I want to make it clear that I'm trying to highlight differences, not objectively claim one is better than the other. What I'd like to encourage is understanding from both sides, but since I am on one side, I naturally understand it better.
Yes, I can see that some people do that, though if done properly, you're not specifically looking for evidence just to support it, you're looking for any evidence at all that's connected with it, in order to evaluate it in light of all pros and cons. Though in fact, in my case, usually I don't even have a theory to begin with. I'm not even sure if what I have is a problem. That's the thing... I don't like to even try to define something until I've enough data on it to see where it fits into the bigger picture.
Sure - I seperated the part you quoted from the rest of my post because that's how I feel (so I kinda answered the question objectively first, then subjectively).
But this description is the problem. If you don't know what you are looking for, how can you know when you have found it? Is it just when you find something? That doesn't sit well with me

I need to know what you are looking for, so that I can then answer it properly - that is, give you data on the topic you want to solve.
(Illustration: You recently asked about bullying and different personality types in another thread. There is huge amounts of research on bullying now, including a lot on personality influences (though, not much with MBTI, but there never is

) Did you want to know? Or is it just for discussion? Which would feed the theory, which would create more theories to work on? Which is your preference? I could answer your questions on personality and theories in a very dry way, and even explain what the sub-scales show so that you don't have to theorize much. You wouldn't like that, would you...? Yet, that will provide a great deal of data around the problem, but it will be structured and not leave much room for you to theorize inside.)
I do understand the need to understand something, the drive to gather data, to model it. I'm not
that different. But we do have a strong divergence when it comes to defining something. To me, the theory serves as a tool to solve a problem. As above, with the bullying, I know the answer to a high enough degree of confidence that I don't need to theorize. Do you want to reach that point, really? I never feel that NPs do want to, and avoid it at all costs!)
That's the gap we face when we talk. You start theorizing, I see you struggling for an answer, and so I give you an answer. At the same time, I can provide you an answer, and then you'll ask me what it means, and I'll answer "exactly what it says", and I'll get frustrated with your need to find meaning - you'll be thinking about how closed minded I am. Both of our feelings end up being correct - they don't come from being right, but not being able to close the gap. One of us can't be ourselves in that situation.
Opposing functions (and we) will always be in conflict, in that regard. It probably becomes worse coming from me, who uses Ti like a sledgehammer for any problem (way more than the INTPs

). I will have structure, even if it means I'm going to try building a giantic rib frame for a personal submarine on a whim.

My methodology is very well honed, and it doesn't agree with your well honed method. Doesn't mean that we aren't useful to each other, but we might not be useful when we work on the same thing, in the same room!
That's the thing... many abstract theories ARE useful. You can start out with a "what if?" and then look for ways to find out whether your "what if?" is possible, desired, necessary, etc. all of which would never have happened if you'd insisted that no theorizing was possible without FIRST knowing all those facts.
Do you consider a theory abstract after it has been validated? I don't.
"What ifs" are not my strong point, naturally. But this is intentional for me (well, 'I don't regret it intentional'). If you come up with 1000 theories, but only 100 are validated, is it better or worse than me coming up with 100 theories, but 90 being validated.
What if you assign a low confidence to the validation? (Meaning, you don't really know which of the 100 are actually good?) However, I know that NPs tend to be happy with this, because they don't really think that much of their own theories. I can't afford that luxury - no practical application means the theory doesn't exist (and then, if I need a theory to apply, it has to work, so I need very high confidence.)
Again, just to avoid making it sound like I'm attacking Ns in general, this is my perspective, and my frustration. Yes, abstractions are useful, but they can't be useful for *me*, now, in my project or life. I can't take an N theory and use it because I can't *trust it*. I need to validate it, but when I validate it, I'm removing the whole N-centric approach.
Both Ns and Ss reach this point, just in a different order.
I would say that N theories are useful because of their farther reach, but my general feeling is to never trust an obvious N theory - that is, if I can tell where it came from, it probably isn't refined enough. I'm looking for middle ground between abstract and validated.
(The other side of this is that many "S" theories are so obvious that there is no point calling them theories. It is all a matter of scale. Too big to be validated isn't useful, too small isn't useful... but inbetween? Perfect.)
It wouldn't be necessarily rude to ask that question, but y'know... perhaps you could allow for the idea that the 'throwing around' actually *is* the process of putting thought in? Especially with NP's, it's done in the open air. What you're hearing is the PROCESS, not the result. I don't personally get why you would see it as such a bad thing that somebody just thinks aloud... do we all have to have our thoughts sit an entrance exam before they're allowed expression??
Putting 'thought' in doesn't impress me as a methodology, and that's the issue. I'd rather be out gathering the validation long before you guys have discussing something. It would be like facing an entrance exam and instead of studying, you guys think about the exam a lot.
I answer "what do I need to know to understand this" first, then go hunting for it. When you guys don't do that, it feels like you are just spinning wheels pointlessly. That's where the
Smallest link between two points comes in. I want to make as few intuitive leaps as I can, which means more data to fill in the gaps. To hear that data is often
not welcome in these situations... that's a very heavy statement to put in. It isn't always true - depends on the balance on the individual/the theory being worked out. But to be biased against
data during the forming process isn't a high compliment to me.
But we are different people, with different methods. I respect the outcome of Ns in the big picture, but dealing with them locally requires understanding, and tolerance. Better that they vet their own ideas first, then bring them to me to review... I can't handle being part of the process as well!
Hence, fewer, long posts, on topics that I can contribute and not get rejected
---
Also note that there are fallacies to the preferences as well - Ss might invalidate theories that they can't see in their own life (often falling prey to statistical issues, like clumping, etc). I'm not pointing out these because they are intrinsic to the method we use anyway.