The Berkeley High Jacket


Newsletter

The best of the Jacket, delivered to your inbox.

News Print
November 20, 2026 Login

Math fields vol. VI: The even bigger numbers: Googology II

Niko Stephens on November 20th, 2026

Last time, we left off at a little known number called Graham’s number, which we got using a power tower repeatedly over and over to achieve higher and higher numbers, until we hit an arbitrary limit. But that number is small. We can go much, much bigger than that. 

To do that, we first must begin with a little game about trees in a forest. In this game of TREE(k), you use nodes of varying colors to make as many trees as possible. A tree is a construction of colored nodes and edges with no loops, where each node has exactly one path to get to the first node at the top, aka the root (orientation does matter, as flipping a tree upside down makes it different). 

To start, you first pick a number, k. This number is how many different colors of nodes you get. Now you must construct as many trees as possible. But there are some rules you must follow while making your trees. Each tree can have at most as many nodes as its turn number, so tree 1 has one node, tree 2 can have up to two, etc. Also, when you make a new tree, it must not contain any previous trees. This includes both color and ancestry, so if, say, tree 2 had a root with some color and a child node of a different color, there will never be a node with that different color that has descended from a node with the first color. 

Now that we know the rules, we can play. If we pick TREE(1), we can make one tree, as we only have a red node to work with, and we can only use it once, so our game ends there. TREE(2) lets us use two colors, red and green, so we start with one red node. We no longer can use red, so we have “burned” it. We can make the second tree two greens, taking full advantage of being able to use two nodes. Our third tree must only contain one green, as if we do any trees that are just made up of two or more greens, it will just contain tree 2, so the game ends there. TREE(3), however, unlike the previous two steps, is much bigger. 

So big, in fact, that Graham’s number is basically zero compared to it. The only way we know that the number is finite is because of Kruskal’s tree theorem. There are so many numbers in between the two that we have skipped, but we must continue. The last number we will talk about today is one of the biggest well-defined non-salad numbers that we have: Large Number Garden Number. Extremely oversimplified, the function f(n) is a “number-hunting machine” that searches through an endless tower of increasingly powerful math universes to add up every number you can describe in (n) symbols or fewer, repeated 10 times to itself with an input of 10 ^10. 

This number is so big that it is literally just uncomputable, which is where we encounter the Halting problem, so there is no program we can write to get there. Also, because of Gödel’s incompleteness theorems, any mathematical system strong enough to define numbers at this level cannot prove its own consistency, meaning we can never be 100 percent mathematically certain that LNGN doesn’t hide a contradiction.