Asking “good” questions is challenging and is perhaps one of the most important things that we do as scientists. The type of question that a scientist asks—or the absence of a question at all—plays a large role in how they appoach their subject. Of course what qualifies as a good question is highly subjective, and as a result, any commentary on question quality in what follows is a reflection of my own biases. My point here is less to make value judgements on what is an interesting vs uninteresting question, and more to understand what goes into the process of formulating and pursuing a scientific question, while noting some apparent trends across physics and biology.

It is also true that in many cases, the nature of a question changes as one tries to answer it: it is hardly the case that science consists of a neat sequential process of asking and then answering questions. Many times random explorations can later be given meaning by projecting onto them a post facto question that only made sense to ask in hindsight—for that reason science that is not solely question-driven is surely also a valuable thing. In fact being too tied to any one way of doing science seems like a recipe for disaster: the fact (or at least my belief) that there are infinite ways of being a good scientist means that in an ideal world different people occupy many different niches.

What counts as a question?

I was recently at a workshop on microbial ecology where, during a session devoted to framing open problems in the field, the discussion leader wrote on the board a number of questions to spur contemplation. These questions were along the lines of “How can we better measure metabolic activity of complex ecological communities?” or, “How can we integrate information from microbial genomics, physiology, and ecology to characterize natural communities?”

For the sake of argument, I’ll classify these two questions as “type-A” questions. Although they start with the word “how,” which might point towards an inquiry about mechanism, they are ultimately methods questions, that must be answered with a new tool or toolbox. In the case of measuring metabolic activity, maybe this new tool is an advancement in spatially resolved metabolomics. In the case of integrating different modalities of biological information, perhaps advances in sequencing or data analysis or the ability to measure different physiological outputs could play a role in the answer. But in both cases, the questions are rather technical in nature: their asking and answering is meaningful only if you (1) have a large amount of context on what the state of the field is and (2) are invested in accruing information about the natural world for its own sake. This type of question fuels our remarkable desire to categorize and classify natural world. Their answers amount to new data points on the coexistence and characteristics of microbial communities, but do not necessarily superimpose an interpretation onto these data points.

Type-A questions are essential for the driving a field: they are the questions whose framing pushes us to increase the sheer volume of what we know. However, eventually a different type of thinking becomes important: once we know many facts about a subject, the scentific endeavour pushes us to condense these facts into a (hopefully) small number of principles. I would argue that this drive to distill can be fueled by a different type of question: for the sake of argument, “type-B” questions. Some examples of type-B questions are: “how do bacteria sense and navigate their environment reliably?” or “how does evolution generate extensive microbial diversity?” Type-B questions are essentially about mechanism: I would say they capture the spirit of the “how” interrogative more than the type-A questions. Type-B questions push us toward some combination of empirical measurement and theoretical synthesis to provide a cogent explanation for some observed natural phenomenon.

The type-A questions (or “methods” questions) can be phrased as in service of some broader science questions (which might ultimately be a type-B question). In other words, the majority of scientists only cares about measuring metabolic activity insofar as it helps to understand some natural phenomenon like the coexistence of microbial diversity. By the end of our discussion at the workshop, it seemed to me that, starting from type-A questions and asking further what concepts these type-A questions were servicing, we had essentially converged to one type-B question, namely: how does so much microbial diversity coexist in nature, and why is it continually generated by evolution?

This is crucially a question that is motivated by an obersevable phenomenon i.e. that of the microbial diversity present in nature. In its asking, one makes an observation about the natural world and then seeks to understand why it is the way it is, which necessitates both methods development, but also information synthesis and distillation.

In more mathematical sciences such as physics, different types of questions that are not quite type-A or type-B seem to emerge. One can ask questions about mathematical models which aren’t directly related to observation or experiment. For example, in asking if the solutions to the Navier Stokes equation are smooth for all time, one is less interested in the actual dynamics of fluids, and more interested in the behavior of the mathematical structure that we use to describe them. After all, the whole concept of smooothness breaks down if you look closely enough at a real fluid, which is composed of discrete molecules.

Again I would argue that these questions about models tend to be in service, even if only vaguely, of some type-B question which is derived from a natural phenomenon. Here the link between phenomenon and question can become tenuous very fast but is ultimately, I think, important for motivating choice of model to study.

“Why” Questions

Having discussed a number of “how” questions, and how the use of that word alone is not dispositive for the spirit in which the question is asked and answered, it is particularly interesing to consider the nature of “why” questions and whether they are answerable at all.

In physics, the purview of “why” questions depends on where they are asked. In emergent phenomena, i.e. “why does aluminum lose all electrical resistance below a certain temperature?” “why” and “how” questions are not so easily differentiated: the answer to a “why” question calls for a mechanistic description in terms of known physical forces, which is the same treatment elicited by the answer to a “how” question. But, riding up the tower of subsequent “why” questions, assuming that the varying shells of emergence can be explained in terms of a few effective quantities derived from their fine-grained descriptions, one eventually arrives at so-called fundamental why questions, i.e. “why is the strength of the coulomb force what it is?” or “why is there more matter than antimatter in the universe?” These questions seem to push the boundary of what is scientifically answerable, and spur the development of apparently-unfalsifiable explanations such as the anthropic principle.

In biology, I would argue that similar types of “fundamental” “why” questions all lead to evolution. There are also “why” questions that are basically mechanism questions, e.g. “why does the circadian rhythm in many organisms oscillate robustly even in the absence of external forcing?” but here the emphasis is key. I would posit that the question of “why the circadian rhythm oscillate robustly”, while partially addressed by a detailed description of the protein interactions that drive oscillation, ultimately can only be satisfactorily answered by an understanding of how evolution drove the circadian rhythm—and its mechanism—to be what it is in modern biology. Indeed, Dobzhansky’s dictum that “nothing in biology makes sense except in the light of evolution” is perhaps true, but I would add that many things in biology don’t (yet) make sense even in the light of evolution! Ultimately it seems to be an eventual goal to understand not only how biology works, but why it is the way it is, i.e. how evolution gave rise to it. The question of how evolution creates biology can be asked in all subfields, from neuroscience (“how did a system which can so plastically learn and adapt?”) to development (“how did robust organismal developmental programs arise?”)

Can these questions be answered at all? If evolution is unpredictable, then perhaps not: perhaps the only possible answer is a just-so answer: things are the way they are because a series of random events led them to be here. But this is surely not true for all features of biology. What aspects of life have some air of inevitability about them, such that we expect them to arise time and again through evolution? Identifying and answering these “why” questions can give us some perspective on what features of biology we might be able to understand in an existential way, in addition to a mechanistic way.