Research

Purpose of Thinking; Positive and Negative Arguments; Clear Goals

Elliot Temple

Feb 5, 2023 • 9 min read

Summary: Thinking is complicated but worth working on and improving. Positive arguments aren’t as good as negative arguments, because we care whether an idea is broken (and will fail) or not. One negative argument can imply an idea is broken; a dozen positive arguments cannot rule out the idea being broken. In order to improve at decisive criticism, we need to define goals more clearly. To remain flexible, we need to consider multiple goals instead of just one. It’s easy to give decisive, pass/fail judgments to candidate houses in terms of a clear goal like “at least 3 bedrooms, a pool, and costs under $300,000 (plus works for background context goals like having a roof)”.

What’s the purpose of thinking? Thinking helps us understand the world and guide our actions. Random actions wouldn’t even succeed at feeding myself, while intelligent actions are powerful enough to land a man on the moon. And if I didn’t think, my understanding of the world would be sort of like an infant’s.

Thinking lets me achieve goals like walking across a room or acquiring and driving a car without crashing. Thinking also helps me decide which goals to pursue.

The world is very complicated. There’s too much to think about. We can’t pay close attention to everything. We can’t even pay a little bit of attention to absolutely everything. How can we narrow our attention without missing something important? How can we exclude a bunch of junk from our thoughts without excluding great ideas? How can we know what is junk without thinking about it first? But if we think about it first then we may spend too much time thinking about junk. We must focus on what’s important but it’s easy to screw that up.

And people have many other good questions about thinking: How can I think better? How do I decide between several contradictory ideas? How do I know which ideas are better than others? How do I know what’s true or false? How do I come up with good ideas? How can I be a fast, effective learner? How can I be creative and come up with ideas that other people don’t? How can I be a good problem solver? How can I be smart or wise? How can I improve my memory? How can I think faster without losing quality? How can I teach ideas to others or make persuasive arguments? How can I think objectively instead of being biased? How can people rationally resolve disagreements? How can people avoid making mistakes?

There are so many issues that it’s hard to think about them all. Let’s try to organize this by starting simple and building on it.

There is a physical world. It exists. It’s an objective reality independent of our minds. All people share the same reality which has facts and objective truths about e.g. what physical objects exist at what locations and what laws of physics govern them. We can know things about the world by using our senses to observe and our minds to think. But we can make mistakes. We can have knowledge but we can’t have guarantees of correctness. We’re fallible. We can keep improving our knowledge indefinitely instead of reaching perfection and being done thinking. Any of our beliefs could potentially be worth revisiting and revising if we got new information. Some people disagree with these things, but they’re pretty common ideas, and they’re explained and argued a bunch elsewhere. I have reasons to believe them. Let’s assume them here and focus on other stuff.

We learn by guesses and criticism. We come up with ideas, e.g. by brainstorming, and we look for errors. This is literally an evolutionary process. These claims are pretty controversial. I won’t go into explaining evolution here. But why look for errors instead of for merits? Why look for bad things instead of good things? Why focus on the negative instead of the positive? Or why not both?

You can look for both, but finding good things or bad things about an idea are different processes. It’s asymmetric. One isn’t just the mirror of the other. They work differently. There’s a reason to emphasize negative.

No matter how many good traits of an idea we correctly find, they can never tell us there are no errors. If an idea is correct in a million ways, it’s still probably not perfect. Perfect basically means infinitely correct. We can’t establish our ideas are perfect. If an idea has lots of good attributes, it could still have a major, relevant flaw. That’s actually pretty common. An idea can be great in 50 ways, yet there’s one big catch…

However, we can establish ideas are flawed. That’s because even one notable flaw is a big deal. The standard isn’t infinitely flawed. There’s no need to show an idea is wrong in every way. We don’t run into the same problem when criticizing.

If we have five positive arguments in favor of an idea, and they all deal with important issues (not minor details), and they’re all true, it could easily be the case that we still shouldn’t act on or believe the idea.

If we have five negative arguments against an idea, and they all deal with important issues (not minor details), and they’re all true, it could not easily be the case that we should act on or believe the idea anyway. That’d take something, unusual, e.g. desperation: every alternative idea we have is even worse.

We want ideas that aren’t broken. We want ideas that actually work mostly as intended and don’t have major negative side effects. If we didn’t do critical thinking, things would go horribly wrong pretty often.

Positive arguments can technically all be converted to negative arguments (if they make sense). So we only need critical thinking. However, using positive arguments is fine. People find it convenient and natural because they’re used to it. Just be careful about selective attention. Often an idea is good at something but the alternatives are too. Positive traits are often shared by many or all of the ideas being considered, so finding that a specific idea has a positive trait doesn’t mean it’s better than the other ideas. On the other hand, important negative traits often aren’t shared by the other ideas. Ideas we’re considering often aren’t all broken in the same way. (Sometimes they are and we need to think of some new ideas.)

Both positive and negative arguments must be considered in context: What is the goal? What ideas have we thought of to achieve this goal? We want to avoid ideas that will fail at the goal (this requires critical thinking) and compare the remaining ideas in some way to figure out which one is best (if optimization is worth the effort – for less important goals we’ll often just go with the first idea we think will work).

There is a fairly standard view along these lines. Critical thinking eliminates ideas that are false/wrong/refuted (that’s phase one), and then (for phase two) we use both positive and negative arguments to determine which of the remaining ideas are better or worse. This gives negative arguments a bigger role than positive arguments, since they’re exclusively used in the first phase.

I have an original proposal, which is to use only the first phase and avoid the second phase of comparing ideas by degrees of how good they are. With better decisive criticism, we can narrow things down enough. And if we don’t, and it’s worth the effort to optimize, then instead of using phase two we should return to phase zero, which is goal selection. Also the critical techniques are powerful enough that we may refute all the ideas. That’s pretty common, contrary to the expectation that only a small fraction of ideas can be decisively refuted. In that case, we have to brainstorm more, and if we get stuck there’s a specific, repeatable technique to get unstuck by making things easier.

Let’s clarify the process.

First we need a goal. Then we brainstorm ideas to achieve it. Then we think critically about whether the ideas will succeed at the goal. We do this in context. We have a situation, including our life in general, and our solution/idea/plan needs to work in that context. E.g. if our goal is about winning a chess game, we still have the background context goal of not dying. We want a way of playing chess that doesn’t get us killed. That may sound trivially easy. And you’re right. Most background goals are easy. If they were hard for us, they’d get more active attention. But sometimes background context makes a difference and arguments are allowed to bring it up. We do have other goals we want to simultaneously succeed at, like safety.

So, goal, brainstorm, criticism in context. And people’s main concern is there will be a bunch of ideas that survive criticism and we (often, when it’s worth the effort) need to rank them by how good they are or otherwise compare them. There’s also the potential case where we don’t have any good ideas.

So I think people need to get better at criticism so they can refute way more ideas. And they should do it with decisive criticisms – reasons an idea fails at its goal (or a background context goal), rather than non-decisive reasons the idea is less than perfect.

How? People already try to be good at critical thinking? Do I have a major breakthrough on that? Not directly. I think people need clearer goals and then criticism will be way easier.

If goals unambiguously specify what counts as success and failure, then they maximally enable criticism. A criticism just has to say what outcome an idea/plan/solution will get and then show it’s in the failure category not the success category. And that’s a decisive criticism, not a degree-of-goodness criticism.

People already know clarity is important. Are their goals pretty clear already? No! People work on goals like “get a good pet” or “find and buy a good house”.

If you defined “good house” to mean 3 bedrooms and a pool for under $300,000, and nothing else is relevant, then it’d be really easy to judge and refute houses. That one has two bedrooms; it’s refuted. That one has no pool; it’s refuted. That one costs over $300,000 ; it’s refuted. A super clear goal like this makes criticism easy! This sort of goal enables giving ideas/plans/solutions a pass or fail grade and makes judging them way easier. A fail grade is a decisive criticism.

So why don’t people use clear goals like this? They have flexible goals. They might be willing to pay a bit more or give up the pool. And their goals are complicated: they actually want more from the house besides those three criteria, e.g. it should have a roof. Having a roof is a background context goal which we may not need to state and focus attention on, but we’d know it really is a goal if asked (“Do you want a roof, too?”). If we saw a house with no roof, we’d recognize the problem.

So instead of having one vague goal, consider several clear and specific goals, and check which ideas succeed at which goals. The way to think is to evaluate (idea, goal, context) triples. Does that idea work at that goal in that context?

When we act, we should act on an (idea, goal, context) triple with a pass evaluation not a fail evaluation. This is a fairly unambitious claim. Lots of epistemologies get into a lot of trouble trying to say things that are harder to defend than that. They make more complex claims and run into some hard problems. But the idea of not acting on an idea, for a specific goal/purpose, that you think won’t work (at the very thing it’s supposed to work for) makes total sense.

So instead of evaluating how good an idea is at a vague goal, we should evaluate whether it passes or fails at several specific goals. Making the goals specific enables decisive, pass/fail criticism and evaluation. Evaluating for multiple goals gives flexibility. Instead of refusing to define our goal clearly to keep flexibility, we should define many goals that we might be happy with and investigate them. It’s like brainstorming for goals rather than having all that complexity in one big, unclear goal. Imagine if people did that with ideas/solutions; they wouldn’t make any specific suggestions because they wanted to be flexible and they thought they could only have one idea instead of being allowed to brainstorm a list of ideas, so they had to keep their one idea vague enough that they could change it into any idea they might want.

This leaves open the issue of choosing which goal to pursue. Some goals are better or more ambitious than others in some sense. Are we just moving the problem of evaluating goodness to the goal? E.g. Getting a house for under 120k is a pretty good goal, and getting a house for under 100k is a better and more preferable goal.

Sign up for more like this.