Let’s map out a 2D space for making decisions. Along one axis there’s every object, and along the other axis is every possible action you could choose. Most of these combinations will be impossible. For example you could combine “glass of water” on the first axis with “drink” on the other axis and it would make sense. Or combine “Chicago” and “go to” and it would make sense. But you can’t drink Chicago. Talking about axises might make it sound like we’re going for some formal mathematical description of making choices, but we’re not. There’s just too many possibilities along each axis. We’re not going to be calculating anything, this is just a tool for discussing the ideas that I’ve defined elsewhere. And to be able to discuss them using consistent ideas and terminology.
So we have our 2d plane of every possible action we could take. Now let’s extend along the third axis, this will be our expected benefit (or cost) of taking that action. We can imagine it as a landscape where we’re trying to find the highest point, assuming of course that similar object/action combinations are clustered near each other with similar kinds of outcomes. Again, most of the landscape will be impossible to reach, it’s made of impossible combinations. And a lot of the landscape will be valleys, all the choices that will make us worse off.
And we can extend this idea one step further, at each point where we make any choice, any combination of action and object, we can imagine traveling to that point and rising or falling to whatever height it occupies. And then from that point we can repeat the process, creating a new grid of all possible choices from that point, and traveling to another point, and repeating to end up with a string of actions. This might be a plan or a solution to a problem or a random walk. But we can sum together the outcomes of these stacked landscapes to end up with a net impact from the choices taken in order.
We can imagine it like playing chess, but the grid we’re looking at isn’t the board, instead on one axis is all of the pieces (all the objects in the world), and on the other axis is every square on the board (every possible action). Most combinations will be impossible, no piece can move to any square. But for every combination where one of our pieces could be moved on our turn there’s an outcome, higher for an improved position and lower for a worse one. All the possible combinations of moves creates a stacked series of outcomes that show the change in likely outcomes as we follow that series. It’s forming a 4d chess board (and I’m aware of the idea of 4d chess, but I still like this description).
Now that we have this idea, a series of points of all possible choices connected together to form a series of actions, we can define some ideas based on this framework.
- Imagination: a series of linked choices that seem like they could be possible in some scenario
- Plans: a subset of all imaginable series of choices, ones that seem actually possible in our current scenario
- Concept: a way of grouping together points along one or more axises. For example the concept of “trees” would be every point that includes an object that shares the characteristics of being woody and perennial. Or we could look at the concept of a single tree, and this would be a series of points of every action that could be taken with a single object, and it would be all the linked actions forwards and backwards, the things that could’ve potentially created the object now, and the actions that could happen in the future. And we can group potential actions together too, like the concept of “traveling” would include a subset of all possible actions. A concept would also includes a series of steps forming a plan or solution to a problem
- Intelligence: the process of picking a potential series of actions to solve a problem. Obviously it’s impossible to compare them all, it’s just mathematically too many combinations. Instead intelligence means picking two series of points and comparing them, and then repeating that process to try and find a better and better plan/solution.
- There’s two obvious factors that would make someone more or less intelligent:
- How quickly you can make comparisons between two concepts:
- How complex of concepts you can hold in mind and compare at once
- If you can quickly compare a lot of individual steps, that might be one way to be intelligent. But also you could compare two entire complex series of steps at a time, and even if that was slower, it might let you evaluate more possibilities, faster.
- Also, there’s the idea of creativity, or figuring out which series of steps might be possible and comparing new ones that might not get considered otherwise.
If we flatten our landscape from 4d down to 3d to make it easier to imagine, we can think of it like shining a spotlight on a dark series of hills and valleys. A small spot light that can look at a small amount of slope, but moves quickly might be just as good as a large spot light that can illuminate a larger area but moves much more slowly. We can also extend this idea of a person being intelligent by searching the landscape to a group trying to work together to find the best solution/highest point. How quickly they can search, how well they can communicate what they find, and how far they spread out, without being so dispersed that they miss important features, re all ways that groups can be better or worse at solving problems.