One of my favorite comics is XKCD. These are short comic strips with sick figures, very often dealing with a scientific or a philosophical topic. On of my favorite comic, one of the favorite strips is on the purity of science.
It reflects the idea that mathematical representation of reality is more real, or pure, than the representation itself. In it the path of purity goes like this: sociology-> psychology-> biology-> chemistry-> physics and after a large gap, you have mathematics.
I am sure Randall Munroe wrote the strip in jest, but I think it reflects an animistic nature of mathematics. A belief that mathematics somehow weaves reality into existence. Of course if you asked theoretical physicists, or mathematicians, whether they believed this, they would probably say "of course I don't believe this". But if you were to insist what they believed the true nature of reality is, they would be bound to fall back to mathematics and to equations describing experimental phenomena.
I think this mode of thinking is putting the cart before the horse. Mathematics is an immensely powerful tool to understand reality, but mathematics isn't reality in itself. There are no imaginary numbers in the real world despite their usefulness. The idea of believing mathematical tools to be equal to reality is like believing that our favorite actors live on the screen from which we watch them.
In the same sense, there is no physical wave function. This doesn't mean that it isn't a useful tool to accurately describe how physical phenomena behave, but real world is not probabilistic.
We can take an example of water. The primary supramolecular shell of water is a double sphere of 2898 nm. But this state is dependent on its surroundings. For instance, if you add water-miscible solvent, such as ethanol to water, the ethanol molecules form a supramolecular shell around a core of water. If one wished to accurate describe what would go on in this process, my best bet is that you would require the Schrödinger equation to describe how you get from a state from where water was not mixed with ethanol to the state where ethanol had formed supramolecular orbitals around water. However, if you don't care for the process and only care what was the state in the beginning and the state afterwards, the Schrödinger equation becomes unnescessary.
In my case, I can indirectly see the end state by dynamic light scattering. The only problem is that my data is in the form of the light scattered by colloidal lignin particles in mixtures of water and ethanol. But because I have to admit that my skills in handling the Schrödinger equation is pretty much nil, I have no way of expressing my chemist's intuition in a publishable way.
That is, the dirty secret of natural science is the need to not be wrong. If you cannot formulate an equation that describes your data, you are not allowed to speculate on the data. If your data relates to something relevant in the manuscript, you are allowed to say something rather circumspect, not really claiming anything. However, what you can's say is that: " This graph might mean x, but I cannot be certain of this". Peer review weeds out informed guesses, until all you have left are platitudes.
But this is not how scientific discoveries are made. You begin with a hypothesis and test it in all manner of conditions and only subsequent results show whether your hypothesis was right, wrong or more often something in between. Natural sciences does not have the equivalent of qualitative research. I do not mean qualitative analysis, which is more or less based on data that is still based on quantitative data. What I mean is that you cannot publish ideas that are still in the process of formation.
In my case, the idea of the supramolecular shell began with the graph below. It describes how one of our LignoSpheres behave in mixtures of ethanol and water. This graph is at the same time replicable, but unique to the sample. This means that to obtain exactly the same data point one would have to replicate the experiment exactly the same way each time. Additionally, I observed that while the sample was stable for quite a few weeks, eventually the data points (obtained by replicating the measurement with old dispersions) started to deviate from the original results.
So what does this graph mean? Well, first of all, the y-axis is the diameter (or more specifically the z-average diameter) of our lignin spheres, as measured by dynamic light scattering. And the x axis is just the ratio of water to ethanol, into which I diluted a droplet of a dispersion of LignoSpheres. Is there any theory to explain why the sample would suddenly show much larger sizes at very specific ratios of ethanol and water? No, as I was the first one to observe this phenomenon.
What to do next? Of course follow the scientific method. Form a hypothesis, test it and analyze the data. At first I had an idea that the graph related only to the LignoSphere sample, because how could it reflect anything else? Only after a long time of further experiments, it seemed that this relation was actually something to do with the interaction of ethanol and water. The problem was that there was no theory to explain such peculiar behaviour.
I was able to use the graph to make further hypotheses and test them. Sometimes the hypotheses were clearly wrong, sometimes they seemed to match very well with the expected outcomes and sometimes the results were vague.
It took a long time for me to transfer the above graph to the concept of the supramolecular shell. To do that, I used a shortcut version of the scientific method. As my research was always leading to commercial products, I couldn't reveal the raw data to the public, to let them figure out what the graph might mean and snatch the ideas from us. So I was working for a long time in a valley of uncertainty, where I made hypotheses upon hopotheses, all based on incomplete data. But I knew that if the hypotheses were completely off, it would show in our R&D.
The rapid scientific method
The problem with the rapid scientific method is that you can go badly wrong, if you go on long with it. At some point you need to stop and publish.
Earlier this year I still had the conundrom that I already had the basic idea of the supramolecular shell, but only an intuitive idea of how to explain it. It dealt with hydrogen bonding and hydrophobic interaction. The basic idea was that the polar end of ethanol would be attracted to water, whereas the non-polar end would not. This would cause ethanol to form a crust over a core of water. The problem with this idea is that the way scientific papers are written, I couldn't just put the graph above into an article and state: "This is roughly what I think is going on. Perhaps someone will figure out a more formal description."
The reason why this would have led to the rejection of such a paper is that I would have no way of knowing whether I am right or wrong. And the peer reviewers would be right in saying that. I was still on the level of a chemist's intuition.
In peer review there is no such thing as Ideas Worth Spreading, as they say in TED. If you have an idea, you keep it to yourself, until your idea has been formulated into a proper theory that can be verified or falsified. It doesn't matter if you write the manuscript in a way where you state clearly that you don't know whether you are right, but you assume x.
We hear of physicists (and sometime chemists) using thought experiments to come up with experiments that aid in laboratory experiments. The problem is if the person coming up with the thought experiment is neither able to translate it to an equation or an experiment. Currently, there is no proper path for a publication of just a thought experiment. It has to be formalized and tested to warrant publication.
In my personal opinion we as a scientific community, should be much more tolerant of incomplete and partially wrong ideas. Oftentimes the most important observations when conducting research are things that didn't work out as planned. Even when you are able to publish results that could be considered failures in relatively respectable journals, such as happened to me, it is not advised to speculate why your initial hypothesis failed.
Backtracking to how I turned the idea of the ethanol crust to the concept of supramolecular shell, I had to make a lot of untestable thought experiments, assumptions and altogether ignore the scientific method. Only when I had the idea of the supramolecular shell figured out, I had to figure out how it connected with my experimental observations.
This presented a dilemma. When you use intuition to aid in research, you leave gaps that seem unexplainable. You hope that you are able to explain the leaps of faith you had to make, but end up being unable to.
This is what happened to me. At first I tried to write a manuscript with all of the bits and bobs included, which had helped me figure out the theory. This led to the rejection of the first manuscript as speculative.
This wasn't all bad, though. Having to explain why I thought the supramolecular shell was feasible, I had to go back to what Einstein had to say about the photoelectric effect and figured out that the problem was with the kinetic theory of gases. So, in the end I had to introduce mathematics and reference to existing theories of physics, but neither of these was really the source of the theory.
And I ended up not including the bits and bobs in the revised manuscript, even though they would have explained a lot of why thought that there would be a such a thing as a supramolecular shell. The reason being that I could not show how the bits and bobs were connected to the final discovery. So what I submitted for peer review was only what had any chance of being verified or falsified.
The most mathematical part of my theory is armchair topology. If I had taken a single class of topology when I minored in mathematics, I could say that I recalled coursework from twenty years ago. But I hadn't. And it's almost as bad with quantum mechanics. I studied quantum chemistry for a physical chemistry course, but hadn't really needed it in the twenty years since, despite working on all sorts of fields of chemistry.
And getting back to XKCD. A supramolecular shell is phenomenon of physical chemistry, which is a sub field of chemistry, not physics. If it turns out that supramolecular shell explains the fundamentals of many more phenomena than just the properties of water, it just might be that the order of purity of sciences needs a bit of a reshuffle.
Komentarze