Graph theory begins with a beautiful legend. I have no reason to suppose it’s false, except my natural suspicion of beautiful legends as origin stories. Its organization as a field is traced to 18th century Köningsburg, where seven bridges connected the banks of a river and a small island in the center. Whether it was possible to cross each bridge exactly once and get back where one started was, they say, a pleasant idle thought to ponder and path to try walking. Then Leonhard Euler solved the problem. It’s impossible.
Graph theory arises whenever we have a bunch of things that can be connected. We call the things “vertices”, because that’s a good corner-type word. The connections we call “edges”, because that’s a good connection-type word. It’s easy to create graphs that look like the edges of a crystal, especially if you draw edges as straight as much as possible. You don’t have to. You can draw them curved. Then they look like the scary tangles of wire around your wireless router complex.
Graph theory really got organized in the 19th century, and went crazy in the 20th. It turns out there’s lots of things that connect to other things. Networks, whether computers or social or thematically linked concepts. Anything that has to be delivered from one place to another. All the interesting chemicals. Anything that could be put in a pipe or taken on a road has some graph theory thing applicable to it.
A lot of graph theory ponders loops. The original problem was about how to use every bridge, every edge, exactly one time. Look at a tangled mass of a graph and it’s hard not to start looking for loops. They’re often interesting. It’s not easy to tell if there’s a loop that lets you get to every vertex exactly once.
What if there aren’t loops? What if there aren’t any vertices you can step away from and get back to by another route? Well, then you have a tree.
A tree’s a graph where all the vertices are connected so that there aren’t any closed loops. We normally draw them with straight lines, the better to look like actual trees. We then stop trying to make them look like actual trees by doing stuff like drawing them as a long horizontal spine with a couple branches sticking off above and below, or as * type stars, or H shapes. They still correspond to real-world things. If you’re not sure how consider the layout of one of those long, single-corridor hallways as in a hotel or dormitory. The rooms connect to one another as a tree once again, as long as no room opens to anything but its own closet or bathroom or the central hallway.
We can talk about the radius of a graph. That’s how many edges away any point can be from the center of the tree. And every tree has a center. Or two centers. If it has two centers they share an edge between the two. And that’s one of the quietly amazing things about trees to me. However complicated and messy the tree might be, we can find its center. How many things allow us that?
A tree might have some special vertex. That’s called the ‘root’. It’s what the vertices and the connections represent that make a root; it’s not something inherent in the way trees look. We pick one for some special reason and then we highlight it. Maybe put it at the bottom of the drawing, making ‘root’ for once a sensible name for a mathematics thing. Often we put it at the top of the drawing, because I guess we’re just being difficult. Well, we do that because we were modelling stuff where a thing’s properties depend on what it comes from. And that puts us into thoughts of inheritance and of family trees. And weird as it is to put the root of a tree at the top, it’s also weird to put the eldest ancestors at the bottom of a family tree. People do it, but in those illuminated drawings that make a literal tree out of things. You don’t see it in family trees used for actual work, like filling up a couple pages at the start of a king or a queen’s biography.
Trees give us neat new questions to ponder, like, how many are there? I mean, if you have a certain number of vertices then how many ways are there to arrange them? One or two or three vertices all have just the one way to arrange them. Four vertices can be hooked up a whole two ways. Five vertices offer a whole three different ways to connect them. Six vertices offer six ways to connect and now we’re finally getting something interesting. There’s eleven ways to connect seven vertices, and 23 ways to connect eight vertices. The number keeps on rising, but it doesn’t follow the obvious patterns for growth of this sort of thing.
And if that’s not enough to idly ponder then think of destroying trees. Draw a tree, any shape you like. Pick one of the vertices. Imagine you obliterate that. How many separate pieces has the tree been broken into? It might be as few as two. It might be as many as the number of remaining vertices. If graph theory took away the pastime of wandering around Köningsburg’s bridges, it has given us this pastime we can create anytime we have pen, paper, and a long meeting.