WEBVTT 00:00:03.393 --> 00:00:04.621 Science, 00:00:04.621 --> 00:00:07.958 science has allowed us to know so much 00:00:07.958 --> 00:00:10.984 about the far reaches of the universe, 00:00:10.984 --> 00:00:14.179 which is at the same time tremendously important 00:00:14.179 --> 00:00:16.245 and extremely remote, 00:00:16.245 --> 00:00:18.704 and yet much, much closer, 00:00:18.704 --> 00:00:20.795 much more directly related to us, 00:00:20.795 --> 00:00:23.263 there are many things we don't really understand. 00:00:23.263 --> 00:00:25.392 And one of them is the extraordinary 00:00:25.392 --> 00:00:28.718 social complexity of the animals around us, 00:00:28.718 --> 00:00:30.734 and today I want to tell you a few stories 00:00:30.734 --> 00:00:32.742 of animal complexity. NOTE Paragraph 00:00:32.742 --> 00:00:36.092 But first, what do we call complexity? 00:00:36.092 --> 00:00:37.579 What is complex? 00:00:37.579 --> 00:00:41.006 Well, complex is not complicated. 00:00:41.006 --> 00:00:44.454 Something complicated comprises many small parts, 00:00:44.454 --> 00:00:46.884 all different, and each of them 00:00:46.899 --> 00:00:50.003 has its own precise role in the machinery. 00:00:50.003 --> 00:00:52.814 On the opposite, a complex system 00:00:52.814 --> 00:00:55.455 is made of many, many similar parts, 00:00:55.455 --> 00:00:57.463 and it is their interaction 00:00:57.463 --> 00:01:00.783 that produces a globally coherent behavior. 00:01:00.783 --> 00:01:04.619 Complex systems have many interacting parts 00:01:04.619 --> 00:01:08.045 which behave according to simple, individual rules, 00:01:08.045 --> 00:01:11.394 and this results in emergent properties. 00:01:11.394 --> 00:01:13.282 The behavior of the system as a whole 00:01:13.282 --> 00:01:14.950 cannot be predicted 00:01:14.950 --> 00:01:17.102 from the individual rules only. 00:01:17.102 --> 00:01:18.912 As Aristotle wrote, 00:01:18.912 --> 00:01:21.972 the whole is greater than the sum of its parts. 00:01:21.972 --> 00:01:24.434 But from Aristotle, let's move onto 00:01:24.434 --> 00:01:28.124 a more concrete example of complex systems. NOTE Paragraph 00:01:28.124 --> 00:01:30.080 These are Scottish terriers. 00:01:30.080 --> 00:01:33.831 In the beginning, the system is disorganized. 00:01:33.831 --> 00:01:37.632 Then comes a perturbation: milk. 00:01:37.632 --> 00:01:41.482 Every individual starts pushing in one direction 00:01:41.482 --> 00:01:44.791 and this is what happens. 00:01:44.791 --> 00:01:47.617 The pinwheel is an emergent property 00:01:47.617 --> 00:01:49.520 of the interactions between puppies 00:01:49.520 --> 00:01:53.430 whose only rule is to try to keep access to the milk 00:01:53.430 --> 00:01:57.037 and therefore to push in a random direction. NOTE Paragraph 00:01:57.037 --> 00:02:01.012 So it's all about finding the simple rules 00:02:01.012 --> 00:02:03.770 from which complexity emerges. 00:02:03.770 --> 00:02:06.710 I call this simplifying complexity, 00:02:06.710 --> 00:02:08.845 and it's what we do at the chair of systems design 00:02:08.845 --> 00:02:10.822 at ETH Zurich. 00:02:10.822 --> 00:02:14.527 We collect data on animal populations, 00:02:14.527 --> 00:02:18.338 analyze complex patterns, try to explain them. 00:02:18.338 --> 00:02:20.957 It requires physicists who work with biologists, 00:02:20.957 --> 00:02:23.680 with mathematicians and computer scientists, 00:02:23.680 --> 00:02:26.500 and it is their interaction that produces 00:02:26.500 --> 00:02:28.214 cross-boundary competence 00:02:28.214 --> 00:02:29.792 to solve these problems. 00:02:29.792 --> 00:02:32.064 So again, the whole is greater 00:02:32.064 --> 00:02:33.464 than the sum of the parts. 00:02:33.464 --> 00:02:35.614 In a way, collaboration 00:02:35.614 --> 00:02:39.105 is another example of a complex system. NOTE Paragraph 00:02:39.105 --> 00:02:40.981 And you may be asking yourself 00:02:40.981 --> 00:02:43.798 which side I'm on, biology or physics? 00:02:43.798 --> 00:02:45.909 In fact, it's a little different, 00:02:45.909 --> 00:02:47.498 and to explain, I need to tell you 00:02:47.498 --> 00:02:49.840 a short story about myself. 00:02:49.840 --> 00:02:51.567 When I was a child, 00:02:51.567 --> 00:02:55.676 I loved to build stuff, to create complicated machines. 00:02:55.676 --> 00:02:58.413 So I set out to study electrical engineering 00:02:58.413 --> 00:02:59.965 and robotics, 00:02:59.965 --> 00:03:02.058 and my end-of-studies project 00:03:02.058 --> 00:03:04.984 was about building a robot called ER-1 -- 00:03:04.984 --> 00:03:06.914 it looked like this— 00:03:06.914 --> 00:03:09.285 that would collect information from its environment 00:03:09.285 --> 00:03:12.783 and proceed to follow a white line on the ground. 00:03:12.783 --> 00:03:15.162 It was very, very complicated, 00:03:15.162 --> 00:03:18.146 but it worked beautifully in our test room, 00:03:18.146 --> 00:03:21.599 and on demo day, professors had assembled to grade the project. 00:03:21.607 --> 00:03:24.509 So we took ER-1 to the evaluation room. 00:03:24.509 --> 00:03:26.819 It turned out, the light in that room 00:03:26.819 --> 00:03:28.638 was slightly different. 00:03:28.638 --> 00:03:30.969 The robot's vision system got confused. 00:03:30.969 --> 00:03:32.730 At the first bend in the line, 00:03:32.730 --> 00:03:36.469 it left its course, and crashed into a wall. 00:03:36.469 --> 00:03:38.556 We had spent weeks building it, 00:03:38.556 --> 00:03:40.229 and all it took to destroy it 00:03:40.229 --> 00:03:42.885 was a subtle change in the color of the light 00:03:42.885 --> 00:03:44.481 in the room. 00:03:44.481 --> 00:03:45.996 That's when I realized that 00:03:45.996 --> 00:03:48.323 the more complicated you make a machine, 00:03:48.323 --> 00:03:50.362 the more likely that it will fail 00:03:50.362 --> 00:03:52.925 due to something absolutely unexpected. 00:03:52.925 --> 00:03:54.755 And I decided that, in fact, 00:03:54.755 --> 00:03:57.768 I didn't really want to create complicated stuff. 00:03:57.768 --> 00:04:00.710 I wanted to understand complexity, 00:04:00.710 --> 00:04:02.698 the complexity of the world around us 00:04:02.698 --> 00:04:05.103 and especially in the animal kingdom. NOTE Paragraph 00:04:05.103 --> 00:04:08.423 Which brings us to bats. 00:04:08.423 --> 00:04:11.474 Bechstein's bats are a common species of European bats. 00:04:11.474 --> 00:04:12.887 They are very social animals. 00:04:12.887 --> 00:04:16.178 Mostly they roost, or sleep, together. 00:04:16.178 --> 00:04:17.857 And they live in maternity colonies, 00:04:17.857 --> 00:04:19.397 which means that every spring, 00:04:19.397 --> 00:04:22.655 the females meet after the winter hibernation, 00:04:22.655 --> 00:04:24.744 and they stay together for about six months 00:04:24.744 --> 00:04:27.230 to rear their young, 00:04:27.230 --> 00:04:30.035 and they all carry a very small chip, 00:04:30.035 --> 00:04:31.906 which means that every time one of them 00:04:31.906 --> 00:04:34.963 enters one of these specially equipped bat boxes, 00:04:34.963 --> 00:04:36.606 we know where she is, 00:04:36.606 --> 00:04:37.775 and more importantly, 00:04:37.775 --> 00:04:40.338 we know with whom she is. 00:04:40.338 --> 00:04:44.032 So I study roosting associations in bats, 00:04:44.032 --> 00:04:46.477 and this is what it looks like. 00:04:46.477 --> 00:04:48.919 During the day, the bats roost 00:04:48.919 --> 00:04:51.223 in a number of sub-groups in different boxes. 00:04:51.223 --> 00:04:53.152 It could be that on one day, 00:04:53.152 --> 00:04:55.372 the colony is split between two boxes, 00:04:55.372 --> 00:04:56.672 but on another day, 00:04:56.672 --> 00:04:58.913 it could be together in a single box, 00:04:58.913 --> 00:05:01.229 or split between three or more boxes, 00:05:01.229 --> 00:05:04.156 and that all seems rather erratic, really. 00:05:04.156 --> 00:05:07.359 It's called fission-fusion dynamics, 00:05:07.359 --> 00:05:09.072 the property for an animal group 00:05:09.072 --> 00:05:11.250 of regularly splitting and merging 00:05:11.250 --> 00:05:12.911 into different subgroups. NOTE Paragraph 00:05:12.911 --> 00:05:15.473 So what we do is take all these data 00:05:15.473 --> 00:05:17.135 from all these different days 00:05:17.135 --> 00:05:18.639 and pool them together 00:05:18.639 --> 00:05:21.256 to extract a long-term association pattern 00:05:21.256 --> 00:05:23.761 by applying techniques with network analysis 00:05:23.761 --> 00:05:25.382 to get a complete picture 00:05:25.382 --> 00:05:27.919 of the social structure of the colony. 00:05:27.919 --> 00:05:32.184 Okay? So that's what this picture looks like. 00:05:32.184 --> 00:05:34.578 In this network, all the circles 00:05:34.578 --> 00:05:37.355 are nodes, individual bats, 00:05:37.355 --> 00:05:38.938 and the lines between them 00:05:38.938 --> 00:05:42.602 are social bonds, associations between individuals. 00:05:42.602 --> 00:05:45.280 It turns out this is a very interesting picture. 00:05:45.280 --> 00:05:47.262 This bat colony is organized 00:05:47.262 --> 00:05:49.130 in two different communities 00:05:49.130 --> 00:05:50.969 which cannot be predicted 00:05:50.969 --> 00:05:53.218 from the daily fission-fusion dynamics. 00:05:53.218 --> 00:05:56.768 We call them cryptic social units. 00:05:56.768 --> 00:05:58.384 Even more interesting, in fact: 00:05:58.384 --> 00:06:00.748 Every year, around October, 00:06:00.748 --> 00:06:02.309 the colony splits up, 00:06:02.309 --> 00:06:05.007 and all bats hibernate separately, 00:06:05.007 --> 00:06:06.468 but year after year, 00:06:06.468 --> 00:06:09.541 when the bats come together again in the spring, 00:06:09.541 --> 00:06:12.131 the communities stay the same. NOTE Paragraph 00:06:12.131 --> 00:06:14.851 So these bats remember their friends 00:06:14.851 --> 00:06:16.681 for a really long time. 00:06:16.681 --> 00:06:19.155 With a brain the size of a peanut, 00:06:19.155 --> 00:06:21.280 they maintain individualized, 00:06:21.280 --> 00:06:23.422 long-term social bonds, 00:06:23.422 --> 00:06:25.146 We didn't know that was possible. 00:06:25.146 --> 00:06:26.905 We knew that primates 00:06:26.905 --> 00:06:29.473 and elephants and dolphins could do that, 00:06:29.473 --> 00:06:32.101 but compared to bats, they have huge brains. 00:06:32.101 --> 00:06:34.500 So how could it be 00:06:34.500 --> 00:06:36.451 that the bats maintain this complex, 00:06:36.451 --> 00:06:38.139 stable social structure 00:06:38.139 --> 00:06:41.671 with such limited cognitive abilities? NOTE Paragraph 00:06:41.671 --> 00:06:44.560 And this is where complexity brings an answer. 00:06:44.560 --> 00:06:46.701 To understand this system, 00:06:46.701 --> 00:06:49.498 we built a computer model of roosting, 00:06:49.498 --> 00:06:51.516 based on simple, individual rules, 00:06:51.516 --> 00:06:53.951 and simulated thousands and thousands of days 00:06:53.951 --> 00:06:55.970 in the virtual bat colony. 00:06:55.970 --> 00:06:58.094 It's a mathematical model, 00:06:58.094 --> 00:07:00.048 but it's not complicated. 00:07:00.048 --> 00:07:03.146 What the model told us is that, in a nutshell, 00:07:03.146 --> 00:07:06.332 each bat knows a few other colony members 00:07:06.332 --> 00:07:08.820 as her friends, and is just slightly more likely 00:07:08.820 --> 00:07:11.330 to roost in a box with them. 00:07:11.330 --> 00:07:13.774 Simple, individual rules. 00:07:13.774 --> 00:07:15.486 This is all it takes to explain 00:07:15.486 --> 00:07:17.875 the social complexity of these bats. NOTE Paragraph 00:07:17.875 --> 00:07:19.593 But it gets better. 00:07:19.593 --> 00:07:22.441 Between 2010 and 2011, 00:07:22.441 --> 00:07:25.894 the colony lost more than two thirds of its members, 00:07:25.894 --> 00:07:28.880 probably due to the very cold winter. 00:07:28.880 --> 00:07:32.024 The next spring, it didn't form two communities 00:07:32.024 --> 00:07:33.295 like every year, 00:07:33.295 --> 00:07:35.498 which may have led the whole colony to die 00:07:35.498 --> 00:07:37.593 because it had become too small. 00:07:37.593 --> 00:07:42.966 Instead, it formed a single, cohesive social unit, 00:07:42.966 --> 00:07:45.698 which allowed the colony to survive that season 00:07:45.698 --> 00:07:48.802 and thrive again in the next two years. 00:07:48.802 --> 00:07:50.580 What we know is that the bats 00:07:50.580 --> 00:07:53.487 are not aware that their colony is doing this. 00:07:53.487 --> 00:07:57.033 All they do is follow simple association rules, 00:07:57.033 --> 00:07:58.382 and from this simplicity 00:07:58.382 --> 00:08:00.823 emerges social complexity 00:08:00.823 --> 00:08:03.663 which allows the colony to be resilient 00:08:03.663 --> 00:08:06.644 against dramatic changes in the population structure. 00:08:06.644 --> 00:08:09.338 And I find this incredible. NOTE Paragraph 00:08:09.338 --> 00:08:11.422 Now I want to tell you another story, 00:08:11.422 --> 00:08:12.977 but for this we have to travel from Europe 00:08:12.977 --> 00:08:16.025 to the Kalahari Desert in South Africa. 00:08:16.025 --> 00:08:18.052 This is where meerkats live. 00:08:18.052 --> 00:08:19.552 I'm sure you know meerkats. 00:08:19.552 --> 00:08:21.658 They're fascinating creatures. 00:08:21.658 --> 00:08:24.647 They live in groups with a very strict social hierarchy. 00:08:24.647 --> 00:08:26.106 There is one dominant pair, 00:08:26.106 --> 00:08:27.488 and many subordinates, 00:08:27.488 --> 00:08:29.202 some acting as sentinels, 00:08:29.202 --> 00:08:30.539 some acting as babysitters, 00:08:30.539 --> 00:08:32.436 some teaching pups, and so on. 00:08:32.436 --> 00:08:35.757 What we do is put very small GPS collars 00:08:35.757 --> 00:08:37.282 on these animals 00:08:37.282 --> 00:08:39.157 to study how they move together, 00:08:39.157 --> 00:08:42.874 and what this has to do with their social structure. 00:08:42.874 --> 00:08:44.364 And there's a very interesting example 00:08:44.364 --> 00:08:47.080 of collective movement in meerkats. 00:08:47.080 --> 00:08:49.447 In the middle of the reserve which they live in 00:08:49.447 --> 00:08:50.656 lies a road. 00:08:50.656 --> 00:08:53.889 On this road there are cars, so it's dangerous. 00:08:53.889 --> 00:08:56.173 But the meerkats have to cross it 00:08:56.173 --> 00:08:58.747 to get from one feeding place to another. 00:08:58.747 --> 00:09:03.498 So we asked, how exactly do they do this? 00:09:03.498 --> 00:09:05.334 We found that the dominant female 00:09:05.334 --> 00:09:07.955 is mostly the one who leads the group to the road, 00:09:07.955 --> 00:09:11.227 but when it comes to crossing it, crossing the road, 00:09:11.227 --> 00:09:13.578 she gives way to the subordinates, 00:09:13.578 --> 00:09:15.355 a manner of saying, 00:09:15.355 --> 00:09:18.037 "Go ahead, tell me if it's safe." 00:09:18.037 --> 00:09:19.701 What I didn't know, in fact, 00:09:19.701 --> 00:09:22.843 was what rules in their behavior the meerkats follow 00:09:22.843 --> 00:09:25.768 for this change at the edge of the group to happen 00:09:25.768 --> 00:09:29.618 and if simple rules were sufficient to explain it. NOTE Paragraph 00:09:29.618 --> 00:09:33.609 So I built a model, a model of simulated meerkats 00:09:33.609 --> 00:09:35.522 crossing a simulated road. 00:09:35.522 --> 00:09:37.394 It's a simplistic model. 00:09:37.394 --> 00:09:40.234 Moving meerkats are like random particles 00:09:40.234 --> 00:09:42.456 whose unique rule is one of alignment. 00:09:42.456 --> 00:09:44.862 They simply move together. 00:09:44.862 --> 00:09:48.046 When these particles get to the road, 00:09:48.046 --> 00:09:49.988 they sense some kind of obstacle, 00:09:49.988 --> 00:09:52.072 and they bounce against it. 00:09:52.072 --> 00:09:53.228 The only difference 00:09:53.228 --> 00:09:55.270 between the dominant female, here in red, 00:09:55.270 --> 00:09:56.755 and the other individuals, 00:09:56.755 --> 00:09:59.309 is that for her, the height of the obstacle, 00:09:59.309 --> 00:10:01.814 which is in fact the risk perceived from the road, 00:10:01.814 --> 00:10:03.763 is just slightly higher, 00:10:03.763 --> 00:10:05.424 and this tiny difference 00:10:05.424 --> 00:10:07.262 in the individual's rule of movement 00:10:07.262 --> 00:10:09.708 is sufficient to explain what we observe, 00:10:09.708 --> 00:10:12.268 that the dominant female 00:10:12.268 --> 00:10:13.702 leads her group to the road 00:10:13.702 --> 00:10:15.372 and then gives way to the others 00:10:15.372 --> 00:10:18.235 for them to cross first. 00:10:18.235 --> 00:10:21.886 George Box, who was an English statistician, 00:10:21.886 --> 00:10:24.848 once wrote, "All models are false, 00:10:24.848 --> 00:10:26.907 but some models are useful." 00:10:26.907 --> 00:10:30.104 And in fact, this model is obviously false, 00:10:30.104 --> 00:10:34.072 because in reality, meerkats are anything but random particles. 00:10:34.072 --> 00:10:35.709 But it's also useful, 00:10:35.709 --> 00:10:38.458 because it tells us that extreme simplicity 00:10:38.458 --> 00:10:41.816 in movement rules at the individual level 00:10:41.816 --> 00:10:44.167 can result in a great deal of complexity 00:10:44.167 --> 00:10:46.105 at the level of the group. 00:10:46.105 --> 00:10:50.161 So again, that's simplifying complexity. NOTE Paragraph 00:10:50.161 --> 00:10:51.609 I would like to conclude 00:10:51.609 --> 00:10:54.426 on what this means for the whole species. 00:10:54.426 --> 00:10:56.090 When the dominant female 00:10:56.090 --> 00:10:57.656 gives way to a subordinate, 00:10:57.656 --> 00:10:59.773 it's not out of courtesy. 00:10:59.773 --> 00:11:01.280 In fact, the dominant female 00:11:01.280 --> 00:11:03.799 is extremely important for the cohesion of the group. 00:11:03.799 --> 00:11:07.311 If she dies on the road, the whole group is at risk. 00:11:07.311 --> 00:11:09.547 So this behavior of risk avoidance 00:11:09.547 --> 00:11:12.348 is a very old evolutionary response. 00:11:12.348 --> 00:11:16.217 These meerkats are replicating an evolved tactic 00:11:16.217 --> 00:11:18.450 that is thousands of generations old, 00:11:18.450 --> 00:11:20.864 and they're adapting it to a modern risk, 00:11:20.864 --> 00:11:24.189 in this case a road built by humans. 00:11:24.189 --> 00:11:26.584 They adapt very simple rules, 00:11:26.584 --> 00:11:28.873 and the resulting complex behavior 00:11:28.873 --> 00:11:31.829 allows them to resist human encroachment 00:11:31.829 --> 00:11:34.277 into their natural habitat. NOTE Paragraph 00:11:34.277 --> 00:11:36.079 In the end, 00:11:36.079 --> 00:11:38.779 it may be bats which change their social structure 00:11:38.779 --> 00:11:41.163 in response to a population crash, 00:11:41.163 --> 00:11:42.562 or it may be meerkats 00:11:42.562 --> 00:11:45.764 who show a novel adaptation to a human road, 00:11:45.764 --> 00:11:48.449 or it may be another species. 00:11:48.449 --> 00:11:51.242 My message here -- and it's not a complicated one, 00:11:51.242 --> 00:11:54.006 but a simple one of wonder and hope -- 00:11:54.006 --> 00:11:57.099 my message here is that animals 00:11:57.099 --> 00:11:59.523 show extraordinary social complexity, 00:11:59.523 --> 00:12:01.964 and this allows them to adapt 00:12:01.964 --> 00:12:05.445 and respond to changes in their environment. 00:12:05.445 --> 00:12:08.213 In three words, in the animal kingdom, 00:12:08.213 --> 00:12:10.987 simplicity leads to complexity 00:12:10.987 --> 00:12:12.470 which leads to resilience. NOTE Paragraph 00:12:12.470 --> 00:12:14.754 Thank you. NOTE Paragraph 00:12:14.754 --> 00:12:21.434 (Applause) 00:12:30.694 --> 00:12:32.647 Dania Gerhardt: Thank you very much, Nicolas, 00:12:32.647 --> 00:12:35.926 for this great start. Little bit nervous? 00:12:35.926 --> 00:12:37.570 Nicolas Perony: I'm okay, thanks. 00:12:37.570 --> 00:12:40.030 DG: Okay, great. I'm sure a lot of people in the audience 00:12:40.030 --> 00:12:41.894 somehow tried to make associations 00:12:41.894 --> 00:12:43.718 between the animals you were talking about -- 00:12:43.718 --> 00:12:45.774 the bats, meerkats -- and humans. 00:12:45.774 --> 00:12:46.982 You brought some examples: 00:12:46.982 --> 00:12:48.717 The females are the social ones, 00:12:48.717 --> 00:12:50.430 the females are the dominant ones, 00:12:50.430 --> 00:12:52.103 I'm not sure who thinks how. 00:12:52.103 --> 00:12:54.998 But is it okay to do these associations? 00:12:54.998 --> 00:12:57.798 Are there stereotypes you can confirm in this regard 00:12:57.798 --> 00:13:01.071 that can be valid across all species? 00:13:01.071 --> 00:13:02.674 NP: Well, I would say there are also 00:13:02.674 --> 00:13:04.626 counter-examples to these stereotypes. 00:13:04.626 --> 00:13:07.766 For examples, in sea horses or in koalas, in fact, 00:13:07.766 --> 00:13:11.464 it is the males who take care of the young always. 00:13:11.464 --> 00:13:16.505 And the lesson is that it's often difficult, 00:13:16.505 --> 00:13:18.257 and sometimes even a bit dangerous, 00:13:18.257 --> 00:13:20.929 to draw parallels between humans and animals. 00:13:20.929 --> 00:13:23.035 So that's it. 00:13:23.035 --> 00:13:25.881 DG: Okay. Thank you very much for this great start. 00:13:25.881 --> 00:13:27.961 Thank you, Nicolas Perony.