You probably know Nikola Tesla, the man who according to a series of sites you can find around invented basically everything. Not quite, however. If you go and read a proper Tesla biography, such as W. Bernard Carlson’s Tesla: Inventor of the Electrical Age you will discover that many things said about Tesla are myths.
However, there is consensus in attributing Tesla the invention of the induction motor and of rotating magnetic fields. Now, suppose Tesla hadn’t ever been born. And I present you with a question: Would we have induction motors? And what makes you think what you just replied? There are two main views you may hold:
One, that we would not have induction motors, as we don’t have the inventor among us anymore. So modern technology would be what is sans the induction motor, and any technology that depends on it.
And two, that basically nothing would have changed. Others would have invented it roughly at a similar time, so modern day technology would be the same.
I argue that reality is basically the second case. Now I will elaborate a little bit more about why it is so. I think you had the intuition that the second case is closer to reality, but that’s not necessarily always the case. Now consider this: What if Archduke Franz Ferdinand had never been assassinated? How would have the world changed? No World War I? And no World War II? I think here it seemed as if one man did decide the course of History.
So why do we think that the Great Man Theory holds for some historical events, but doesn’t quite seem the case for innovation and technology?
It is because History is all about circumstances, and the interplay of very complex forces that vary overtime, with no definite end result. In Franz Ferdinand’s case, maybe there would have been other assassination attempts, but the fact one attempt happened would make it harder for others to follow, as extra caution would be taken. Circumstances that enabled the assassination would be no more. Or take Hitler. What if he hadn’t been born? Would a similar person appear in his place? Maybe, if the conditions of post-Versailles Treaty Germany held like that, it would be the case. Or if Germany’s economic conditions recovered before the next Hitler appears, then there would be no Hitler 2 at all, again, because circumstances.
In contrast, innovation is a different process. All innovation is trying to do more things better, at a lower cost, and that’s basically an optimization process driven by Physics. Innovation in this sense is filling a bucket with water, and the bucket’s volume is fixed. When the bucket is filled, the best mean for a given end exists, given Physics. It is possible to fill this bucket faster or slower, and it is also possible for water to evaporate (to loss knowledge), but empirically, the stream of water going into the bucket is ever increasing, and the evaporation rate is rather low.
It could be argued that there are diminishing returns into play here: with a given set of resources you could do more innovation in the past than now. Now you need to spend quite a lot of money to acquire equipment to be able to build upon what we know and further our knowledge. Granted, but another empirical regularity is in play against this: The share of GDP spent on R&D rises with economic growth and then stays relatively constant. And given the exponential growth of GDP, R&D spending will also follow an exponential pattern.
What could change is the path of technological development. Suppose no World War 2, and so rocketry didn’t develop for military purposes as fast and it develop later instead. But then, scientists devoted to rocket science would have been researching other things, and those would have advanced first.
As a conclusion, what I’m saying is that although some resources devoted to research and application of technology, no single especial person is required for advance to occur. If you go back in time and kill of the great historical geniuses, now you have read this -and if you read some History of Science and Technology- you won’t worry about the present state of technology.
Now let’s talk about governments. Governments can increase the supply of funding available for R&D up to whatever levels they want to. They can build research labs, particle colliders, fund universities, and so on. Governments have one apparent difference from other agents in that they can direct incredibly large sums of money without regard for future profits. So what if there are beneficial discoveries that won’t be researched unless they are government funded, because the ways to reach them are so indirect that there is no apparent connection between research and the final product, and no profit calculation can be done?
Suppose discovery D can only be done by building a rather large particle collider, and that discovery D can be applied to improve Computerised Axial Tomography scanners. Now assume that this fact is not known. So who would bother building a particle collider? An initial reply is that science is not always done with an end in mind, but it is also propelled by curiosity, and a desire to understand the world. So physicists would keep working on theoretical physics, theories would accumulate, and demand for empirical testing would emerge. Even with a given level of fixed % of science philanthrophy over GDP, as GDP rises, even very expensive projects can be undertaken that way. And then you would get the D discovery. Unless, perhaps, it is the case that you can get D via other cheaper ways. You never know where a new discovery is coming from. The LHC, for example, costed around 7.5 billion euros. There are quite a few billionaires, or corporations, who could pay for that, if there were no colliders already built.
But another argument against Big Science is available, namely that it doesn’t empirically deliver. Our comprehension of the world may be enlarged by them, but with respect to useful technology, there are almost nothing that Big Science has given us. If one goes and tries to find such thing, one usually finds that in the construction of the project, new materials had to be developed and similar things. But that could have also, and easily be reached via normal R&D. I mean that the facts we discover with these instruments (Say, the Hubble Telescope, the International Space Station, the LHC, Tevatron, etc) do not have technological applications. And I say this as someone who really love science and technology, and building machines to understand the universe, even if that understanding has no use in directly improving our lives. It always amuses me when scientists working on areas far removed from practical application try to defend their field saying that the by-products of Big Science are useful. That’s nonsensical. They should defend the value of pure scientific understanding directly. But in doing so, they are conceding the truth: namely that the research they are doing won’t probably be useful in a material sense. And material usefulness is what is usually put forward to get government grants. Suppose I ask people “Hey, can I take this money from you for this research you won’t even understand? It will save millions of lives” vs “Hey, can I take this money from you for this research you won’t even understand? It will help us know the universe better”. It doesn’t pack the same punch neither from a rhetorical or ethical point of view.
This does not mean that governmental funding of science cannot accelerate discoveries. I’ll get to that in future posts. But it means that there are no important discoveries from a usefulness point of view that require governmental action. Given that for policy recommendations values are needed, if you hold that governmental action needs very good reasons to be justified, there are no sufficient reasons to justify it in the field of science. Now, if you think of government as a tool, and you like to make research happen faster, you can justify some spending. But for the reasons presented here, and elsewhere, this is hard to do with a consequentialist cost-benefit analysis. You can’t just calculate expected utilities for things like these, and if you could, there would be rather low. Lower than regular science projects, to which you would chose to devote resources. So in this case, policy would look like throwing money to researchers in universities or research institutes, maybe conditional on them making public their findings.
The conclusions reached here may seem extreme, but I think they are reasonable. There are surely some objections that could be raised against this view, and I will be happy to answer them -and change my opinion, if one of them succeeds.
The full argument, so that you can see where mistakes may be:
- The realm of physical laws is fixed
- So the realm of possible (technological) innovations is fixed
- Generally, what is learned is not forgotten, from a global perspective
- So innovation is cumulative
- If there is a significant amount of resources devoted to innovation, every innovation will be developed at some point
- There is a significand amount of resources devoted to innovation by default
- Growing GDP’s
- Rising or stable share of R&D spending over GDP
- Rising absolute amount of resources spent on R&D
- Therefore, every innovation will be developed at some point
- Reasons in 6 are government-independent
- Therefore, no government is required for significant amount of resources to be devoted to innovation (Conclusion 1)
- Therefore, every innovation will occur without government funding (Conclusion 2)
- Big Science does not lead to significant technological innovations
- Empirical argument: Historically, not the case
- There where Big Science does lead to significant technological innovations, it is possible to fund them privately at some point given Point 6
- So even apparently pointless but understanding-expanding discoveries will be discovered, given the innate curiosity of human beings.
- Big Science fails a cost-benefit analysis against Small Science
- Governments can, in principle, make research advance faster
- IF government action requires strong justification, science spending is not justified. (Conclusion 3a)
- IF government action is merely instrumental, and faster advancement of science is a worthy goal, then some government funding of science is justified, but not in the form of Big Science projects, for cost-benefits reasons, unless the furthering of understanding is regarded as an end in itself (Conclusion 3b)
I expect the most vulnerable premise to be 8. Some may want to challange 11.1 and say Big Science has done a lot of useful things. Good luck with that. And a further inquiry into 15 deserves attention. There are studies on this. Another criticism would be to grant the argument, but argue that progress would be really slow.