[MUSIC] Hi, welcome back to the third session of the lesson about terrible simplifiers. Again, we're giving examples of problems that you find when you're trying to crunch number in relation to the characteristic of a complex system. In this session, we will deal with the issue of scale. We want to show that the usefulness of the representation of a complex system depend on how we can handle the challenge force by scale. In relation to locate things or describe things, that in space and time. The problem is generated by the fact that complex system are operating simultaneously across different levels of organization and they can only be observe using different scale, that which imply different description domain. Since this seems a little bit esoteric, I will try to give you examples to show practical example of this problem. So let's imagine that I show you this object and ask you what this object is? And then, second, how big is this object? Looking at that, it could be a component of a radio or some circuit, or something else. If I put in relation this object to something else, I can detect that this is a remote sensing picture of a place, if you know the place is the University of Arizona. So this is the auditorium of the top part of the auditorium of the Arizona State University. Unless you are contextualize the image, so there's no way that you can map this recognition so the field that deal with the issue of having multiple events and multiple descriptive domain depending on the scale is called Hierarchy Theory. And one of the first statements saying a strong thing about that was done by Mandelbrot, the father of the fractals. And he said, you cannot measure the land of a segment of coastal line, you should don't first define the scale of the map that you will be using. In the sense, the measurement depend on your political choice of a scale at which I represent your object. If you're having a very rough representation of the UK, then you will have a short coastal line. If you want to have a very detailed representation of the coastal line. Then you will have a much longer a coastal line. He published it very, very, very nice, in which he was showing that the assessment of the border between two country is different. And the smaller country as an assessment of the border with the other country, which is longer because they're using scale, they are more detailed. And so, what it is the point, that if I go to very low, a high, very high resolution. The level of meters, then I cannot even define what is the coastal line because the coastal line is moving. It's moving because of tides, and then it's moving because of waves. So, at the level of centimeter, the land of the coastal line cannot be defined because it's no longer there. So this bring us to the definition non equivalent descriptive domain, depending on what type of animal you're focusing on, in order to be able to detect some element you will have to use a matter of recognition or a presentation and may imply you can no longer see all things. For instance, if I taking picture of a person with a microscope, I can take zillion of pictures but they will never see the nose of that person. Let's have an example, let's imagine that you are asked to get a person to the airport you don't know, so you ask can you send a picture of this person. And you'll get this picture and say okay, this is the picture of the person that you have to pick up at the airport. And of course this is a cell. It's a part of the, I believe that I got it in my old institute. It's a [INAUDIBLE] part of an intestine or something like that. But, however, this type of a picture is not useful for picking up the person at the airport. So you will write again and say, nice job, guys, can you send me a picture with a large domain, so that I can recognize the person. And you will get this picture and say, the person you have to pick up at the airport is indicated by the white arrow is this person here? Again, you can write and say no, no, no joke please can you send me the photo of the head of the person. You can get this type of picture and again this is not the picture that you want to have to pick up a person at the airport. The picture you want to have is this one, so you can recognize the person when you see here. What is the meaning of this presentation? If you want to know how a person is working in terms of intestine. [LAUGH] The internal functioning of the organs. This is the right type of picture you want to have. If you want to understand the social behavior, this is a rally of the green party so you understand that this person is involved in protecting the environment. But even this type of picture is not good for getting a person recognizable at the airport. But you can detect that this person as a matter of fact this is the x-ray of my wife. What is the issue here? Depending on the descriptive domain that you're using, you will see things that you would not see other things. And this is important. Be aware that every time you manage to finally get what you're looking for. What is the face of the person to get at the airport? You no longer can add information about all the others type of information you can get by using different descriptive domains. So if you're using two non equivalent descriptive domain, it is because you're using two different perceptions of the world. What we will call different narratives in the next lesson. What is the problem that if you're using different perception of the world, not is not possible to have a reduction of [INAUDIBLE] representation to one another. Let's say into this [INAUDIBLE] domain for example, the monetary value of teddy bear economic reading cannot be used to describe the sentimental value of for the child. You cannot have the value of your mother in monetary terms, because it would be at also the central is associated with the fact that every time we're defining a before and after. Or topological relation, we're adopt the referential framework. And the referential framework will define whether or not we can translate our assessment to another representation. Let's imagine to have an example, we can say that Maine which is a state of the United States is this one on the East Coast. And we can make an experiment scientific experiment we can call 500 people randomly selected on the phone book of Maine from long run ask what time is it, and you can define that this is on the East Coast. At the same time, you can go inside Maine at the level of Lincoln County and the fact that the coastal line there is oriented to the south you can get satellite picture and define that. Then you can go inside Lincoln County and then find a village, in this case this is Colonial Pemaquid is a real place I used to go there on vacation. And you can see that the orientation of the costal pemaquid is east toward west. Finally you can go to a particular beach of Colonial Pemaquid and Polly's beach. And find that the orientation of the course align east toward north. Let's imagine you are at this point scientific controversy because different scientists assign the of. This sounds a north in the west. And every scientific experiment will not solve the controversy. You could have a person going under beach with a compass in their hand saying, I am really grand on feet on the ground. I can tell you that this is absolutely no and I'm telling you I was mentioning with a compass. Another person here, an ecologist, can go look at the trunks of the trees and see where is the orientation. That's the matter, it's not the problem of the type or the certification of the state of democracy here. The problem here, that you have a puritanical definition of what is coastal line, and what is the orientation of coastal line, which is not a clear one. If you're looking for a scientific tool that tells you, what is the right time to make a phone call to Maine from Los Angeles. Then Maine is on the east coast. If you're looking for buying a house. Is the porch looking at the sunset? That, the real type of information is the one provided by the compas. But these as nothing to do so whether or not the statement is true because you could have different statement through interaction to different definition of orientation of coastal line. Where you get into the impossibility of reducing considering more those because when we're losing the narrative of the iscos We are assuming that the space is a globe. Mainly on the east coast because there is an east and the west coast in the United States over a global world. Whereas, if you're defining with the compass where is no other. You're assuming that the space is a plane. Is flat, and which basically, these meridian touch here and here, they have are parallel. So you have an irreducible difference in the definition of your topological relation of the reference, at the same happens if you're trying to define time using external references. So, you could have winter, spring, summer, and fall. And then you could have different grain in extent, evening, night, morning, and afternoon. Then you could have on another scale a 24 hours clock with one hour grain. Or we can have with that. What happened here? You cannot say, if evening is coming before spring because evening and spring are incommensurable types in relation to relation to their relative position in simple time. Each of these a clock, a sand clock, all the seasons, all the weeks with the calendar. Are a framework to make possible to define at before and an after. But they have to be integrated in order to be able to say whether or not. Friday is before or after summer or if May, is before or after 3:45 AM. We need the possibility of integrating the perception of before and after came in from different, theoretical framework of an axis. So this is the last slice and is the final lesson, that if you want to deal with a complex system, you have first of all to focus on relation across descriptive domains before then working on relations within descriptive domains. And this is what reduction is [INAUDIBLE]. The reduction works only within descriptive domain, at a given scale with a given variables and this is really wrong. One way of looking at it is imagining a fresco. This is a huge fresco in Sienna, and you can see this is an allegory of the good and bad government. In this case, it's a good government. And you have in this fresco houses, you have people, you have different groups of people doing different things. And let's imagine that you want to paint this fresco, what you do? To paint this fresco you have different type of problems here. While you put your houses, when you do your horses. And then, you have this group of people here for instance they're doing something. And then, of course if you are painting these things you have, as well, be able to paint the hand of this person. So what is the problem? That in order to have a fresco which is nice to see and is well done, you have, first of all, continuously to have an idea how big are the houses compared with the horses. How big are the horses compared with the people there, how is the relationship among the people at this level, what are the relation of the finger with the hand at this level. But in order to do this operation in a harmonical way, you have to continuously move at the beginning to define in big terms, know how big are the houses, how big are the people, how big is a group how big is the hand within the group? And only after having an idea of this, you can start getting into details. So the problem is that reduction and you have a questions a mother's who have very good analysis of the hand and relation a mother to do how these people are interacting. Or but what you want to do before getting the fine tuning of quantitative analysis is be able to locate where the different looks are doing what they're doing and how big are they compared with their owners. And whether or not you're capable of providing and integrate overview of the different parts
