The Problem of the Many
plato.stanford.edu
plato.stanford.edu
If you're a pilot, a cloud is a way to talk about something that reduces your visibility and causes turbulence. If you aren't carrying an umbrella, a cloud may be a sign you're about to get wet. The fact that these phenomena share the name "cloud" is only meaningful to the extent that they arise from the same sort of physical processes. But trying to rigidly assign a particular arrangement of physical processes to a single entity "cloud" leads to the nonsense you see here.
Nobody using the word "cloud" cares about this assignment. They care about whether there'll be turbulence, or whether they're likely to get wet, or whether the cloud looks like a bunny rabbit or whatever. The solution to the paradox is to realize that not all concepts are analyzable to this degree, and that that's OK.
Practically, this is a problem that computers encounter whenever they are given an image and told to identify the boundaries of some object depicted in it. Instead of water droplets, we have pixels, and we have to figure out exactly which pixels are part of which objects (and to what degree, in case of partial transparency).
Having said that, I don't have high hopes that solving this philosophical "problem" will help us come up with better image processing algorithms. What we need is a scientific explanation of how human brains process visual stimuli, not a philosophical explanation of what we mean when we say there's a cloud.
But helping to solve them? Don't get your hopes up. We still haven't really solved any of the questions that Socrates came up with 2400 years ago.
-- Eliezer Yudkowsky
rigidly assign a particular arrangement of physical processes to a single entity "cloud" leads to the nonsense you see here.
I have difficulty understanding what you mean and therefore I will need to analyse your use of terminology if I am to respect your opinion. You make some implicit assumptions which are very challenging. E.g. you say
A cloud exists when someone can say "that's a cloud" and a listener thinks "yes, that is a cloud", or that fact is somehow meaningful to them in their life.
I am getting the sense of a consensus argument, but then you have to address when does the listener actually exists.
I am not saying these concepts are interesting to everyone, or that they definitely can be given a complete unquestionable answer, but to call them nonsense and turn your back is not needed when you can just turn your back.
I'm not sure I understand why over-analysis is needed. The "cloud" concept isn't entirely arbitrary. Most clouds (unless they're in a photograph or painting or being described abstractly using language) arise from the same physical processes and behave similarly (raining, darkening the sky and so on). This is how "cloud" gets its meaning in everyday language: there really is a shared phenomenon there.
The error happens when we try to understand the individual physical processes underlying a particular phenomenon as part of a single object. Even though the concept of separate objects is useful in everyday life, it no longer has meaning when you start trying to decide which cloud a given raindrop belongs to. Why not just abandon the now-useless concept of cloud and start talking about the raindrops as individual entities with their own behavior? Of course, we could look closely at the raindrops and see they are made of molecules: really the concept of "raindrop" is just another useful but ultimately imprecise bit of language. And so on...
For some time, for me, the grammatical structure of my sentences formed such a weak relationship to the interpretation of meaning of language, that a single word of a single sentence in a single paragraph could compose many meanings, and that doing this over and over throughout the paragraph could compose many more meanings, until all comprehension of what was intended to be expressed seemed to be completely lost. Communication had distinct dual meanings often, sometimes many, many more. The grammatical structure serves as an abstract form, then each word relationship is applied and toyed around with until associations that have nothing to do with the topic at hand are formed. I would be lost in attempting to speak back to someone. When I tried to speak, I would run off on a tangent that received blank stares at best. I was convinced people were purposefully messing with my head, but it was only the extraction of a single word from a single sentence that projected itself into my imagination and then distorted itself into a web of intricate knots that continued to build one after the other.
Some people call this telling stories. For me, it was the way I perceived my reality, even though my perception held in thought never matched my reality.
I have a habit of escaping into complex mathematics, so at least I have the illusion of intelligence (although it turns out, I am very good at complex mathematics, and this is useful). But it's very difficult, living like this. The single and the many is a real problem. You can think you know what you are talking about, but until you actually become convinced that every sentence can be interpreted completely differently by the listener, and by some form of magic I call compassion of others noticing how completely aloof I am, you manage to exist in society as member that actually contributes something. Also, when you intersect with people really frequently over short gaps that are spaced out by really long gaps of 'misunderstanding one another', then it's freaky and causality gets all tangled and you can sometimes get convinced that people can read your mind.
You may say cloud, but when I first read this article, I thought of the internet clouds, instead of water clouds. Now, this is only a small delta change between word choice. Imagine that a single sentence can be interpreted in millions of ways, and it can be continued on in conversation, in other sentences in millions of ways, and no one actually has a clue of what is being spoken about, but we all think we do. This is why I prefer to stare at my whiteboard with complex mathematics.
So much of how we talk to each other is about faith that we are understood. If you lose that faith, you start to question whether other people know who you are and what you mean. In reality, I don't think anyone can ever know, although some people may come close.
Formal language tries to eliminate ambiguity, because throughout history many people have realized what you have. There's nothing wrong with eliminating ambiguity, but it is important to recognize its purpose and place, and to get comfortable with informal language where it is needed. Artistry also lies in knowing what to leave undone, after all.
During a discussion I will often find myself go on multiple tangents to test the meaning of a phrase I am listening to, lopping with each new word coming in. And surprisingly to most I will sometimes arrive at multiple meanings I grasped and incidentally can't decide on which to pick!
While this is really embarrassing at times, I however consider this capacity of seeing a difference where most don't: a gift. The brain grows on this branching capability and I believe it is exactly this that enables me to understand a conversation faster than average.
So I have a somewhat autistic level of conversation but also have access to uncommon thought patterns just through casual conversation.
Thank you for sharing this.
That said, if there is clearly _one_ cloud, I think most annotators would agree that there is one cloud (and not, say, infinitely many).
So going from that, you can frame it as a constraint optimization problem. You want the largest possible collection of droplets to be a cloud, without accidentally defining all the clouds in the world into a single cloud. There has to be a loss function for the cloud-ness of a set of droplets based off how dispersed the droplets are in it.
Think about the fill bucket in Microsoft Paint. A single pixel hole allows the entire image to get painted one color. We don't want our definition of cloud to leak along the single droplets that exist in the air to define the entire atmosphere as a cloud, but we definitely want to group certain things together as clouds.
Hopefully that is food for thought for someone who is better versed at the specifics of anything I just said!
P.W. Anderson, "More is Different" (1972), http://robotics.cs.tamu.edu/dshell/cs689/papers/anderson72mo...
(One of my favourite essays on philosophy of science, by scientists. Perhaps just after Wigner's "The Unreasonable Effectiveness of Mathematics in the Natural Sciences".)
I really hate it when philosophers make "arguments by intuition". It's almost always used to justify denying some aspect of reality, or even worse, some ethical tradition.
Precisely. When you see what you're calling an argument by intuition, you can be sure that the writer will soon arrive at a comfortable proposition that makes you and him feel good and doesn't overturn any sacred cows.
Elsewhere in this thread a commenter claimed philosophy hadn't solved any of the problems posed by Socrates thousands of years ago. In fact, it has. The problem is many of these problems are subject to "proofs by intuition" which make us feel good and let us dismiss the counterevidence supplied by reason and thereby continue the "debate" in perpetuity.
Take for example, free will. The philosophical answer to the question of free will is that there is no question: the idea is confused nonsense. But (at least nowadays) we have a "feeling" of free will, and it upsets people to imagine they're not the captain of their fate, so the argument by intuition prevails. A similar objection goes along the lines of "but what would we do about criminals if they don't have free will! it would be unfair to punish them!" ergo free will exists because otherwise it would makes us feel bad.
I don't mean to pick on free will - the same goes for most of the other old questions. The argument by intuition as you say is worst in ethics - anything that doesn't "feel right" is wrong.
Is there a problem? Seems to me if there were an actual and important problem whose solution were dependent on "what a cloud is" then you would have no shortage of conceptualizations. What then matters is to what extent they are useful for resolving problems.
I believe this is quite normal and our conceptualizations of ideas change as the problems we face change. For instance, you can always ask what does Justice, the Idea, mean. People have developed concepts over time and applied them. The success leads to further problems, asking again what Justice means, and further concepts.