> The kind of absolute truth you are talking about may make sense when discussing philosophy or religion.

Exactly—except you are talking about it, too. When you say “discovering underlying principles”, you are implying the idea of absolute truth where there is none—the principles are not discovered, they are modeled, and that model is our fallible human construct. It’s a similar mistake as where you wrote “explain”: every model (there should always be more than one) provides a metaphor that 1) first and foremost, jives with our preexisting understanding of the world, and 2) offers a lossy map of some part of [directly inaccessible] reality from a particular angle—but not any sort of explanation with absolute truth in mind. Unless you treat scientific method as something akin to religion, which is a common fallacy and philosophical laziness, it does not possess any explanatory powers—and that is very much by design.

Now we come back to words gaining their meaning from usage.

You are assigning meanings to words like "discovering", "principles", and "explain" that other people don't share. Particularly people doing science. Because these absolute philosophical meanings are impossible in the real world, they are also useless when discussing the reality. Reserving common words for impossible concepts would not make sense. It would only hinder communication.

I can see what you mean. Then perhaps you could give a non-circular definition of what you mean by “underlying principles” and how that is different from any other prediction or model to deserve this distinct and quite strong-sounding term? or what you mean by “explain” that is different from “predict” or “model” to warrant such a distinctive term, and where exactly such explanatory activity fits within scientific method?

Communication is inherently circular, and words don't have definitions. But people are often capable of discovering what a particular word means in a particular context. And other people can sometimes help that by giving useful explanations.

Science is pretty much the same. We can often discover how the reality works, and use the discoveries to explain many things. Somehow that keeps happening all the time, even if we can never be fully sure about anything.

Any word can be given a definition, that’s how we communicate. A non-circular definition is a way to define what you mean by a term to another person.

Again: scientific method does not explain. Religion or philosophy are about explaining. Scientific method is about experimentation and making testable predictions. What experiments we perform is determined by how we understand the world, and if there is any subsequent explanation about “how things really are” (a.k.a. “the underlying principles”) then it has nothing to do with scientific method which does not make such claims by design; that is untestable/unfalsifiable beliefs and a product of either philosophical or religious thinking.

Since you insist on using specific words “explain” and “discover”, rather than more conventionally used in science “predict” or “model”, it implies they mean something different to you. I have provided the meanings of “explain” and “discover” I am familiar with, as it applies to the discussion at hand (which is about the philosophy of scientific process, underlying principles and truths about objectively existing reality). If you refuse to identify the meanings you are using those words in, I take it that you concede whatever point you had.

I've never met anyone capable of communicating with well-defined terms. Or giving definitions that actually match the real-word usage of the term. And all definitions are ultimately circular, because the number of words is finite. In any chain of definitions, you will eventually have to use a term you were trying to define.

What you call the scientific method is a philosophical construct that has little to do with actual science. And philosophers disagree on whether it's a good ideal for science. Given that it's neither a good description of science nor a universal ideal for science, I wouldn't focus too much on it when discussing science.

> And all definitions are ultimately circular, because the number of words is finite.

I can’t help thinking I’m talking to an LLM or a troll.

If you use a complex term that needs definition in a casual discussion, it’s most likely none of the words you use in the definition would themselves require definitions—and if this was to happen repeatedly, the conversation would halt long before we would be running out of words. It’s enough to avoid circularity within a couple of levels in good faith.

Anyway, I’m not sure whether we disagree or not or what exactly we are arguing about. My point is “ML making predictions is not a threat to us getting at underlying principles, because natural science (scientific method, predicting things) in general does lead us to any provable facts about those principles, and because ML would make predictions within an incorrect/incomplete model that we gave it.” In that, by “underlying principles” I mean some statements about objective reality. If we are on the same page here, we can continue discussion, otherwise let’s not.

It's an analogy. All communication is ultimately circular, and we can never be sure that we understand the terms the same way as the other party. Still, people often seem to be able to communicate.

Similarly, scientific method cannot discover the underlying principles or explain the nature. It can only rule out principles and explanations. Regardless, science seems to come up with principles and explanations all the time.

And that's because scientific method is not science. It's a theoretical model for a subset of (often ritualized) activities within science. Actual science is more than that. It can do things scientific method cannot, because it's less focused on philosophical ideas such as absolute truth or provable facts.

In my experience, scientific method is like a picture of an elephant drawn from a written description. Drawn by someone who has never seen the animal or a picture of it, and who has no idea what kind of an animal it is. There are some recognizable elements, but it definitely does not look like the real thing.

Sorry, what’s less focused, scientific method or “actual science”?