1 00:00:09,190 --> 00:00:14,920 Hello I'm Eve Mitleton-Kelly, I'm the founder and director of the complexity 2 00:00:14,920 --> 00:00:18,520 research program at the London School of Economics. 3 00:00:18,520 --> 00:00:23,680 What we will be looking at over the next two videos are the generic characteristics of 4 00:00:23,680 --> 00:00:24,680 complex systems. 5 00:00:24,680 --> 00:00:30,250 In the first video, we will discuss the characteristics that we 6 00:00:30,250 --> 00:00:36,230 have come to understand through systems theory that are essential for us to build on to then 7 00:00:36,230 --> 00:00:39,920 develop a fuller understanding of complexity theory. 8 00:00:39,920 --> 00:00:44,780 So let me show you what we are going to look at. 9 00:00:44,780 --> 00:00:50,249 So the first video will look at the connectivity, inter-dependence, 10 00:00:50,249 --> 00:00:51,710 feedback and emergence. 11 00:00:51,710 --> 00:00:56,120 And in the second video we will look at the other characteristics. 12 00:00:56,120 --> 00:01:01,749 Let's start with connectivity. 13 00:01:01,749 --> 00:01:05,780 But first of all to understand complex behavior. 14 00:01:05,780 --> 00:01:15,700 Complex behavior arises from interaction, and complexity theory focuses on relationships. 15 00:01:16,000 --> 00:01:21,700 The theory does not focus on the entities themselves but on the relationships between them. 16 00:01:21,780 --> 00:01:29,700 The other key characteristics of complex theory is that complex systems can create new order. 17 00:01:29,800 --> 00:01:33,380 This is the jargon, and I hope that you will understand what that 18 00:01:33,380 --> 00:01:38,040 means over the course of the next two video's. 19 00:01:38,040 --> 00:01:41,910 Connectivity and interdependence, what I would like to emphasize is that there 20 00:01:41,910 --> 00:01:44,420 are degrees of connectivity. 21 00:01:44,420 --> 00:01:49,660 In other words, connectivity between entities does not remain 22 00:01:49,660 --> 00:01:55,000 the same over time, it can vary in quality and in intensity. 23 00:01:55,000 --> 00:02:01,270 And when I talk about entities, I mean anything from to individuals interacting, 24 00:02:01,270 --> 00:02:03,880 to groups, to whole organizations, 25 00:02:03,880 --> 00:02:05,360 to whole economies. 26 00:02:05,360 --> 00:02:11,090 So entities refers to all those, simply because all the principles of complexity 27 00:02:11,090 --> 00:02:16,349 we will be discussing are scale invariant, they apply at all scales. 28 00:02:16,349 --> 00:02:17,580 So let's go back. 29 00:02:17,580 --> 00:02:22,150 Connectivity, there are two things to actually look at. 30 00:02:22,150 --> 00:02:28,160 One is strength of coupling, which is how strongly related the interacting 31 00:02:28,160 --> 00:02:29,520 entities may be. 32 00:02:29,520 --> 00:02:33,010 And the other one is epistatic interactions. 33 00:02:33,010 --> 00:02:40,300 Now this term comes from biology, but let me explain what it means in a human context. 34 00:02:40,480 --> 00:02:47,000 Some of you would have gone through a process of actually hiring people, 35 00:02:47,000 --> 00:02:50,459 you look at their cv, you interview them, 36 00:02:50,459 --> 00:02:57,459 and you'll point someone that you feel is going to make a very good contribution to 37 00:02:57,780 --> 00:02:59,209 the rest of the team. 38 00:02:59,209 --> 00:03:02,640 Six months down the line, you find that that person has not actually 39 00:03:02,640 --> 00:03:04,610 made that contribution. 40 00:03:04,610 --> 00:03:13,600 Now one of the possibilities is that the other members of the team have not actually alowed 41 00:03:13,600 --> 00:03:17,650 that new member to make that contribution. 42 00:03:17,650 --> 00:03:26,000 So what that means in the definition, is that the fitness contribution made by one 43 00:03:26,000 --> 00:03:30,310 individual will depend upon related individuals. 44 00:03:30,310 --> 00:03:36,310 The opposite of course will also be the case, and it could be that the person that may appear 45 00:03:36,310 --> 00:03:41,340 at the beginning to be shy or withdrawn, in the right environment, 46 00:03:41,340 --> 00:03:45,610 may then flower and actually make a very, very good contribution. 47 00:03:45,610 --> 00:03:49,489 So that is epistatic interactions. 48 00:03:49,489 --> 00:03:53,140 As I said, is the fitness contribution made by one individual 49 00:03:53,140 --> 00:03:59,380 that will depend upon related individuals. 50 00:03:59,380 --> 00:04:08,300 We of course are constantly being encouraged to increase our connectivity to increase our networks. 51 00:04:08,489 --> 00:04:13,629 However, we need to be aware of what might happen if 52 00:04:13,629 --> 00:04:20,629 we push that too far, because intense interconnectivity also creates 53 00:04:20,799 --> 00:04:23,819 quite intricate dependences. 54 00:04:23,819 --> 00:04:30,560 So the system becomes to be very dependent, or the different parts become very dependent 55 00:04:30,560 --> 00:04:34,710 on each other, and these dependences cannot be pulled apart. 56 00:04:34,710 --> 00:04:44,200 The outcomes are often non-deterministic and the point here is that complexity does not 57 00:04:44,240 --> 00:04:47,900 argue for ever increasing connectivity. 58 00:04:47,919 --> 00:04:56,900 Why? Because if we push the system too far into being connected it then becomes too inter-dependent 59 00:04:57,000 --> 00:04:59,550 and it becomes fragile. 60 00:04:59,550 --> 00:05:02,560 And this may lead to complexity catastrophe. 61 00:05:02,560 --> 00:05:07,680 Let me give you an example. 62 00:05:07,680 --> 00:05:13,260 An American group has done a study, the study was done in 2010. 63 00:05:13,260 --> 00:05:19,660 They took 500 corporations with the highest stock trading volume and they were analyzed. 64 00:05:19,660 --> 00:05:25,030 What they did is they looked at five economic sectors, 65 00:05:25,030 --> 00:05:27,620 technology, which you see in blue. 66 00:05:27,620 --> 00:05:29,510 Oil, in dark grey, 67 00:05:29,510 --> 00:05:33,650 other basic materials light grey, finance linked to real estate, 68 00:05:33,650 --> 00:05:35,750 dark green, and other finance, 69 00:05:35,750 --> 00:05:36,600 light green. 70 00:05:36,600 --> 00:05:43,000 Now please observe 2003 and 2008. 71 00:05:43,000 --> 00:05:52,300 In 2003 each sector, each economic sector was interconnected within 72 00:05:52,300 --> 00:05:58,900 itself, but you can see the sectors quite separate, they are distinct. 73 00:05:59,139 --> 00:06:03,430 In 2008, that's only 5 years later, 74 00:06:03,430 --> 00:06:09,100 you cannot see that distinction, that clear distinction between the 5 economic sectors. 75 00:06:09,199 --> 00:06:13,460 Furthermore, what is happened is that finance linked to 76 00:06:13,460 --> 00:06:16,840 real estate, which is right at the center of it, 77 00:06:16,840 --> 00:06:23,600 which is the dark green, is at the very heart of the global financial system. 78 00:06:23,620 --> 00:06:30,620 What that meant was that the system had become too interconnected, 79 00:06:30,970 --> 00:06:36,280 too interdependent and that had made the system fragile. 80 00:06:36,280 --> 00:06:41,009 So it did not take much to actually topple the system. 81 00:06:41,009 --> 00:06:48,009 And that is the danger of intense interconnectivity and interdependence. 82 00:06:49,780 --> 00:06:56,199 The second finding was that the sectors, as they came together, 83 00:06:56,199 --> 00:06:58,560 changes in one affected the other. 84 00:06:58,560 --> 00:07:04,850 And you can see from this graph how the 5 sectors are actually moving together and this 85 00:07:04,850 --> 00:07:11,850 is just a period of just ten years. 86 00:07:12,080 --> 00:07:19,900 The question is why? Why does this happen? And I hope that by understanding the characteristics 87 00:07:19,930 --> 00:07:26,900 of complex systems we will come to understand why these phenomena actually take place. 88 00:07:28,000 --> 00:07:34,200 One of the first characteristics I want to explain when we move beyond connectivity and 89 00:07:34,240 --> 00:07:36,220 interdependence is feedback. 90 00:07:36,220 --> 00:07:41,710 There are two types of feedback, and I want to actually explain the technical 91 00:07:41,710 --> 00:07:45,789 difference between positive feedback and negative feedback. 92 00:07:45,789 --> 00:07:48,820 Now, counter intuitively, 93 00:07:48,820 --> 00:07:57,000 positive feedback tends to cause system instability, while negative feedback tends to cause system stability. 94 00:07:57,069 --> 00:07:58,700 Let me explain what that means. 95 00:07:58,729 --> 00:08:00,960 First of all let's look at positive feedback. 96 00:08:00,960 --> 00:08:04,410 Now, this is a stampede. 97 00:08:04,410 --> 00:08:09,699 Now imagine what happens when a few sheep are frightened, 98 00:08:09,699 --> 00:08:12,440 they start running, the faster they run, 99 00:08:12,440 --> 00:08:15,009 the more they panic, the more they panic, 100 00:08:15,009 --> 00:08:16,479 the faster they run. 101 00:08:16,479 --> 00:08:22,139 The first few sheep that started will then attract other sheep and more sheep will follow 102 00:08:22,139 --> 00:08:26,289 and you can then see what actually happens. 103 00:08:26,289 --> 00:08:28,080 You get a stampede. 104 00:08:28,080 --> 00:08:32,120 So it is actually making the system unstable. 105 00:08:32,120 --> 00:08:37,919 So positive feedback tends to cause system instability, 106 00:08:37,919 --> 00:08:43,190 and it can very quickly lead to a bank run or even a global financial crisis. 107 00:08:43,190 --> 00:08:49,320 And of course what happens is there is positive feedback in the loss in confidence, 108 00:08:49,320 --> 00:08:50,959 so it feeds upon itself. 109 00:08:50,959 --> 00:08:56,520 A little loss of confidence leads to more loss of confidence and that keeps on increasing. 110 00:08:56,520 --> 00:09:04,500 So positive feedback feeds on itself and makes the difference greater. 111 00:09:05,330 --> 00:09:11,400 So what we see there is of course the number of cattle running, 112 00:09:11,459 --> 00:09:16,080 the overall level of panic increasing and feeding into itself. 113 00:09:16,080 --> 00:09:19,270 Now let's look at the opposite. 114 00:09:19,270 --> 00:09:24,560 Negative feedback tends to make a system self regulating and it can produce stability. 115 00:09:24,560 --> 00:09:27,050 Let me then explain that. 116 00:09:27,050 --> 00:09:31,850 This is a very simple mechanistic system. 117 00:09:31,850 --> 00:09:34,370 It's a ballcock. 118 00:09:34,370 --> 00:09:39,680 So when you press the lever the water will empty, 119 00:09:39,680 --> 00:09:46,680 when the water empties, the valve opens and the water comes in and 120 00:09:47,360 --> 00:09:48,990 it refills the system. 121 00:09:48,990 --> 00:09:55,360 As the water rises, the ballcock rises and then, 122 00:09:55,360 --> 00:10:01,440 when the water reaches the right level, the valve is closed and no more water comes 123 00:10:01,440 --> 00:10:02,250 into the system. 124 00:10:02,250 --> 00:10:10,200 That is a very efficient, very simple system and it has a single equilibrium point. 125 00:10:10,540 --> 00:10:20,300 Now what our mistake is, is that very often we actually make the assumption 126 00:10:20,330 --> 00:10:29,600 that what applies to a very simple mechanistic system will also apply to a complex system. 127 00:10:29,660 --> 00:10:36,660 The assumption there is that the right amount of correction can be applied in the most timely 128 00:10:36,660 --> 00:10:39,240 manner and that is not the case. 129 00:10:39,240 --> 00:10:44,940 The other assumption is that there is a single equilibrium point. 130 00:10:44,940 --> 00:10:51,920 Now an economy, which is a complex system, 131 00:10:51,920 --> 00:11:00,500 may have both positive and negative feedback at the same time and it will have multiple equilibria. 132 00:11:00,560 --> 00:11:04,700 Not just a single equilibrium point. 133 00:11:04,700 --> 00:11:10,089 Now, what we then need to understand is the next 134 00:11:10,089 --> 00:11:13,190 principle of emergence. 135 00:11:13,190 --> 00:11:15,420 Now, most of us, 136 00:11:15,420 --> 00:11:22,420 if you have done any kind of reading in complexity, you will have come across emergence. 137 00:11:22,709 --> 00:11:29,470 And the idea of emergence is that individual agents, 138 00:11:29,470 --> 00:11:34,000 this could be individual people, as I said, 139 00:11:34,000 --> 00:11:40,370 groups etcetera, interacting together create something which 140 00:11:40,370 --> 00:11:54,300 is both unpredictable and it has a bottom-up effect. 141 00:11:54,300 --> 00:11:57,990 But that is only half the story. 142 00:11:57,990 --> 00:12:01,459 Most of us I think are aware of only half the story. 143 00:12:01,459 --> 00:12:05,570 Research on the brain has actually shown us a second process. 144 00:12:05,570 --> 00:12:11,329 The second process said that once the emergent comes into being, 145 00:12:11,329 --> 00:12:16,120 there are two things the happen. 146 00:12:16,120 --> 00:12:21,589 It affects the agents in two ways, it can both constraint certain behaviors, 147 00:12:21,589 --> 00:12:24,540 while at the same time it can open up new possibilities. 148 00:12:24,540 --> 00:12:27,110 Let me give you an example. 149 00:12:27,110 --> 00:12:29,959 Culture is an emergent process. 150 00:12:29,959 --> 00:12:35,720 It arises through the interaction of everyone in a particular organization or society. 151 00:12:35,720 --> 00:12:41,700 I am a member of the London School of Economics, there are certain things I would not dream of doing. 152 00:12:41,740 --> 00:12:47,800 In other words, my behavior is constraint through the emergent 153 00:12:47,890 --> 00:12:49,790 process which is the culture. 154 00:12:49,790 --> 00:12:54,279 However, at the same time there are doors open to me 155 00:12:54,279 --> 00:12:57,890 because I'm a member of the London School of Economics, 156 00:12:57,890 --> 00:13:01,760 that would not be open to me as an individual. 157 00:13:01,760 --> 00:13:07,660 So can you see when we put the two processes together how dynamic that process is. 158 00:13:07,660 --> 00:13:18,300 So we got individual agents interacting and creating the emergent at the macro level. 159 00:13:18,380 --> 00:13:25,100 But once the emergent has been created it then affects the interacting entities in those 160 00:13:25,120 --> 00:13:26,640 two different ways. 161 00:13:26,640 --> 00:13:33,640 So we have that constant very dynamic process happening. 162 00:13:35,279 --> 00:13:42,279 So emergent properties can be processes, can be qualities can be patterns. 163 00:13:42,730 --> 00:13:49,560 They arise from interaction and they cannot always be predicted. 164 00:13:49,560 --> 00:13:51,519 They are not additive or cumulative. 165 00:13:51,519 --> 00:13:53,529 Let me explain what that means. 166 00:13:53,529 --> 00:13:56,450 If you were to take a group of people and ask them, 167 00:13:56,450 --> 00:14:00,370 invite them, to a brainstorming session. 168 00:14:00,370 --> 00:14:05,029 What comes out of that brainstorming session will be quite different then if you were to 169 00:14:05,029 --> 00:14:09,260 take exactly the same people, put them in separate rooms, 170 00:14:09,260 --> 00:14:15,420 give them exactly the same question and then compare the outcomes. 171 00:14:15,420 --> 00:14:26,400 You cannot add the individual answers and come to the same outcome as you will get from 172 00:14:26,450 --> 00:14:31,800 the group working together in the brainstorming session. 173 00:14:31,839 --> 00:14:36,320 So what that means is that emergence is a systemic property , 174 00:14:36,320 --> 00:14:41,760 it is a property of the system working together, interacting together, 175 00:14:41,760 --> 00:14:44,070 to create the emergent property. 176 00:14:44,070 --> 00:14:46,600 And is not additive or cumulative, in other words, 177 00:14:46,600 --> 00:14:50,620 it is more than the sum of the parts. 178 00:14:50,620 --> 00:14:52,610 It is also, think about it, 179 00:14:52,610 --> 00:14:59,610 as a process of transition from micro agent interaction to macro-structures, 180 00:15:00,240 --> 00:15:04,500 and macro-structures are the emergent processes, qualities, patterns. 181 00:15:04,510 --> 00:15:05,900 Learning, culture, 182 00:15:05,990 --> 00:15:12,990 innovation are all emergent processes, but also new ways of organizing, 183 00:15:13,380 --> 00:15:19,089 and new organizational forms can also be emergent. 184 00:15:19,089 --> 00:15:24,040 When we look at the challenges of managing complex systems, 185 00:15:24,040 --> 00:15:29,529 we will then focus on that much more. 186 00:15:29,529 --> 00:15:36,180 So to summarize, what we have looked at are the four basic 187 00:15:36,180 --> 00:15:41,240 characteristics of complex systems. 188 00:15:41,240 --> 00:15:45,130 These have already been articulated by systems theory, 189 00:15:45,130 --> 00:15:47,550 which are connectivity, interdependence, 190 00:15:47,550 --> 00:15:49,480 feedback and emergence. 191 00:15:49,480 --> 00:15:54,329 And in the next video we will look at the other characteristics, 192 00:15:54,329 --> 00:15:56,870 which arise from the theories from complexity. 193 00:15:56,870 --> 00:15:57,689 Thank you.