The Roman Philosopher Lucius Anneaus Seneca (4 BCE-65 CE) was perhaps the first to note the universal trend that growth is slow but ruin is rapid. I call this tendency the "Seneca Effect."
Showing posts with label networks. Show all posts
Showing posts with label networks. Show all posts

Monday, November 22, 2021

The Mousetrap Experiment: Modeling the Memesphere

 Reposted with some modifications from "The Proud Holobionts"

 Ilaria Perissi with our mousetrap-based mechanical model of a fully connected network. You can find a detailed description of our experiment on ArXiv

You may have seen the "mousetrap experiment" performed as a way to demonstrate the mechanism of the chain reaction that takes place in nuclear explosions. One of its earliest versions appeared in Walt Disney movie "Our Friend, the Atom" of 1956. 

We (myself and Ilaria Perissi) recently redid the experiment with 50 mousetraps and 100 wooden balls. And here it is. It was fun, except when (and not so rarely) one of the traps snapped on our fingers while we were loading it.

But why bother redoing this old experiment (proposed for the first time in1947)? One reason was that nobody had ever tried a quantitative measurement. That is, measuring the number of triggered traps and flying balls as a function of time. So, we did exactly that. We used cell-phone slow motion cameras to measure the parameters of the experiment and we  a system dynamics model to fit the data. It worked beautifully. You can find a pre-print of the article that we are going to publish on ArXiv. As you can see in the figure, below, the experimental data and the model go reasonably well together. It is not a sophisticated experiment, but it is the first time that it was attempted.

But the main reason why we engaged in this experiment is that it is not just about nuclear reactions. It is much more general and it describes a kind of network that's called "fully connected," that is where all nodes are connected to all other nodes. In the set-up, the traps are nodes of the network, the balls are elements that trigger the connection between nodes. It is a kind of communication based on "enhanced" or "positive" feedback.

This experiment can describe a variety of systems. Imagine that the traps oil wells. Then, the balls are the energy created by extracting the oil. And you can use that energy to dig and exploit more wells. The result is the "bell shaped" Hubbert curve, nothing less!  You can see it in the figure above: it is the number of flying balls "produced" by the traps.

We found this kind of curve for a variety of socioeconomic system, from mineral extraction to fisheries (for the latter, you can see our (mine and Ilaria's) book "The Empty Sea." So, the mousetraps can describe also the behavior of fisheries and have something to do with the story of Moby Dick as told by Melville.

You could also say the mousetrap network is a holobiont because holobionts are non-hierarchical networks of entities that communicate with each other. It is a kind of holobiont that exists in nature, but it is not common. Think of a flock of birds foraging in a field. One bird sees something suspicious, it flies up, and in a moment all the birds are flying away. We didn't have birds to try this experiment, but we found a clip on the Web that shows exactly this phenomenon.

It is a chain reaction. The flock is endowed with a certain degree of intelligence. It can process a signal and act on it. You can see in the figure our measurement of the number of flying birds. It is a logistic function, the integral of the bell-shaped curve that describes the flying balls in the mousetrap experiments

In Nature, holobionts are not normally fully connected. Their connections are short-range, and signals travel more slowly through the network. It is often called "swarm intelligence" and it can be used to optimize systems. Swarm intelligence does transmit a signal, but it doesn't amplify it out of control, as a fully connected network does, at least normally. It is a good control system: bacterial colonies and ant colonies use it. Our brains much more complicated: they have short range connections but also long range ones and probably also collective electromagnetic connections. 

One system that is nearly fully connected is the world wide web. Imagine that traps are people while the balls are memes. Then what you are seeing with the mousetrap experiment is a model of a meme going viral in the Web. Ideas (also called memes) flare up in the Web when they are stimulated it is the power of propaganda that affects everybody.

It is an intelligent system because it can amplify a signal. That is that's the way it reacts to an external perturbation. You could see the mousetraps as an elaborate detection system for stray balls. But it can only flare up and then decline. It can't be controlled. 

That's the problem with our modern propaganda system: it is dominated by memes flaring up out of control. The main actors in this flaring are those "supernodes" (the Media) that have a huge number of long-range connections. That can do a lot of damage: if the meme that goes out of control is an evil meme and it implies, say, going to war against someone, or exterminating someone. It happened and keeps happening again as long as the memesphere is organized the way it is, as a fully connected network. Memes just go out of control.

All that means we are stuck with a memesphere that's completely unable to manage complex systems. And yet, that's the way the system works. It depends on these waves of out-of-control signals that sweep the web and then become accepted truths. Those who manage the propaganda system are very good at pushing the system to develop this kind of memetic waves, usually for the benefit of their employers. 

Can the memesphere be re-arranged in a more effective way -- turning it into a good holobiont? Probably yes. Holobionts are evolutionary entities that nobody ever designed. They have been designed by trial and error as a result of the disappearance of the unfit. Holobionts do not strive for the best, they strive for the less bad. It may happen that the same evolutionary pressure will act on the human memesphere. 

The trick should consist in isolating the supernodes (the media) in such a way to reduce their evil influence on the Web. And, lo and behold, it may be happening: the great memesphere may be rearranging itself in the form of a more efficient, locally connected holobiont.  Haven't you heard of how many people say that they don't watch TV anymore? Nor they open the links to the media on the Web. That's exactly the idea. Do that, maybe you will start a chain reaction in which everyone will get rid of their TV. And the world will be much better.