Good post by on why starting a new field is hard. Direct economic incentives (bigger grants) are too expensive. Selling researchers on importance is tough absent an acute crisis like Covid. Cc
Conversation
Misses the biggest factor though imo — new paradigms. Article hints at that in passing — more pivoty seminal publications get more citations — but doesn’t unpack it. You can’t just “start a new field” because a topic is important. You have to start it *with* something.
1
13
A new paradigm promises to produce tons of new results fast, at low cost. New fields take off when researchers sense an explosion is imminent and they can ride it to big results. I’d argue the best researchers index on interestingness not importance. ssnationalstudies.files.wordpress.com/2009/11/thatsi
3
1
15
I’d argue that the burst of activity triggered by Covid was at least partly due to a “paradigm effect” rather than importance/social good. Many recognized that a huge natural experiment was underway and would paradigmatically reframe everything we thought we knew in many fields.
Replying to
I’m not actually a big fan of Kuhnian paradigm shift model, but I think it applies here to first order (I’m more in the Feyerabend methodological anarchy camp).
Basically good researchers don’t work on what’s “important” but on what they have a promising attack for.
4
11
Unless you define “important” the way Hamming did. The problem with all programmatic top-down efforts to “start a new field” is that they tend to go after important in the wrong outsider sense flagged by Hamming here. cs.virginia.edu/~robins/YouAnd
1
2
12
It’s not actually possible to drive planned roadmaps at any useful level of detail below say “go to the moon” and allocating Apollo level budgets. Below that, you gotta pour fuel on emerging paradigm sparks, not force “research time travel” down people’s throats with money/funds.
1
1
14
This is one reason I’m a skeptic of Progress Studies type thinking. Appeals to economists and billionaires but rests on an illusion of methodological legibility and governability that the history of science casts serious doubt on.
2
17
A “tell” that you’re trying to force a new field is uninspired, extrapolative naming as in “5th generation computing” which couldn’t avert the AI winter despite billions in funding by Japan in the 80s vs. “deep learning” as a natural paradigm that attracted funding.
2
8
You can’t “make fetch happen”
One of the warning signs of Web3 is that it is called Web3. I do think there’s a new paradigm there but the bane suggests a weakness that’s being patched over by funding forces.
1
24
But I’m not overindexing in this name heuristic. Web 2.0 despite being named by Tim O’Reilly by extrapolation was a new paradigm. Semantic Web despite sounding like a paradigm, wasn’t really. It was wishful thinking.
1
14
Also values/manifestos are not paradigms.
Sustainability is not a paradigm
Progress is not a paradigm
Equality/justice are not paradigms
Maintenance over building is not a paradigm
Paradigms don’t prioritize problems, they suggest ways of looking at phenomena
1
1
12
The basic tension in a formulation like “progress studies” is that it applies a lens from “low paradigm” fields (social sciences) to “high paradigm” processes (technological evolution). It’s like trying to study chemistry by composing music about it.
1
6
Replying to
This is an important point and I am seeing variations of this insight all over the place.
2

