So, I was reading Hollingsworth et al. 'Transforming socio-economics with a new epistemology'  and fully understood one of the things that has been bothering me lately. It seems that a lot of people have difficulty grasping feedback loops. Reading this article made me realize why.
It basically states that the original scientific paradigm originates from Newton (Science I) and is based on causality and reductionism. In the past 150 years a new paradigm of science has been emerging, one based on evolution of systems, complex and adaptive systems and so on.
He does not say it out aloud, but one key difference in thinking is that Science I has definitely a linear causal models as its modus operandi. Science II, on the other hand, is based on causal feedback loops.
Consider a falling apple. The apple is falling because of gravity (quantum-gravity, or einsteinian curvature of space time, if you insist), but the gravity is not happening because the apple is falling. While it is true that the gravity fields are changing when the apple moves, the effects of that are small compared to systems with true, and potentially complex feedback loops.
The counter example would be evolution of a predator and prey. The predator-prey and predator-predator dynamics most likely brought about groups. Groups brought evolutionary pressures for memory and understanding of social dynamics. A number of other things happened with the result of sky rocketing intellect in this one group of primates. We're still running on this curve; it may even be that one result of this curve is artificial general intelligence.
Look at the environment and society around you. Do you see linear causalities, in the sense of A causes B? Or do you see more or less complex feedback loops, in the sense of change in A causes a change in B, which causes a change in A. Apart from autofeedback loop, this is the most simple feedback loop there is. There are many, many more, all with different properties and with potentially hugely complex and dynamic structure.
What does this all mean for science?
One way to look at this is to say that in earlier times, the primary need was for us to understand the main (linear) causalities that exist. This work is the necessary work of understanding the nuts and bolts of the system, understanding each causal arrow in the whole. But we do understand a lot about those, in some cases enough that it is much more important to start looking at those feedback systems and the kinds of effects they have.
 Hollingsworth and Müller. Transforming socio-economics with a new epistemology. Socio-Economic Review (2008) vol. 6 pp. 395–426