Excuse me, where is the mainstream?

More than once I went away from a discussion, confused about what exactly my conversation partners meant with “physics mainstream”. The “mainstream,” it seems, is typically employed as reference point for why some research projects get funded and others not. If it’s not “mainstream physics”, I gather, it’s difficult to get it funded. But can we come up with a good explanation for what is “mainstream”?

My first attempt to define the “mainstream” would be by the context it is most often used, the ease by which a research topic can be funded: the easier, the more mainstream. But on second thought this is not a helpful definition because it’s not based on properties of the research itself, which makes it circular. We could as well say a topic attracts funding easily because it’s mainstream, and so we are none the wiser. What is it that puts a research area into the main of the stream to begin with?

Reference to “mainstream science” is often made by pseudoscientists who are using the term in an attempt to downgrade scientific research and to appear original. (Ironically they then often boldly use and abuse vocabulary from the unoriginal mainstream. Google for quantum healing to see what I mean.) In that case the “mainstream” is just all that deserves to be called scientific research.

The way that pseudoscientists use the expression is not what I want to discuss today. It’s the way that researchers themselves speak about the “mainstream” that has left me wondering if it is possible to make this a meaningful, and useful, terminology. The way the expression is used by researchers it seems to have connotations of popularity, fashionableness, timeliness, and attracting large numbers of people. Below I have tried to make sense of each of these properties, and then I’ll offer the best definition that I could come up with. I invite you to submit your own!

Public attention

At any given time, some topics are popular. In physics there is presently direct detection of dark matter, quantum computing, topologic insulators and cold atom gases, for just to mention a few. But in many cases these popular topics constitute only a small fraction of the research that is actually happening. The multiverse, to name another example, is in reality a fringe area of gr-qc that just happens to capture public attention. The same is the case for the black hole firewall. In fact, in many cases what makes headlines in the press are singular or controversial findings. It’s not the type of research that funding agencies have on their agenda, which is why popularity is not a good defining property for the mainstream.

Fashionableness

High energy physics is a very fad-driven area of physics and quite often you can see a topic appearing and gathering momentum within a matter of months, just to then exponentially decay and hardly be mentioned some years later. Anybody remembers unparticles? The pentaquark? The so-called OPERA anomaly? These fads aren’t as extreme in other areas of physics (or so I am told) but they exist, if less pronounced.

Fashionableness indeed seems to some extent correlated with the ease of getting funding. But a trend must have been around for a while and have attracted a base of research findings to appear solid and worthy of funding, so that cannot be the whole story. This brings me to the next point.

Occupation number

The more people work on a topic, the easier it is to make a case that the topic is relevant and deserves being funded. Thus the number of researchers in an area seems a plausible measure for it being mainstream. Or does it?

The total number of people is in fact a highly misleading quantity. A topic may attract many researchers because it’s very rich and there is a lot that can be done. Another topic might just not support such a large number of researchers, but this says more about the nature of research in an area than about its relevance or its promise.

String theory is an example of a research area that is very rich and supports many independent studies, which I believe is reason it has become so dominant in quantum gravity – there’s just a lot that can be done. But does that make it mainstream research? Nanoscience is another example of a research area that attracts a lot of people, but does so for an entirely different reason, that being the potential of developing patents and applications. Throwing them both together doesn’t seem to make much sense. The total number of people does not seem a good defining property for the mainstream either.

The important factor for the ease of obtaining funding is not so much the total number of people, but the amount of tangible open problems that can be attacked. This then leads me to the next point.

Saturation level

A refinement of the occupation number is the amount to which a research area attracts people that study presently open questions on the topic. Mainstream physics, then, would be those areas that attract at least as many researchers as are necessary to push forward on all presently open questions. Since this is a property relative to the number of possible research projects, a small research area can be mainstream as much as a large one. It has some aspects of fashionableness, yet requires a more solid base already.

This definition makes sense to me because ease of funding should have something to do with the availability of research projects as well as their promise, which would be reflected in the willingness of researchers to spend time on these projects.

Except that this definition doesn’t seem to agree with reality because funding usually lags behind, leaving research areas overpopulated: It is easy to obtain funding for projects in areas whose promise is already on the decline because funding decisions are made based on reports by people who work in the very same area. So this definition, though appealing at first sight, just doesn’t seem to work.

Timeliness

Thus, in the end neither popularity, fads, the number of people, nor the availability of promising research topics seem to make for a good definition of the mainstream. Then let me try something entirely different, based on an analogy I used in the Nordita video.

Knowledge discovery is like the mapping of unknown territory. At any time, we have a map with a boundary beyond which we do not know what to expect. Applied research is building on the territory that we have mapped. Basic research is planning expeditions into the unknown to extend the map and with it the area that we can build on.

In either case, building on known territory and planning expeditions, researchers can take small steps and stay close to the known base. Or they can aim high and far and risk both failure and disconnect from colleagues.

This image offers the following definition for mainstream research.

Mainstream research is the research that aims just far enough to be novel and contribute to knowledge discovery, but not so far as to disconnect from what is already known. It’s new, but not too new. It’s familiar, but not too familiar. It’s baby steps. It builds on what is known without creating uncomfortable gaps.  It uses known methods. It connects. It doesn’t shock. It’s neither too ambitious nor too yesterday. It’s neither too conservative nor too tomorrow. It is what makes the community nod in appraisal.

Mainstream research is what surprises, but doesn’t surprise too much.

We previously discussed the relevance of familiarity in an entirely different context, that of appreciating musing. There too, you want it to be predictable, but not too predictable. You want it to have just the right amount of complexity.

I think that’s really the essence of what makes a field mainstream: how tight the connection is to existing knowledge and how well the research is embedded into what is already known. If a new field comes up, there will be a phase when there aren’t many connections to anything. But over the course of time, given all goes well, research on the topic will create a map of new territory that then can be built upon.