(The ideas in the first part of this post on categories was heavily influenced by Scott Alexander’s essay on categories. It’s fascinating, and I’d recommend you read that first.)
In the world of close-up and stage magic, many people have, at different points in time, attempted to classify magic into numerous categories, from magnetism to divination. It’s interesting to see how these intuitive categories jar with what we know about physics today. Granted, many of these efforts took place a while ago, but scientists have had fairly good approximate models of reality for a long time (thanks to Galileo, Newton, etc.).
For example, magician Dariel Fitzkee differentiates between productions, vanishes, restorations, and penetrations. At their heart, though, all four of these involve manipulating the presence or absence of matter.
This is interesting because, despite the fact that these categories don’t conform to the fundamental way things are, they still correspond to some notion of “different intuitive things”. Making a coin vanish elicits a different reaction than causing a playing card to change color, which in turn feels far away from cutting and restoring a length of rope. Our shared sense of human culture probably plays something into this.
Though there is still vigorous debate over the taxonomy of magic today, it’s fair to say that Fitzkee’s work (and others) has proven to be a key cornerstone in magical theory. It’s inspired many more ideas, and perhaps the resulting effects are where much of its value comes from.
Why does classification seem to be so useful?
We use it in just about every field, from poetry analysis to Pokemon typing. The most obvious reason seems to be that our world is filled with lots of different things, at least on a level we care about. Because of this variation, things interact with other things differently. We may find it useful to classify algorithms with different orders of growth, for example, because we have time and space constraints, and our classification helps with this. The space and time partitions we use, in turn, may also be defined in terms of other things. That’s all fine, as the categories make things easier on us.
There’s a general sense of abstraction happening here, where we creating categories helps us navigate what would otherwise be a very variegated mix of experience. By instead moving through a world of “floors”, “people”, “plants”, and other “objects”, we reduce unnecessary complexity, making it easier to focus on the important things in life.
Perhaps more interestingly, though, it appears that most of the thoughts I have fall into one of just a few categories, and classification is one of them:
In Defense of the Obvious basically says that there exist some causal structures that lead following obvious advice to good outcomes. It’s showing that a causal link exists between several items.
Overriding Defaults is about how we can classify a group of apparently unrelated behavior under a common label. This one is simple classification.
Thingspace of Thingspace, in essence, draws parallels between 3-dimensional space and our intuitions about the world. A metaphor is the key thought here.
Mildly echoing Hofstadter who posits that analogies are at the heart of everything, it looks like causal links, classification, and analogy form the bulk of the “whoa” insight-porn-y thoughts that come out of my brain. Given our deterministic universe, as well as our tendency to care about future selves, causal links make a lot of sense. I also sort-of plausibly explained away classification. And I’ll let Hofstadter’s own words explain the whole analogy thing.
But what about all the other type of thoughts I could have? Introspection seems to be a different thing altogether; imaginations and daydreams are also fairly alien. For the most part, however, it seems like many thoughts can be easily shoved into one of the above three areas.
As my friend Julian points out, though, all my analysis of the structure of these thoughts happens after the fact. Playing “mad-libs” with these skeleton structures will likely not yield anything fruitful. In hindsight, it’s easy to draw targets around the arrows themselves.