means.
1:02:21 MM: Well, so the value of a future harm or benefit rapidly declines with distance from the now, and the function that describes that decline is hyperbolic.
1:02:34 SC: One over x.
1:02:38 MM: So Cowen and others, there's others who have discussed this in the literature, think that this discount rate is inimical to our future [laughter] survival, and I think that's correct.
1:02:53 SC: 'Cause if you don't care that much about people far away then...
1:02:55 MM: Then you act the way we are know.
1:02:58 SC: Live in the moment.
1:02:58 MM: Which is living for the moment, and our future generations, and possibly the next generation be damned. And that's deeply problematic. And one could ask, is humanity worth surviving? That's another level up, but if you care about humanity...
1:03:15 SC: We can grant that for the purposes of this conversation.
1:03:18 MM: Okay. [chuckle]
1:03:18 SC: Humanity or it's descendants would be a good thing to have around.
1:03:21 MM: The pro-octopus person would say, humanity be damned. Let's...
1:03:24 SC: I've also met people who think that it's better not to have humanity around, in all seriousness.
1:03:30 MM: Yes.
[chuckle]
1:03:30 SC: But I'm not gonna invite them onto the podcast. Some bridges I'm not gonna cross. Yes, but I just think that we... I'm just putting this out there. Maybe I've said it before in the podcast, I do think that I'm sympathetic to the idea that we should care more than we empirically do about people far away and people in the future, but I'm not sympathetic to the idea that we should care about everyone equally. No matter where they are, when they are. I think that it's sort of mathematically ill defined to do that, and I think that there's a practicality issue.
1:04:02 MM: Yeah, there is.
1:04:03 SC: I don't have any $5000 suits, but I certainly am not giving away all of my worldly goods in the way that we best benefit mankind. And I think that it's kind of impractical to imagine people doing that. There is something to be said for balancing being good to the world and being good to yourself.
1:04:20 MM: I fully agree. And in fact, I'm gonna forget the name of the philosopher, he's Canadian, he's written this paper on the paradoxes that arise from zero discounting. And there are many, it's...
1:04:32 SC: Yeah. It's not really practical.
1:04:33 MM: It's not practical in many of the same ways that Peter Singer's approach, so being the temporal version has inherits it seems some of the same paradoxes that the spatial version that Peter Singer pushes as...
1:04:49 SC: But I get it as limiting concepts reminding us how far away we are from that ideal. It's very useful to talk about these things.
1:04:56 MM: Yeah, I think if we even went 1/10 of the way we would do much, much better than we are.
1:05:02 SC: Okay, so how can we get people to better... This is the practical inspiration of what you're doing, you wanna get people to be better at taking the future seriously.
1:05:03 MM: Exactly. So my initial thought was maybe if I could really quickly figure out what was going on in the brain to support the horizon we have, I could just invent a prosthetic that everybody would put on and suddenly, they'd be taking their recycling out and...
1:05:26 SC: Or a pill.
1:05:27 MM: Or a pill. I subsequently decided that neuroscience proceeds too slowly.
[chuckle]
1:05:34 SC: The world will be 1,000 degrees.
1:05:36 MM: That was my naive, optimistic self 10 years ago, and then now, I'm now the leading climate... People say we have a decade to two decades to really do something or else it's catastrophe. I believe that's correct. So I've been pursuing cultural technology to try to address the problem. And it was inspired by a book called "The Myth of the Rational Voter" by Caplan. And in that book, he talks about the concept of rational irrationality. The irrational rationality is that if the belief of... If false belief, the cost of false belief is low, then you're gonna have irrationality. So it's sort of an economist take on epistemology. Low-cost, high-irrationality. And so, that got me thinking, well, how can we make something that is outside of our perceptual bound, outside of our sensory envelope that we really ought to care about? How can we make it real? How can we give, get what is sometimes referred to as skin-in-the-game for something like that?
1:06:55 MM: And so, myself and a couple collaborators, Moran Cerf at Northwestern, have been pursuing a study, which we just completed, a little over 150 people participated, half of them were climate dentialist and half of them were believers in climate change, where we had them play a climate prediction market that we created. So there are bets over a period of 30 days on things like, "Will California have more wildfires in the coming month than they did a year ago at the same time?" And whole bunch of things like that. And people got $20 of money to play this market. And if they do really well, they could potentially earn a decent amount from the study. And so, we just finished it, actually, weeks ago. We've collected, we've got all the data sewn up, and I can't wait to get back to analyze it. There's a number of things I'm excited about. One is that I do think that there's this issue of, "Well, how much in the next 30 days can you say is actually due to climate change, and so on and so forth?"
1:08:01 SC: Sure, that's one of many questions, right?
1:08:04 MM: However, there are other things that lead me to think that this could be successful regardless of whether or not that's a problem. For example, as a result of this sports betting, there's been a lot of research on what effect has sport betting have on people's behavior. And the data's clear, it more than triples participation. Now, causality there is unclear, right? So if you're a sports buff, then perhaps, you're gonna watch... You're gonna both bet and watch three times more NFL games, which is what the data show. But there seems to be huge engagement effect. Once you bet money on something happening on an outcome, you get engaged. And we'd just like people to get engaged with climate information. So one of the questions we asked in our survey, pre and post-study survey is, "How many climate-related stories did you attend to in your media scape before and after the study?" So we'd like to see whether there's engagement effect, which could be quite different from whether or not you get skin-in-the-game effects.
1:09:08 SC: I recently came across a study, or it was not that, but similar, and I actually found another podcast. Barry Lam has a Hi-Phi podcast, Hi, P-H-I, for philosophy. And so, he quotes the study, I think it's by Bullock. I'll get the reference, and I'll put it on the website. There's this very well-known difference in beliefs about certain factual things between Democrats and Republicans, or Conservatives and Liberals. It's like, "Did the budget deficit go up or down during the Clinton presidency?" It's a fact, everyone knows there's a number. What do you think the number is? And there's a gap. And they said, "Okay," 'cause they give the survey to some people and they see what the gap is. And then, they give the survey to another bunch of people, and they say, "We'll give you a dollar if you get the answer right." [laughter] And suddenly, the gap goes away, basically.
1:09:58 MM: Yes. Yes, yes.
1:09:58 SC: So people are much better... There's an expressive component to giving opinions about things like that, even when they're facts. And when you actually put the skin-in-the-game, charging them a dollar, then they're more likely to get aware.
1:10:09 MM: Well, you're probably aware of this. So Hanson, who came up... Well, one of the people who came up with prediction markets...
1:10:15 SC: Robin Hanson.
1:10:15 MM: Robin Hanson at George Mason has tons of examples of this, where it seems to take a trivial amount of money to actually close that gap. So it's almost like it's unlinked from the monetary value, just the fact that you have any, actually causes a huge change in your epistemology.
1:10:36 SC: Yeah. No, I'm familiar with this from playing poker, because poker is the most boring game in the world if you're not playing for money.
1:10:41 MM: Yeah, right.
1:10:42 SC: 'Cause people don't act rational. They're like, "Yeah, all right, let's do whatever." They'll play for a couple of pennies, and like, "Oh, I wanna... "
1:10:47 MM: Well, that's a fabulous metaphor for people's position on climate change.
1:10:51 SC: That's right, yeah. So okay, so I can see how putting a little bit of skin-in-the-game makes people act a little bit more rationally. How do we implement that to save the planet?
1:11:00 MM: Well, so the idea would be if we could, say, universalize such a market, suppose there's this tiny 1% tax, carbon tax, and it goes to people being able to play on this prediction market, it'd have a huge number of beneficial effects. One effect would be that you get with some of the crowd effects. So there's a whole bunch of things happening to the climate right now that a very, very small subset of which are being attended to by scientists, things like, for example, ice roads in Alaska having to close much earlier than normal. Things like my friend and Pasadena's Garden has been overtaken by this invasive species that likes warmer climates.
1:11:53 SC: The whole wine industry is changing dramatically.
1:11:55 MM: Either that... There's that...
1:11:55 SC: They're making champagne in Britain.
1:11:57 MM: There's that. So all of these, you can create these positions on the market, and now it becomes an instrument by which we can see what climate change effects are occurring, and it becomes a prosthetic, if you will, for understanding what changes are occurring over what time scale, and that the market might be a means by which we attune our behaviors to longer temporal sort of the...
1:12:30 SC: Sort of imagining essentially forcing people to play in this market, or bribing them to do it?
1:12:37 MM: Well, I suppose I haven't thought that far in advance, but if one could universalize the market and have a small amount of a carbon tax going to everybody's pocket, and they could either just leave it there as sort of direct profit, or they could play this market and possibly amplify based on their knowledge, then the people who want to play the market will do so, and those who don't.
1:13:03 SC: I mean, it relies on the idea that in some sense we have the cognitive capacity to plan far in the future and get it right, but there are shortcuts that we often take that prevent us from doing.
1:13:14 MM: Precisely right, that I feel we need to bring it into... We need to bring these very slowly looming phenomena, be the climate change or a future in AI. We need to bring it into the planning range, our native planning range.
1:13:30 SC: Right.
1:13:30 MM: And the mechanism for that would be to change the cost, so to make it costly not to.
1:13:36 SC: Right. We just need to... Change the incentive structure.
1:13:38 MM: That would... Yeah. Change the incentive structure, but in a way that we can get our arms around spatial temporally speaking, and I feel like a prediction market would be one way, it's probably just one of many ways, just the first one that we thought was worth testing, but there are other approaches, and the principal is basically get skin in the game, as in your skin now, during your life for something that might otherwise take very, very long. A very, very long time.
1:14:09 SC: Climate change is the obvious example of our failure to think on long horizons, but my favorite example is actually solar flares, I don't know if you're very familiar with this idea, but...
1:14:17 MM: I've heard you talk about this.
1:14:18 SC: Yeah, well I had a lunch with a lawyer who was on some committee to look at this. He's not a scientist or an engineer, but he got the expert testimony and the claim which I'm still not sure. I haven't really dug into the evidence for myself. But the claim was, "Look solar flares happen all the time, but really big ones are rare. We're not sure how rare 'cause we have not been collecting data for solar flares for that long. It is absolutely possible that on time scales of once per every thousand years, we get a solar flare that will be big enough to essentially wipe out the entire power grid of the earth, and millions and millions of people would die."
1:14:54 MM: Yes.
1:14:55 SC: Because they'd be without power for months or something like that, right?
1:14:57 MM: Yes.
1:14:58 SC: And it's not that costly compared to the downside to harden the grid.
1:15:03 MM: I see.
1:15:03 SC: And save that from this.
1:15:03 MM: I see.
1:15:04 SC: But no one is gonna pay the money to do that, when you say there's a one in a thousand chance per year that this could happen. [chuckle]
1:15:09 MM: Yes, yes.
1:15:11 SC: You just can't do that calculation. We're not good at it.
1:15:14 MM: Yeah, it needs to have happened at least once in the modern period.
1:15:18 SC: Yeah. So we're just... Nothing in evolution ever trains you to plan on time scales 10 times longer than a human life time for obvious reasons.
1:15:26 MM: Right. I wonder if that's... Maybe that's a good "Planet of the Apes" movie.
1:15:31 SC: I think that things like that are because scare people a little bit, is useful if it's a good scare. Like, the China syndrome was kinda silly.
[laughter]
1:15:41 SC: And "Day After Tomorrow" was kinda silly for different reasons, but yes, we can apply art and culture to that.
1:15:49 MM: Yeah.
1:15:49 SC: Which reminds me, we'll end the podcast by... Because people need to know this. You are the science advisor on a very well-known, culturally important, TV show.
[laughter]
1:15:49 MM: Caprica.
1:16:00 SC: Caprica, that's right, which was the follow up to [1:16:04] ____.
1:16:04 MM: Yes. Not sure if it falls into all of those categories, but yes, that was a great experience. Being able to go through every script and be a part of that show.
1:16:14 SC: Do you think that there is value... Or how much value do you think there is in scientists engaging in that kind of engagement with popular culture? It's not outreach in the sense that you're not learning neuroscience from watching Caprica, right?
1:16:26 MM: Right.
1:16:26 SC: But maybe by nudging the script in certain ways, you're inspiring people. What is your thought about that?
1:16:34 MM: Yeah, I think that there's lots of ways in which it enriches these shows, and in particular... The script writers aren't necessarily well-trained in science, typically not... And so, when they are thinking about how do I portray this AI breakthrough or this roboticist, they're gonna take off things from the show, which are probably things that they saw in movies and stereotypes that they have from pop culture actually and don't track with what is a much more interesting and fine-grain sort of thing that's going on in reality. And so, what we're able to do is sort of add some nuance there, and but also there's the great thing about Caprica is that it had lots of interesting philosophical aspects to this whole going into VR and the concept of having... I remember talking with the show writers for a long time about how do we have death in VR? And that was a really fascinating problem to puzzle over and eventually decided on something like a game that you got kicked out of once you died.
1:17:48 SC: Yeah. It was a video game really.
1:17:50 MM: And it was a really important game, because in the context of the show, it's a one game in which this person could see their daughter who died and now lives in this virtual reality, and so getting kicked out of show really meant something very meaningful, right?
1:18:03 SC: And did the influence flow the other way? Did working on Caprica in any way influenced you to think deeply about the looming AI menace, the robot takeover?
[laughter]
1:18:13 MM: Well, I think that Battlestar Galactica and Caprica did an amazing job of envisaging this whole arc, right? And there's these beautiful moments of Battle Star, for example, where the Cylon... One of the Cylon whose name is Boomer says, "You know, I'm not sure humanity has a right to continue existing," and just pushes on a challenge for us.
1:18:42 SC: It is a good little challenge for us, right?
1:18:43 MM: It is a good little challenge, and I feel like the way they envisaged the stars that have moral prospects with AI was a fascinating and deeply thought-provoking one. So, yes.
1:19:00 SC: Well, we do hope that your work, and the work of others that we had on the podcast will help stave off the Cylon from taking over at some point.
[laughter]
1:19:07 SC: Malcom MacIver, thanks so much being on the podcast.
1:19:08 MM: It was great to be on the show, thanks.
[music]
Pingback: A Participatory Universe – On Sensing, Consciousness and Imagination – Traders Outpost
What an awesome conversation! Reminded me of Numenta’s work on the role of movement and mapping for cognition (https://numenta.com/neuroscience-research/research-publications/posters/sfn-2018-sensorimotor-object-recognition-using-grid-cells/). Also on my own work vis-a-vis maps and cognition (and this does attempt to attack the hard problem of consciousness) https://www.wisdomneverdies.com/blog/cognition-as-maps-of-maps
Thank you Sean and Malcolm!
What’s the name of that differential geometry book Sean mentions in this episode?
@Charlie: I’m not sure but I think it might be “Applied Differential Geometry” by William L. Burke. (https://www.goodreads.com/book/show/2134261.Applied_Differential_Geometry)
Wonderful topic. Along with cosmology I thoroughly enjoy evolutionary biology. This was so fascinating I missed my exit on the way to work. It went form discovering how electric fish sense prey to the Devonian. Fascinating! I’d love more topics like this!!
Great listen. These points dovetail somewhat with James C. Wathey’s theory about the neurobiological basis of religious belief…he might be a guest worth having on.
Hey There! It might have been said, but who is the Canadian guy that Malcolm mentions about zero discounting? I would love to read something contemporary about how this has fluctuated (if it has) over time in human history. Thanks! Peter
I’m not certain it was this podcast episode but Sean you referred to Noam Chomsky as ‘obnoxious’. I have listened to many of his lectures Q&As, interviews and debates. I can’t say anything I ever heard was obnoxious. He can be a bit dismissive perhaps, of people who display an ignorance of the subject matter in their criticism of him or his viewpoint, which he should do. Have I missed something?
Here it is: Journal of Moral Philosophy. Climate Ethics: Justifying a Positive Social Time Preference, by Joseph Heath. 2017. Also good by same author: The Structure of Intergenerational Cooperation
Source: Philosophy & Public Affairs, Vol. 41, No. 1 (WINTER 2013), pp. 31-66.