tests, predictions, confirmations, hard to vary

There's the relative standing of knowledge between fields, which is
fine. Then there's explanations referencing themselves in terms of being
more hard-to-vary than they were before. Of significance here, the
overwhelming majority of justificationalism in Science corresponds to
doing the same thing, which draws into question the justificationalism
that is ascribed to it.

A non-trivial prediction (i.e. of something new, unanticipated, and
significant about the world) that gets confirmed adds new nowledge and
pushes the explanation up the hard-to-vary scale (because changes now
have to imply that new knowledge as well as all the other things.

Repeatability successively narrows the number of dependent variables,
and hones into their interactions and respective properties creating
upward pressure on various associated hard-to-vary statuses (say, what
the defines the experiment, the granularity of repeatability, precision
of results, etc). All which feedback to the predicting theory as more
upward pressure (because harder to vary prediction/experiment makes the
theory harder to vary for the same prediction.

When a theory leads to something new in technology some part of it fixes
into the laws of nature as snug and hard-to-vary as the last piece in a
jigsaw puzzle. The hard-to-varyness when ths happens transcends the
fragilities of human explanations, attaching instead to the sense and
ways Nature itself is fallible.

When new technologies enable fundamental shifts upward hard-to-vary
pressure may be felt throughout the knowledge creating organism. etc etc
etc