"So much for the tolerant left."
Let's first imagine a society of many cells. These cells might be individual people or they might represent 'cells' of people in a region.
In the first iteration, there are four types of cells: Neutral cells (light gray), minority group cells (dark blue xCells), Intolerant Cells (red iCells), and Tolerant Cells (green tCells). Both tCells and iCells "spread" to adjacent cells, but xCells cannot become intolerant of themselves.
If we agree that intolerance has a higher virality than tolerance, then even a relatively small starting set of iCells eventually results in the destruction of the 'tolerant' society. In the below simulation, the society is seeded with fewer than a quarter as many iCells as tCells.
I wanted a better way of modeling intolerance upon target cells, so let's modify our society by having cells "swap" with willing cells away from positions where they are facing intolerance, to positions of less intolerance.
This relatively simple mechanic results in the striking visual effect of cells "fleeing" intolerance.
While the outcome of this simulation is no different, it's perhaps a consolation that the minority xCells form stable groups within the intolerant society.
Now let's introduce a 5th type of cell, cells that are intolerant of intolerance (light blue, iiCells). They spread their belief with the same virality as intolerance of xCells. Based on the same logic that xCells cannot become intolerant of themselves, iCells cannot be converted to iiCells, and vice versa.
We'll begin with just slightly fewer iiCells than iCells.
While none of these simulations account for deradicalization or other processes of reducing existing intolerance, this is the only simulation where a 'tolerant society' survives, protected by the cells that reject intolerance.
This is a simple model, with light-weight mechanics, which illustrates a phenomenon I hold to be self-evident. Obviously real societies are more complicated, and I make no claim to perfectly representing these processes as they occur in life.
Just for starters: In this society, every cell is neighbored by exactly 8 other cells (the border cells wrap to the other side)—in other words, the universe is uniformly dense. In most real societies, there would be pockets and bubbles connected to the universe as a whole by only a few cells.
The minority cells also don't propagate. And like I mentioned above there is no version of the simulation where large pockets of intolerant cells can be converted to tolerant cells. The iiCells are effectively inoculated against becoming the thing they will not tolerate just as the iCells are inoculated against becoming intolerant of themselves. iiCells can spread tolerance as well, but they can only spread intolerance of intolerance to neighboring tolerant or neutral cells.
Perhaps a more philosophical flaw in the model is that there is no adjustment to be made that produces what we have seen in the long arc of the moral universe: that tolerance does seem to gradually win out over intolerance. Probably, this is at least partly a function of greater education and security... both of which are out of scope for this project.
If you'd like to propose a change or report a bug, the code is hosted on GitHub.