What are incidental effects? They are regularities. Regularities are features shared by interventions. This is not to say that these features are hidden in the world just waiting for us to discover them. Nor is there a talent in our nature that allow us to pick out these regularities for us to exploit the way Hume describes it in Of the connection or association of ideas in his A Treatise of Human Nature.
Hume thought that the mind organizes ideas in a taxonomy. The particular organization of the taxonomy relies on how ideas are brought to mind through different types of relation, the most important being resemblance. The theory assumes that ideas go from particular to general through a sequence of degrees of relationships. Because the reality of an object, for Hume, is not separate from the apprehension of quality or quantity up to a particular degree, it follows that forming the idea of an object is the same as forming an idea. The way by which ideas become general is by accumulating all degrees of quality and quantity associated with an idea. This is not a perfect ability but it is the one suited for the purpose of apprehending the world. It is the task of the imagination to sort through the taxonomy of ideas and extract the contextual ones.
This view is still widely shared. There is an appearance of exploitation of regularities on the part of our species; however, our approach does not require the world to exhibit regularities. Registering regularities is a feature of our biology. Registering, is being conditioned to expect certain effects given stimuli of some kind, not unlike a Pavlovian [45] reflex. Simply, it may be that our capacity to discriminate is limited because of architectural constraints. Discriminating capacity does not come cheap in many ways including processing effort and time. A lengthy discriminating process may reduce the survival rate of our species. Having a notion of potential predator based on a limited discrimination of resemblance is a useful thing to perceive in a split second. An infinite capacity to discriminate would hinder the process of registering regularities hence making very difficult categorizing based on similarities. How could we have syntactic rules if features of utterances would be overwhelmingly different as opposed to having anything in common. We couldn't refer to noun phrases or verbs because object or state would never become apparent as a shared feature. In the case of language, earlier stages of exploitation are of non-linguistic kinds but later stages of exploitations involve mostly linguistic regularities. There are many evolutionary reasons why linguistic regularities would be exploited over non-linguistic ones. A striking example is to be found in people's ability to coordinate each other or large groups into performing complex tasks. These physical effects would not occur if not for the event of language.