Science is a mystery - it happened by accident, unplanned, as a by-product of other cultural changes - and no sooner had people developed a recognition of what had happened, and devised some ideas obout how science worked - than it stopped working and transformed into just-another dishonest, self-serving futile bureaucracy.
But what was the basis of science?
The key was motivation, individual motivation - specifically the motivation to discover more of the truth of things in order to be able to do something.
Only when science is underpinned by the proper motivation of individual 'scientists' will it work.
(Obviously there will always be 'rotten apples' in science; but there cannot be too many nor at too high a proportion, else science will stop being science.)
There is only one proper motivation for science (i.e. truth-seeking), and an infinite number of wrong motivations, which point-off in every other direction than the proper one.
Specifically, each scientist will be motivated by a sub-goal (not to discover the truth about everything, of course; but a particular thing).
The specific motivation would be something like trying to discover something; to invent something like a theory or a device; to solve a problem - e.g. in measuring, predicting, performing some function; to do something like cure a disease, navigate, carry a load.
Something like that.
How much specialization should there be in science?
The answer is determined by motivation - by what each scientist is motivated to do.
There should be as much specialization as is judged to be helpful in doing that thing.
Social aspects of science also flow from motivation.
If the motivational goal can be achieved by a single person them obviously there is no reason for science to be a social activity; but a mutuality of interest may lead to cooperation, which would essentially be informal (based on mutual benefit). This will lead to the proper degree of the right kind of cooperation.
The rewards of science come initially from the intrinsic motivation and secondarily, when there are groups, from the status within the group and from the status of the group.
So that, although there is a pecking order within the group, which is a zero sum game; there is also a sense in which any group bound by mutual interest will award itself status, compared with those people who are not thus interested.
This means that all members of a group bound together by individual motivation may get psychological rewards from the group membership, based on them all making a contribution to that groups motivational goals - so even the lowliest member can feel a reward.
Science as a social process therefore happens when a task (coming from individual motivation) is too great for a single person to accomplish alone using available resources: for example when the making of a new kind of weapon requires the making of a new kind of material - the weapon maker and the material maker may cooperate to achieve this goal, which neither could do alone.
But if motivation is applied via a formal system, then scientists will (firstly) have their motivations re-shaped in order to satisfy system criteria; then (next) the actual selection of 'scientists' will cease to be on the basis of motivation and will instead be on the basis of system criteria - and it will stop being a science.
Because system demands are always partial and biased, and the fastest, easiest and most direct route to satisfying system demands will never be identical to personal motivations, and can be unconstrainedly different.
There is therefore no limit to how non-scientific real science can evolve-to-become, once personal motivation has been abandoned and replaced as the foundation.
Once motivation has been abandoned as the core of science and the principle behind its 'organization', the only thing standing in the path of complete corruption is inertia due to the overlap of generations.
So, all attempts to make science person-proof and independent of individual motivation will destroy real science - although typically the destruction will only become gradually obvious, as earlier generations of individually-motivated scientists lose influence, retire and die-off.
As usual, the benefits of impersonal systematization are immediate and short-term; while the harm is delayed and long-term.
Thus science becomes yet another victim of its own success.