Welcome to the age of diminishing returns

Friday, July 22, 2011

Peak Research

Joseph Tainter's model of decline is based on the idea that civilizations attempt to counter the effect of declining resources by creating more complex structures. That strategy fails to bring the desired results because of the diminishing returns of complexity. The same factors may be causing a decline in the worldwide effectiveness of scientific research, plagued by bureaucracy, strangled by excess of rules and controls, and weighed down by lack of resources. (My detailed interpretation of Tainter's model can be found in a post titled "Tainter's law, where is the physics?")

Life has been very hard for scientists during the last few years. Already, the life of an active scientist was a rat race in which you had to run in circles, trying to get grants that would allow you to pay students and postdocs and they would help you write more papers that then would be used to support proposal that would provide you grants that would allow you to pay students.....  It has always been like that, but in the last few years it has become hell. More and more bureaucracy, tight controls, guidelines to follow, time schedules to keep and less and less money. And, of course, any attempt to do something creative and a little outside the known schemes seems to be becoming impossible to finance.

I think that the situation is like that everywhere in Europe and the US, at least from what I hear from my colleagues: reduced budgets, more paperwork, and the sensation to be running a rat race. I couldn't find data about the worldwide situation, but these data from the US do suggest that we may have peaked in terms of resources available for scientific research or, at least, we are plateauing. (source: Task Force in American Innovation)

But the question of resources may not be the most important one. What I perceive is, rather, a declining quality of the research being performed. I may be wrong, because it is hard to quantify an entity such as "quality of research". But my impression is that it is becoming increasingly difficult to perform original and innovative research within a system that provides resources for doing that only if researchers submit to a series of tight constraints. Not long ago, I was hearing a talk by a well intentioned presenter who endeavoured to teach to young scientists how to successfully apply for research grants. I don't know what the young scientists thought of that. My impression was that the presenter could have been described the rituals of an esoteric cult dedicated to adoring the God Quetzalcoatl; human sacrifices were not requested, but almost.

Don't get me wrong: I am not saying that I don't like to do research any more. I love doing research and I wanted to be a scientist from the time I read my first science fiction novel; I think that was when I was about six. And I am not saying that science is not progressing any more. Absolutely not. I am amazed by the progress being done in many fields, for instance, in climate science. And that is done despite climate scientists being threatened, harassed and insulted for what they are doing.

What I am saying is that the state of scientific research in the world seems to be a nice example of Tainter's interpretation of the diminishing returns of complexity. Tainter had devised his model to explain the fall of civilizations - he had mostly in mind the Roman Empire. His idea is that when civilizations face diminishing resources, they react by building up more and more complex structure to cope with the problem. But there is a diminishing return to complexity; these structures become burdens rather than solutions and help bringing down the whole system (a discussion of mine on Tainter's model can be found here).

Tainter's model has a certain "fractal" quality; that is it applies to subsystems just as it does to the whole system. If you look at the Roman Empire, you see that all its subsectors where declining together. Can you cite a Roman poet after Virgil? Most likely, you can't. Not that there were no poets after Virgil, but  Roman literature declined with the decline of the Empire and we found little or no interest in refined but shallow poems such as those written by Claudian in the 5th century AD.

Something similar seems to be happening in our times with scientific research (and possibly also with literature). It seems that, facing a declining supply of resources, the structures that manage scientific research try to compensate by building up a new layer of bureaucracy aimed at "optimizing" research - just as they are trying to optimize teaching. It means that, when you obtain a research grant, you are told in exquisite detail exactly what you are to do, how, for how long and that any departure from the plan must be justified. It seems that the very concept of "research," intended as looking for something new, is not allowed any more in these plans. You can't get funded unless you already know what you are going to find.

That's the perfect recipe for that "excellent mediocrity" which is the bane of scientific research. It was already a problem with the phenomenon known as "publish or perish", but with the bureaucratization of research it has gotten much worse. I could give you a series of funny (or tragic) examples from my own experience, but let me skip that. It just seems to me the system is becoming innovation-averse; it is like if "research" had become an oxymoron all by itself. The point is, could we do something about that? Don't get me wrong, I am not saying that scientists should be like Dr. Zarkov in the first episode of "Flash Gordon," where he builds an interplanetary spaceship in his basement. Science is a collective task that requires coordination, planning and some degree of control. But how to turn research into something that can change the world for real? Something that may help us to attain sustainability and stop destroying the ecosystem?

That's difficult of course. Bureaucracy is a tool to keep the world as it is, not to change it. So, in perfect Tainter-style, the system works hard to avoid innovation, not to promote it. It is almost impossible to be financed to study resource depletion; that would highlight problems that would require changes and that's a no-no. Instead, it is still possible to obtain research grants as long as there is no risk that the results will threaten the status quo. Hydrogen as a fuel is a good example. It is high-tech, fashionable, sophisticated, popular, environmentally friendly, and it doesn't work. This last characteristic makes sure that its development will bring no changes whatsoever. Absolutely perfect for the bureaucrats who manage the research grant system.

Thinking about that, I feel like the centurion of Kipling's story, the one who dedicated his life to the defense of Hadrian's wall in Northern England.  The Romans could not reform their Empire to survive decline; they fell in the trap of diminishing returns described by Tainter. And if the Romans failed in reforming their empire, who are we to think that we can reform research?  Tainter's law is harsh.

But, if the idea is to make research more creative, we should think creatively. About the problems of the Roman empire, I said that there was only one solution and that was called "Middle Ages". The only way to save the Empire, intended as its culture, art, laws and all what goes under the name of "civilization", was to break it apart. In a way the solution to keep the Empire alive was to kill it. As the Zen Masters say, "when you meet the Buddha on the road, kill him!"

Could we think of something just as drastic for scientific research? Yes, it would mean to leave the tight world of the research grants and find new ways to perform better research - more independent, more creative. Moving, in a way, to an equivalent of "Middle Ages" intended as the break-up of the old, tight structure. And, think about that a moment, maybe there are already examples of this approach. Think of Wikipedia - it was not created by bureaucrats; it is a free sharing of information done by people who work for free. Think of "free and open source software," that has generated the Ubuntu operating system which I am using to write this post. Probably there are more examples of good work that can be performed not because you are paid to do it. It is said, after all, that "things done illegally are done most efficiently."

That doesn't mean that it is illegal to do scientific research without being told to do it by some bureaucrat, at least so far. But I think that some of the best scientific work I've done in my life (or perhaps the least bad work I did) has been done outside the boundaries of the granting system. A couple of examples are the papers on resource depletion that I and my coworkers recently published  (here and here). All done on a strictly zero budget - but so what? Scientific research is about sharing, after all. I think we should at least try this approach.

With the gradual collapse of the Roman Empire, Roman poets such as Claudian could only write elaborate and shallow praises of those who paid them. That wasn't the end of poetry in Europe: at the same time, the great sagas of King Arthur and of Sigfried were being written outside the Empire's borders by unknown poets who knew nothing of the elaborate rules of Latin poetry. So great was their creative energy that their stories were told for millennia after them and are still known today. So, creative energies can survive declining times and that is perhaps true also for scientific research in our times.


Note: you may wonder why I am citing so much the Roman poet Claudian in this post about scientific research. Well, it is because I am preparing a post on the figure of Galla Placidia, the last Roman Empress who lived in the 5th century A.D. The historical sources from that period are scant and, therefore, even the poet Claudian may be used as a source of data. And perhaps I have been too harsh in my judgement; his poems have some charm.... ahem.... occasionally.  


  1. A brilliant post; you are completely right, the life of a scientist is becoming increasingly miserable.

    Not only are scientists straighjacketed by bureaucracy, many also feel the dead hand of corporate sponsorship. I believe that the multiple deficiencies of transport biofuels for instance have been masked for too long by too many scientists failing in their duty to falsify hypotheses.

    James Lovelock (who has worked mostly outside the academic straightjacket)recently said: "I hate academia. Most of the scientists who work there are not free men any more and they can't speak out. That's no way to do science."

    I think good science is incompatible with modern management concepts - imagine asking a Darwin or an Einstein what their 'deliverables' would be. As GB Shaw put it:
    "The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man."

    But to mirror your optimism about new directions, look at the wonderful results from imaginative crowdsourcing approach that is Galaxy Zoo ... http://www.galaxyzoo.org/ ... we can and must do better.

  2. When I was in grad school 25 years ago, our most successful faculty member had the following modus operandi. He would have his research team work in total secrecy sharing information with nobody. When the research was completed, he would write a proposal to do the project he had just completed. With his sterling reputation, the proposal was quickly approved and he would take the funding to work on a new and different secret project. His progress reports always showed results exactly as he predicted. Of course all of the progress reports had been written before the research proposal was written. And thus, the research funding kept rolling in. What a genius!

  3. Alas, Bilbo, recycling old results and selling them again to a different funding agencies is an old trick. It parallels that of writing twice the same paper for two different journals. With the great mass of paper being produced, both tricks may go unnoticed.

  4. Very nice link to the Galaxy Zoo; thanks! And I absolutely agree with you about bioufuels.

  5. Ugo,

    You know; it's a very funny thing about Tainter-type "complexity" --but in digressing here, I do plan to loop back to the peaking of R&D--. The funny part of Tainter complexity is that it always comes with a flip side.

    While a society is undergoing increased complexity, it also undergoes a great "simplification". It's ying and yang, both at the same time.

    Consider Tainter's definition of societal (anthropological) "complexity". It says that we have an increasing number of highly specialized job functions. One person may spend his entire career studying a particular kind of cancer cell and another may spend his work life developing a new kind of cell phone. But in doing so, each develops a kind of cocoon or isolation membrane between himself and the rest of the world.

    The rest of the world becomes extremely "simplified" when viewed through the cocooned shell of the specialist (the complexified individual).

    Ask your average cellular biologist or your average iPhone cell phone engineer about where "oil" comes from or where "money" comes from and they will have a very pat answer:

    "Sorry, I'm no geologist; I'm no economist; I'm no this or that, I'm only this highly specialized me. But given that, I (the specialist still speaking here) am sure that somebody else in this complexified world of ours is taking care of their part of the bargain. So no worries mate."

    So let's jump back now to the interface between politicians and research scientists because each lives inside a same kind of cocoonized shell that every other specialized individual lives in.

    You average politician will say (same pat answer):

    "Sorry, I'm no cellular biologist; I'm no economist; I'm no this or that, I'm only this highly specialized me. But given that, I (the political specialist) am sure that somebody else in this complexified world of ours is taking care of their part of the bargain including making sure that the right kinds of R&D get properly funded and done. I've got many complicated problems spinning in my head, but worrying about R&D funding, sorry that is not my specialty.

    So there you have it.

    As Spiderman's grandfather might have said to the superhero in the movie: "Son, with great complexity, comes great simplicity. Be wary of that".

  6. "coping with resource decline by greater complexity"
    "coping with resource decline by greater efficiency"
    are the same thing!
    Efficiency/complexity is good but orthogonal to the solution (if any)

  7. Part II:

    I was not spinning the Spiderman web above for no reason at all. There is a follow up to dangle off of it.

    Here is one politician's view of how R&D takes place in the cancer research world:


    In case the above hyperlink get's broken, here are a few key snippets out of it:

    "... and is [now] a member of the Washington State Senate." [the writer is a "politician"]

    "I had stage-III lung cancer, and I'm alive today because of a medical miracle."

    [Every politician knows how to get your attention and how play your heart strings]

    "With a streamlined patent system, American inventors will have the incentives and protections necessary to drive innovation, create jobs and stimulate the economy."

    [Here (above) is where the politician has it exactly backwards but is not alone in his simplified view of how and why R&D gets done]

    "But to make miracles like this happen, biomedical companies require the predictability and fairness in U.S. patent law that reform will bring."

    [Wrong. Scientists need funding to do basic research that then leads way down the line to commercialized products. The horse comes before the cart. But hay, I'm no horse-and-cart specialist. So what do I know?]

    [Aside from that, the writer is totally wrong about the American "reform" of their patent laws, but that is a topic outside of this conversation.]

  8. I like your comments Step Back. I would like to be a generalist but academia seems to favour (hyper)specialization, until it doesn't.