There is boring science and there is interesting science.

Quantum computation, for instance, is an interesting science, primarily because it's new: when in 100 years its foundation has been cured and dried, we will look back and identify this period of time, plus or minus 30 years, as the period when the main results came out, when the trailblazers marched on. It will eventually become a boring science, though, unlike those sciences which were forced either by a loose coupling with reality or a deeply philosophical/abstract bent to stay interesting (alchemy is in the former category and evolutionary biology is arguably in the latter).

Note: computational neuroscience is in a sort of superposition of boring and interesting. There's no fundamental theory, and so work in computational neuroscience is interesting insofar as it must propose such a theory to really get anywhere; the majority of such proposed theories are dull and useless, though. Boring interesting science, as opposed to interesting boring science.

The Creation of Interesting Science

One of the most interesting properties of interesting sciences is their volatility: they are prone to major changes effected by small groups of people making hitherto inconceivable paradigm shifts. To make this more concrete, we must first analyze interesting science. By its inherently revolutionary nature, I don't think there can be one overarching framework for the production of interesting science, but I do think we can identify key aspects and preconditions.

There's a certain mode of philosophical thought inherent to interesting science that can be seen in Einstein, in Dirac and Feynman, in Turing and Grothendieck and Noether and Newton. I have trouble articulating it, but it is more or less the 'dissolution of ontology'.

One casts aside the frameworks for thinking about phenomena which have been given to them, and articulates a perpendicular set of frameworks which are capable of recasting the same phenomena in a new light.

There are a billion ways to do this, one can for instance look at the myriad generalizations of any given useful mathematical object, but most of them are, while valid, ultimately useless. There is an additional precondition, which I'd call a sort of "goal-orientedness".

One identifies a problem with the previous framework, a certain phenomenon which it is unable to properly articulate, a certain question which it not only cannot answer but cannot even ask, and uses philosophical faculties to identify how the framework fails to do so. The solution to the failure identified will serve as the seed for the new ontology. This process is echoed by the alchemical process for arriving at the Great Work, the transmutation of lead into gold: the decomposition, the purification, the transmutation, and finally the unification.

I will call the stages of decomposition and purification here — the stages in which the ontology is philosophically dissolved — solvation, and the stage of unification — that in which observation points the way to a new philosophical paradigm which is drawn out from the previous solution — rationation. I've taken these from the Latin words solvendīs and ratiōnis, which seem to more accurately describe the intuition I have in mind for these processes than do any currently existing English word. The process of creating interesting science, then, begins with solvation and ends with rationation. Behind sciendī, solvendīs et ratiōnis. (I do not understand Latin conjugation).

We have left out the intermediary step, though: transmutation. The philosophical solution remaining after the solvation of a paradigm does not in itself contain the basis of the new paradigm to be rationated, nor can observation cleanly identify those parts of it that will serve as the base for the new paradigm: to determine the form of the new paradigm that will emerge from the old is a task left to human understanding, is the task after solvation and before rationation, is transmutation. The transmutation is the key part of the paradigm shift. It is the notion of relativity as such, of quantization as such, and so on.

So, the next stage of analysis: how do we go about solvation? rationation? transmutation?

Two Aspects of Knowing

Science is composed of a spontaneous aspect analogous to the understanding, and a receptive aspect analogous to sensibility. The spontaneous aspect is the alchemical Great Work, the solvation-transmutation-rationation complex, and is in charge of providing science with its intellectual, sense-making aspect, while the receptive aspect provides science with a body of data which is made legible by the spontaneous aspect; receptivity is book-keeping, data-gathering, tying up loose plot lines.

These two aspects require one another: without receptivity, spontaneity does not validate its concepts, and devolves into mysticism. Without spontaneity, receptivity does not organize the data it gathers, and devolves into a mere collection of observations. It is when they are fused together into science that they are capable of studying the world. The means by which the spontaneous and receptive faculties of science are fused is precisely the scientific method: hypotheses are proposed according to some variables within some paradigm, empirically tested, and the variables tuned contingent upon the results. The spontaneous faculty determines the key variables of the paradigm as well as the paradigm itself, whereas the receptive faculty determines how to operationalize and test hypotheses; they work together on determining what hypotheses are to be drawn from the paradigm.

So, to ground the earlier division, "boring science" is science with far more receptivity than is demanded by its level of spontaneity: there is data collection, but there are no essentially new insights. "Interesting science" is that which has a spontaneous faculty in proportion with its representative faculty.

I'd argue that it is extremely rare to see spontaneity outweigh receptivity, since receptivity is much easier and safer for any given "scientist" to pursue, but this is a function of the economic situation of academia as much as human psychology in general. However, it does happen, and this often leads to a sort of "mystical" quality surrounding the field, see e.g. the free energy principle, M-theory, and higher topos theory.

Note that in each of these examples we have a small cabal of trailblazers who have previously been recognized for their genius and therefore have the economic security to pursue these trails which, while ideologically new and refreshing, have little evidence supporting them so far.