People manage to successfully reason using the heuristic rules describing the behavior of macroscale things and events, not needing to derive their behavior from the fundamental laws of nature, and the world seems to behave in a way that allows it. How does this happen? This is not just a consequence of the objects following the simple laws of physics. A door consists of tremendous amount of atoms, yet when I close it, the door just closes. On the high level of description, it is a very simple object. It doesn’t occasionally grow legs and run away, it simply closes.
When trying to extinguish fire, one can try to pour gasoline on it. The fire gets extinguished by water, which is also a liquid. So the gasoline will work as well, won’t it? This is magical thinking, when one acts on surface similarities, and expects the similar consequences. But for objects consisting of a tremendous amount of underlying elements, what is the chance of surface similarities indicating similar behavior in other respects?
The rules of thumb work, because the objects that have the similar causal structure get repeated over and over, resulting from the common cause. The similarity in many properties doesn’t just follow from the presence of few superficial similarities. If I see a flying duck, it might be a robot assassin duck sent by aliens to get me, it just is improbable to the point of being ridiculous. The same cause, the physical process of development from the duck zygote, leads repeatedly to the same system appearing over and over again. By surface similarities I recognize not the other properties that follow from the properties I observed, but a certain causal pattern from the set of causal patterns that exist in my environment and which I learned to recognize in my experience. After the causal pattern is recognized, I can conclude that it has certain other properties.
The examples of repeated patterns in the environment come from many sources. The laws of physics tend to preserve the form of rigid objects, so that the same object will remain in the future, that was there in the past. The stars and other cosmic objects get formed over and over by the same laws of physics during the cosmological development of the universe. On out planet, biological organisms replicate, and so their forms get repeated many times. They mutate, and so slightly different versions of the same system may share many properties. Human culture propagates the ideas and technologies, and so many objects sharing the same properties get constructed for similar purposes. We are surrounded by objects resulting from the common causal processes, and so we are used to seeing the similar consequences from superficially similar events, even if it’s possible to construct systems with the same surface properties that act in an entirely different way.
Applying gasoline to extinguish fire confuses the recognition of the type of causal structure of an artifact with direct recognition of implied properties. Even though both gasoline and water obey the same underlying laws of physics, the causal processes happening during the chemical interactions involved are very different, and so the consequences of applying superficially similar actions to them often won’t result in sufficiently similar consequences. When I don’t expect the object that looks like a duck to be a robot assassin disguised as a duck, I rely on the identification of known causal pattern (a real duck) by few of its observable properties. If I had evidence (indirect experience) of there being in fact robot assassin ducks, I wouldn’t be so sure.
The rules of thumb can be formed about the relationship of many events occurring in sufficiently similar contexts. Groups of events may imply other events, both in the past and in the future, and due to the universality of laws of nature, rules of thumb tend to be reusable, at least where the expected repertoire of causal structures doesn’t change too much.
November 4, 2008 at 5:10 pm |
Hi Vladimir, I’m going to start posting some comments on your blog because I want to get a better understanding of your ideas. I have some catching up to do, though, so I’ll start with old posts.
This post is very interesting, in particular that heuristics generalize best when they apply across phenomena with similar causal histories — and that many things we actually interact with in our world do share deep causal histories (e.g. all ducks), and those are often visible in surface similarities. I think that’s a brilliant point.
I’m not quite sure where you are going with it exactly, though. In particular, in your gasoline vs water example, it seems like the inappropriateness of throwing gasoline on a fire to extinguish it is not really related to their different causal histories… unless perhaps you are saying that unrelated causal histories make the application of a rule of thumb inappropriate in general. So “throw liquid on a fire to extinguish it” is a poor generalization even though for many liquids it works? And, liquid-like substances such as sand, are fine for similar reasons (cutting off oxygen supply to the combusting material). Maybe that is what you mean by “causal structure” but since it is context-specific I’m having a difficult time figuring out how to think about the general case.
Also, I would like to get to the core of your thoughts about causality — what it means and how an intelligence should reason about it. Where would you suggest I look for that?
Thanks
November 4, 2008 at 5:21 pm |
Hmm, one more thing. I’m not sure it is “magical thinking” to guess that gasoline would put out a fire. There is a causal reason that liquids put out fires and therefore generalizing that to all liquids seems reasonable to me. In this case, it turned out that the other liquid was flammable itself (oops!) so it turned out to be incorrect but it doesn’t seem unreasonable.
More unreasonable perhaps would be thinking that water is colorless so throwing another colorless thing like a chunk of glass on to a fire would put it out.
I might be quibbling about the particular example used here, though: if your point is that surface similarities are insufficient by themselves to generate rules of thumb, and intelligent entities should take that into account by performing causal reasoning to confirm the relevance of surface features, I can certainly buy that.
November 5, 2008 at 12:08 am |
Causal patterns are tiny events inside events of rules of thumb, sets of configurations small enough to determine outcomes from physical laws alone, without additionally appealing to actual content of the world. I clarified definition of causal patterns here.
I have a couple of posts later on what I consider a useful notion of causality. Here, I use the word to mean something like a process following physical laws that know no exceptions, the strongest variety of causal relation.
This post summarized what I consider the main reason surface similarity works at all, and why you can chain rules of thumb. Inside events tracked by rules of thumb are narrow threads of causal processes that produce all the surface properties, so by identifying a causal process from surface similarity you not only uncover other current properties, but also properties that hold in the past or in the future, where surface similarity might no longer hold. Reasoning works on heuristic rules of thumb, not on exact physical laws, but the reason heuristics work is that content of the world is arranged in a way that allows them to.
Magical thinking is a necessary evil in some cases, as we can’t see the causal structure of events. Not all reasoning by similarity works, but also not all rules of thumb that apply to unrelated causal patterns fail to work: it’s only a general heuristic that needs additional evidence in each particular case. Some rules of thumb reason from surface similarities but generalize across very different phenomena, and it might be a coincidence when different causal patterns follow the same rule or it might be a rule derived by chaining rules obtained from more basic principles. Some (most, really) rules of thumb are not at all about surface properties, but about highly abstract properties remote from what’s observable by many inferential steps.
The example is confusing, I didn’t like it from the start but preferred it to voodoo medicine for clarity of underlying physics. Chemistry is different, and physical processes are different, so if this kind of similarity knowably worked, it would be for another reason, because of abstract property of being a liquid perhaps, or conclusion from knowledge of chemistry. The main problem here is that coming up with rules of thumb good enough to even bother testing is hard, when search space is huge. You need some kind of evidence that pushes your chances above impossible, and, surprisingly, simple surface similarity is such evidence, all else being equal.