Reasoning from first principles is about differentiating complex systems, breaking them down into component parts so that we can understand how they work.
When trying to understand what a system does, there are two main steps.
The first is understanding the component parts of the system in relative isolation. This can be thought of as the component (or "architectural") view. A general term for the component parts in this view is [idea].
The second is understanding how these component parts connect to each other and affect each other when the system is actually working. This can be thought of as the signal view. A general term for any two connected parts in this view is the [relationship].
Where the component [idea] view helps us to see and understand the potential of the component parts, the signal [relationship] view is where we understand how the system works in a specific configuration.
When working from first principles, what matters is how we think. You could think of this as happening in "imaginary space". Because of this, both the ideas and the relationships that we come up with are "imaginary." They are imaginary representations. As a result, we can choose how we define them.
The things that we define ideas and relationships with are [limits].
Because ideas, relationships and limits are all imaginary constructs, working from first principles can be a trial and error process till we find the ideas and relationships that give us the best results based on what we are trying to do.
The fact that working from first principles involves so much trial and error is actually also its strength. When we redefine ideas and relationships we create a change. This change is geared towards understanding what we are working with.
With practice working with change (by creating it ourselves) we can get better at dealing with change in general.