stimulant

There is a misperception of science/math as “the way” to teach rational, critical thinking. Unfortunately, that puts the cart before the horse: math and science can provide fantastic contexts for rigorous, critical thinking. There’s no doubt that a strong education in the hard sciences can be a straightforward contribution to a student’s ability to think rationally. However, it is a mistake to think that this benefit is intrinsic to the content. If anything, the elegance and structure you see in science and math makes these subjects more prone to subversion by broken learning strategies like pattern matching, wherein students learn to recognize and exploit patterns in problems presented to them without understanding the basic ideas involved. An aptitude for symbolic manipulation can very easily be mistaken for skilled, critical thinking. Indeed, many people’s academic successes are founded upon their ability to manipulate symbols. We should not assume that to ensure that people can think critically, we need to make sure that they understand the quadratic equation.

[Tragi]comically, the humanities and “soft sciences” in fact need more careful critical thinking, lacking the painstakingly constructed scaffolding of mathematics and scientific methodology. Without it, academia’s attempt at self-justification alienates far more than it educates. Consider the public’s reaction to deconstructionist and post-modern thought. The misappropriation by academics of these ideas has led them to become diluted and meaningless. Worse, they’ve made tremendous room for simply bad thinking.

Conversely, the hard sciences have a much narrower domain: for centuries, people have very carefully built up a reliable foundation. Self-consistency, applied over hundreds of years, has created a modular, conceptually clean set of tools upon which to rely. The humanities do not have a fundamental structure built in by anything analogous to physical principles like conservation of energy or mathematical formalisms like index notation.

Maybe in the past, the social sciences had to accept a certain degree of disengagement from hard data, but technology obviates this claim. To see this, let’s step back and think more carefully about what advantages the hard sciences have when it comes to critical thinking and rigor.

An oft-cited aim of the hard sciences is to create models of the way our world works that is as true and self-consistent as possible. For the more fundamental sciences like physics and chemistry, their object has been historically predictive. That is, given a situation, predict what will happen in the system. As the systems considered increase in complexity, science shifts its goal to description, with the implicit acknowledgment that characterization must come before understanding.

So even within science, we see that as the systems considered become increasingly complicated, there is a commensurate shift from understanding principles to characterizing behavior. This is a misleading spectrum: in both physics and chemistry, characterizing the system is an essential step. It is just that it is rarely a problem. Measuring quantities like mass or velocity or electric fields is far simpler than trying to get a handle on the microscale structure of the brain.

It is a mistake to suggest that all systems’ behavior can be predicted given the right theory or enough information. I think that this realization is one of the fundamental shifts that will need to happen before we can start making more progress in the study of complex systems. Our old ideas of what comprises an experiment encourage thinking that predictive principles are the only ones with value. Our study of complex systems is slowly changing that idea, and one domain in which we can see this is the study of complex social and economic systems.

The field of data visualization is exploding. Tools like ManyEyes and GapMinder have not only made strong visualization tools available, but provided access to high-fidelity data sets so that it is possible to turn to them to explore complex socioeconomic questions with a data-driven approach.

We are constantly asked — whether as citizens, parents, consumers, or simply people — to make sense of many, low-credibility sources of information. People frequently note that, “we are drowning in information and starved for knowledge.” The distinction is one between data and conclusions or interpretations. The strength of data visualization is it can make easy visual reasoning about complicated elements of a system, revealing trends or interpretations that would typically be obscured. Visualization tools have gotten to the point that we should integrate their use into our conversation. In the class, the teacher dispenses some information in the article. Then, the class discusses this, essentially brainstorming for the entire class. The internet has made the availability of data and visualizations a non-issue. Questions should be answered with research, rather than exploration.

There is this idea of “literacy:” technical literacy, media literacy, math literacy (or numeracy). Literacy does entail shallow information that can be memorized; literacy requires comfort with processing information. Data are building blocks; by and large, we are terrible at using them to build any substantial ideas. For most, these buildings — conclusions and interpretations — are provided, predigested, by the media or parents or friends or teachers.

Instead, imagine a curriculum built around the tool of data visualization. Learning how to massage data, assess its credibility, and transform it to allow for interpretation. Instead of speculating, what if we were to focus on encouraging data-driven discussion? Most people have a pernicious aversion to saying “I don’t know” (or maybe, “It is not possible to know, with the data we have”) when it comes to “everyday” questions like health care, education, racism, economics, etc. Unfortunately, familiarity with a domain often translates into confidence when reasoning about it, despite the fact that these, familiar domains are the most nuanced.