Can I teach myself scientific creativity?

Textbooks and classes are more than just curated collections of established knowledge. They also suggest what your goal should be:

• Read the textbook and go to lecture

• Take notes, ask questions when you feel confused, and try to understand everything

• Solve the assigned problems for which the material equips you

• Do practical projects and labs for hands-on experience These are all instrumental goals, pointed at the terminal goal of becoming familiar with the material. In real-world scientific research, scholarship has a different terminal goal, which is to let you design and carry out useful experiments. Familiarity with the material is just an instrumental goal that is supposed to contribute to that end. Since useful research is about carrying out novel experiments to chip away at unsolved problems, it’s generally going to be harder to say which specific bits of knowledge are going to be most useful to help you become a better experimenter. Perhaps you’ll hit on a great idea for an experiment, or have a novel theoretical insight, and be able to trace back which bits of prior knowledge were necessary and sufficient to achieve this illumination. Unfortunately, now that you’ve had your eureka moment, the connection between that prior knowledge and your creative insight is no longer very useful. The causal connection between your prior knowledge and your eureka moment is a form of knowledge. Unfortunately, it’s a form of knowledge that becomes obsolete as soon as it is created. All you can do with it is bring others to the same eureka moment you just had, which is what a classroom education is for. You can’t use the prior knowledge->eureka moment connection to teach other people how to crank out their own novel insights. Yet I do believe that creativity can be taught. I worked as a piano teacher for a decade before my turn into scientific research. Although I was classically trained, I looked for ways to teach improvisation and songwriting to interested students. Occasionally, I would find two types of mechanical methods for developing their creativity:

• Processes for improvising that force students to make creative decisions, even when they don’t have any clear ideas.

• Tools that let students generate clear ideas, even when they’re not ready to make any creative decisions. For example, when I taught songwriting, I created a songwriting process (verse/​chorus structure, figuring out a chord progression first, then writing lyrics, then singing them while playing the chords). No particular chord progression is better than any other. You just have to choose some arbitrary chords. Any lyrics are fine, and any melody. It would be fine to generate every facet of the song via a random number generator. It’s only important that there be some content. I also discovered a set of tools for when a student just couldn’t come up with any content for the songwriting process. I’d ask them for basic attributes of a character for their lyrics (gender, name, hair color, mood). Then I’d ask them to describe a problem that character had. Finally, we’d do the Five Whys exercise. Why does the character have that problem? Well, why are they in that situation? And why did that happen? Why? Why? Then we’d just write a song that communicated that information in the lyrics. This worked great. Between the basic songwriting process and tools like this, we never had a day when my student couldn’t write at least one song during her lesson. Occasionally, we would engage in something akin to scholarship. This might involve listening to a song from the radio, reading through the lyrics, and discussing what they mean. Or we might learn a bit of music theory. Sometimes, she would learn songs from the radio by ear. I believe that these activities were also helpful and necessary on some level. But I also never tried very hard to directly connect these "scholarly" activities to her creative songwriting, in the sense of "now that you’ve learned X, you can create Y." When I tried, the results were almost always lackluster. Hypothetically, it seems like we could approach scientific creativity in a similar fashion. We’d try to come up with a reliable process for project planning, where literally any experimental procedure is valid. To complement it, we’d want a set of tools to generate more interesting ideas for the project planning process. Scholarship would need to happen alongside the creative side of the work. But analogously to my work with my piano students, I wouldn’t try to force a connection between the book learning and the creative aspects of my work. This seems like a reasonable strategy. I’ll optimize for scholarly learning and experimental creativity separately. The former I’ll approach using the suite of methods I’ve already developed for classroom learning. The latter will require developing some new processes and tools for scientific creativity. I won’t worry too much about trying to force a connection between these two aspects of my relationship with science, except by tailoring my choice of scholarly topics to align with my creative work. Just as with my piano students, it’ll be most important to keep this fun, creative, and sustainable over the long-term.