Systems are designed to have certain modes of interaction. The assumption by the user is that the observable controls will have the intended result. The reason this is possible is the prior analysis of the entire system and removal of variation of performance from actions which are not intended to have an effect. Sometimes it's necessary to question to what extent this planning was done effectively, but in many cases without prior experience with similar situations the available modes of interaction must be taken as presented.
In a 'living' environment characterized by energy input and systematic replication, there are a greater number of ways to produce results and more ways to make mistakes. It helps to have experience with both designed systems, and natural environments.