Skip to content

Early and late adopters in research

by Michael Nielsen on March 31, 2005

In this post I describe some tentative ideas about two different styles of doing research that I call “early adoption” and “late adoption”. These styles are, of course, caricatures, but I think they’re interesting and enlightening enough to be worth some thought and discussion.

Early adopters are the people who get into new research fields very early on, before those fields are well known or established. They adopt or create a new narrative, one that is initially acknowledged by only a few others.

It’s not difficult to come up with examples of early adopters – consider Einstein on quantum mechanics, Minkowski on special relativity, or Freeman Dyson on quantum electrodynamics.

Such examples are perhaps somewhat misleading, since they create the impression that early adopters are uniformly successful. Of course, almost by definition the great majority of early adopters work in fields that fail and die quickly, and so remain anonymous.

Other more recent examples are the people who worked on quantum information prior to the explosion of interest in 1994, or the people who worked on social networks prior to the big boom in interest starting in the late 1990s.

Other people – the late adopters – get into research fields when those fields have become established, with an agreed upon basic narrative and set of fundamental problems. It’s a lot easier to find examples of late adoption: if you work in a University, then the chances are that ninety plus percent of your colleagues are late adopters, working in well-established research fields.

I’ve been using the loose term “narrative”, without explaining precisely what I mean. I’m actually bundling up several different things into the term:

  • The set of big picture problems that motivate a field, like “how did the Universe begin”, or “how do cells turn into bodies?”
  • The norms that define what it means to have a result in a field, i.e., what it means to make progress, including standards of evidence and of argument. Some consideration shows that even apparently very closely related fields (e.g. quantum optics and quantum information) can have quite different norms.
  • The justification for considering the field important – how it relates to the rest of science, what it can contribute to other fields, how it relates to society as a whole, and so on.

In short, the narrative is the story practitioners tell themselves and others about a field. It’s the first chapter or two of a good PhD thesis. It’s the paragraph at the beginning of the paper that goes “We know that field blah is important because of blah and blah. A major problem in the field is blah, which is important because of blah. We’re going to look at little problem blah, which will enable progress on our major problem.”

In established fields, the narrative is well-known and largely agreed upon by people within the field, although not necessarily without. Most (though not all) papers make only very limited changes to the narrative. By contrast, in new fields, the narrative itself is up for grabs, and early papers are often not only concerned with solving a specific problem, but also (often as a subtext) with constructing the field’s narrative.


A few caveats about the dichotomy between early and late adopters are in order.

First, early adoption does not mean bringing some new set of technical tools into a field to help solve some class of problems, except insofar as those tools help transform the narrative of the field, i.e., modify in some essential way the set of basic questions motivating the field, or the context in which the field is understood, or the norms defining what it means to “make progress” in the field.

A second caveat is that, of course, one person can be both an early and a late adopter. Feynman is a good example, working in established fields such as low-temperature and high-energy physics, and in new-born fields such as the physics of information and nanotechnology. So, while I’ve referred to people as being early or late adopters, these terms really refer to roles that people play, and the same person may play multiple roles.

A third caveat is that to some extent all researchers are early adopters – to write new papers, we need to accept and work on novel problems. But most papers work on problems which are technical variations on previously solved problems, or involve only an infinitesimal change to the narrative, and this is not really what I’m talking about as early adoption.

(Incidentally, I expect that the difference between people who are habitually early or late adopters shows up particularly starkly in the refereeing process, with late adopters being much more resistant to papers that change the narrative of a field, or suggest a new narrative.)

Advantages and disadvantages

There are, of course, advantages and disadvantages to both research roles.

As an early adopter, you can play around with fundamental questions, with a high likelihood of making progress. You often don’t need a lot of background before you can begin doing interesting work. Indeed, I suspect that a good way of finding new fields that are likely to flourish is to look for fields where PhD students are doing much of the best work. Finally, as an early adopter you don’t need to worry so much about being scooped

Of course, there are also many disadvantages of being an early adopter. First, it’s obviously not trivial to pick a research direction that is truly novel and is likely to be of long-term interest. For every field that goes on to become important and well established, there’s a dozen other “promising” nascent fields that fizzle and go out.

Perhaps a more significant disadvantage is that you lose some of the advantages of working in an established field: a large pool of colleagues, conferences, community, grants, jobs, recognition, and all the things entailed.

Outside the bounds of the dichotomy

To finish off, I want to briefly mention two closely related roles that to some extent fall outside the bounds of the dichotomy I’ve set up.

Pioneers: Essentially an extreme version of early adopters. These are the people who lay down the foundations for the narrative of a new field. Think of Shannon’s papers on information theory, Turing’s paper on computation, or Deutsch’s paper on quantum computing. In each case, a very important function of the paper or papers was to outline, in a primitive form, a narrative for a new way of doing science. (All the papers also solved scientific problems, but I think only in Shannon’s case was that the most important function.) This new approach is then taken up and the narrative is fleshed out by other early adopters, until it matures into a full-fledged research subfield.

Solvers of major problems: The most immediate kudos in research usually go to people who solve longstanding problems acknowledged to be of importance in some established field. This sounds like an example of late adoption, and sometimes it is – Andrew Wiles’ proof of Fermat’s last theorem did not, so far as I know, change the basic reasons why people do number theory, or what they consider important in number theory. However, sometimes the solution of such a big problem requires that radically new ideas be employed, and this can change the entire field. For example, to solve Hilbert’s entscheidungsproblem, Turing had to introduce a rigorous mathematical definition for the computer, paving the way for an entire new scientific discipline.

From → General

Comments are closed.