Logic generally is hard and probabilistic logic (with model counting, like ProbLog2) is even harder. But commonsense knowledge and reasoning is quite lightweight. There have been efforts to represent commonsense knowledge like https://www.amazon.co.uk/Representations-Commonsense-Knowledge-Ernest-Davis/dp/1483207706/ but this has been done with the help of the logic and all this leads to inefficiencies that human beings can not afford. There is common pattern that commonsense reasoning and intuition guides even the scientific thought (and this is light step with lot of exploration and less efforts) and only later the ideas are put into more or less logical form (resource intensive step). There is also "Journal of Informal Logic" https://informallogic.ca/index.php/informal_logic but without systems of symbolic execution.
So - are there lighweight symbolic approaches for commonsense reasoning that are lighter than common logical methods but more rational than black box neural networks? Something like commonsense Prolog, commonsens theorem prover.
Maybe there is answer that considers this from the other viewpoint - logic is the same but it is used in the forward chaining mode which is so much cheaper and far more explorative than the backward chaining mode.