I would say that social phenomena share with ecology an evolutionary character, but the reflexivity Soros talks about is present only in social events and so while ecology might be subject to the standard scientific method, sociology is not (and neither is economics). [emphasis added]

Quite. Soros is apparently trying to think himself out of a box of Cartesian proportions and logical operators. "Reflexivity" cannot be the word he's searching for to describe a concept or a intellectual or material world history or even veridical attributes (T, F) of an ideology (e.g. economic literature, scientific method, physical "law") expressed by human events.

We are still waiting, as Jung pointed out, for experimental science to reproduce a thought. Muchless falsify a thought.

Recursion is the word and the procedure that produces, or yields, "truth" (e.g. agreement). By analogy to physical canon, "truth" is a homeostatic observation which tends to understate T,F states of the constituents underlying appearances.

Computer science is particularly helpful in demonstating how (syntax, terms, variables, coefficients, operators) sythesis occurs. Synthesis is a word meaning evolution insofar as the the product of the expression may also be one or more variables and one or more coefficients generated by the procedure. Who knows? All we study are conditional statements --the politics, the law-- which proscribe the limit(s) of the process.

T, F facts denoting any one expressed human behavior, individual or aggregate, are untenable though. Philosophically speaking, that collection of unknown (invisible but not undiscoverable) data is known as "reality" ("uncertainty" or "risk").

Diversity is the key to economic and political evolution.