This is a follow-up to my previous post about the differences between the traditions of integrative, systemsy science. I will use the same –5 to +5 scale as in that post. Negative numbers are not bad but represent people, papers, places, and concepts more to the complex systems / Santa Fe Institute side. Positive numbers represent a stronger flavor of systems science, control theory, etc. (I avoid calling one side “complexity science” the other “systems science,” or similar, because there are plenty of works in spirit belonging to one side but using keywords of the other.) Words confuse the matter, but the attitudes to some key concepts clarify, and one of these concepts is “feedback.”
Feedbacks—loops of influence—is an important concept for all systems-related science, but the attitude towards it varies. At the + side, you don’t need an excuse for pointing them out. Maybe, you even need an explicit reason not to. At the – side, they are just there. “Do you ever see a system without feedbacks?” as some senior colleague put it back in the day.
In general, I think the two traditions need to learn from each other. Feedback effects are worth pointing out more than the – side does, but they do not necessarily warrant a mention by the sole virtue of their existence (as the + side seems to think).
It is understandable that the pluses, with their heavier influence from control theory and engineering, treat feedbacks more explicitly. Probably, at some point, every engineer—deafened by connecting a mic, amp, and speaker—vowed never to forget about them. But there is a deeper reason—humans have a bad habit of presenting knowledge as causal chains that can be long but rarely closed. As scientists, we need to avoid simplifying away feedbacks. Because you rarely can. Removing them would change the explanation (model) too much. I also think it is not too hard of a concept to use in public outreach, let alone academic information dissemination.
Maybe I’m exaggregating. I know several scientists effortlessly moving between the two poles without much of a hassle. There could be tensions*, but usually not caused by the different roles of feedback.
* I never had the chance of seeing Per Bak [–5] (mentioned in the last post), but rumor has it that he had vitriolic polemics with the + side (and probably the – side too). The closest experience I had was at the second network conference I attended (in Santa Fe in 2003). There, John C Doyle [+3] presented his HOT (highly optimized tolerance) model for generating power-law networks. I remember a slide with 10–20 desirable properties (latecomers can be successful, etc.) where HOT checked all but the Barabási–Albert model checked zero. In the following talk, Ricard Solé [–1] had a slide subtly spelling out HOT as “highly overstated theory” that made the audience chuckle in unison. I’m all in favor of Saganesque polemics, powered by deductive reasoning, but scientists are also humans, and we could need some jestful snarks to liven us up at times.