A couple of other loose ideas

Elaine Reynolds/ People on the thread talk about science as a way of thinking and I agree.  However, in my view science is carried out less like the traditional scientific method we all learned about in high school and more like an exploration. It is often a flawed enterprise.  For example, the current connectome project, which seeks to map structural and functional connections in the brain, is not a hypothesis driven project; it is a collection of data points that can be explored for pattern and purpose. This was also true of the much maligned human genome project.  Bias and interpretation is a huge part of science. The idea of scientists as objective viewers of reality is a farce. Unconscious bias creeps in to perturb objectivity. Results are fit together like a jigsaw puzzle and puzzle pieces rarely fit together neatly.  As I said [at our opening event], my advisor taught me to really dig into the pieces that don’t fit into that puzzle, but I am not sure how common that is.  Data is interpreted, including which data to keep and what analysis to pursue and so on.  I see scientists often searching for support of their own ideas, rather than the truth.  And science is in the midst of a “reproducibility crisis” since many experiments especially in psychology can’t be reproduced. The only way science makes progress is through collective action that swamps out the bias and mistakes. I often say that scientific data leans toward the truth.

As science moves towards trying to understand systems, it uses modeling as a way to understand complexity. So in this way, science becomes like art in the sense of modeling reality rather than exploring reality directly.  If art models reality can science use art to create models of complex systems?

And I think we could talk all day about the role of the rational and emotional in science communication. I think our culture is driven by emotion and fear, so maybe it’s no wonder that scientific thought and evidence are taking a back seat.  Maybe scientists should start using emotional rhetoric rather than data as they talk about climate change. Maybe art could help science describe the world we are becoming.

13 Replies to “A couple of other loose ideas”

  1. Elaine, I find your take on things particularly exciting, because as a scientist you see so clearly (and have articulated so strongly) where science has reached a wall, and where it could really use art’s help. For our readers here, I wonder if you could say a few words about the problem of reductionism — about how that tradition within science, as successful as it has been, has proved inadequate when it comes to the problem of consciousness (or emergence in general). I know a lot of artists are interested in the phenomenon of emergence, and you seem to suggest that we’re uniquely qualified to help in this regard. So far, however, much of the work I see that engages the subject contents itself with the visualization of emergent properties (typically in abstract, pattern-oriented work), and I find myself wondering if there’s something we can pursue *beyond* visualization. Your talk on Saturday piqued the interest of many; it might be nice to share some of that with our reading audience.

    1. Thanks Taney for the opportunity to speak on Sat and for facilitating the engaging conversation of that evening and here online. I believe science has been dissecting reality for so long that most scientists are not trained in thinking about systems as a whole. It’s as if they have taken the car apart and can’t put it back together. They can’t figure out how the parts create the mechanical object, and the problem of how the car becomes more than a means of transport is beyond scientific thought. We know how individual neurons work and all the molecular details of their function and we are currently characterizing how all the neurons in the brain are connected to each other. We are beginning to understand how circuits form the informational basis of processing in the brain. But we have a significant understanding gap when it comes to how this processing leads to behavior or cognitive functions that we observe at the systems level. I believe partnerships between artists and scientists could help tackle these problems because they have different training and perspectives. An example of an artist who I believe is making a significant contribution in this way is Ellen Levy (a recent interview is here http://www.artcritical.com/2017/05/30/joyce-beckenstein-with-ellen-k-levy/). Her work on complex systems creates analogies and juxtapositions that are useful in thinking about the neurological processes like attention and information flow. This is more than just using science as an inspiration for work or the output of processes, but imagine how the processes themselves might work.

  2. A few random questions and thoughts.

    When the focus of the art shifts to the output of processes, is it still art? Is it an example of visualization, factual or abstracted?

    Could we consider the realization of a work of art and/or the modeling of current data at a “resting point”? Data is ever changing and many artists consider their work at a stopping point to be intervened at a later time while others may deem it a finished piece.

    1. I am not a person who is comfortable defining art. I think the collection of data hits a resting point much like an piece does. But I think science never feels finished to me. Do artists finish exploring an idea in a piece and then go on to explore an idea further in another piece? That feels like what I do in science.

      1. Good points, Lorrie and Elaine! It’s a cliche, but for me, both science and art are more about processes than outcomes. In art, every piece spawns another. In science, every answer leads to another question.

  3. Thank you, Elaine! I will respond more fully to your intriguing ideas in a separate post, but as a footnote here, I just want to bring up a minor quibble. On the subject of the reproducibility crisis and bias, I agree that science is practiced by humans, and that humans are flawed. But this is not a static condition, and there are many subcultures in science. In my own field of experimental particle physics, we have (in the last two decades) become hyper-aware of our own biases, and we now take blind analyses extremely seriously. We have also adopted a very high bar for claiming discoveries — the so-called 5-sigma threshold, which means the chance of an observation being caused by ordinary processes (the so-called p-value) must be less than 1 in *3.5 million*. In contrast, the published results that spurred the reproducibility crisis only had p-values of less than 1 in *20* (i.e. they were 175000 times more likely than 5-sigma result to be later refuted, which they inevitably were). My point is that not all of science is going through a reproducibility crisis, and it seems to me that steps are being taken to address the problems.

    1. Your quibble is acknowledged and accepted. I did not know about the steps that physics has been taking to improve the rigor of that discipline. As you said, the difference in probabilities that are being accepted by your field and mine are astounding on their own. Rarely do biologist or neuroscientists do tests under blind conditions. I also find that design of experiments and application of statistics is problematic in biomedical science. Interestingly, I use papers published in top journal as exercises for my students to use in critical analysis. But it’s good to hear that this problem doesn’t occur in all fields of science. I guess my main point here was to point out that science is not a rigid set of practices. I felt like some commenters were contrasting science as a discipline filled with reason and I wanted to point out the flaws and interpretation in the practice of science. Tell me if I am wrong but in physics as well as other sciences, one experiment or paper rarely shows anything on its own. Findings have to amassed and be confirmed. And this helps sort through any flaws in individual experiments.

      1. Thanks, Elaine! Yes, statistics is a dry but necessary subject… And yes, science is not rigid at all. You’re right that the right answer is sometimes slow to emerge (for example, take the pentaquark confusion: https://www.symmetrymagazine.org/article/september-2006/the-rise-and-fall-of-the-pentaquark). But also sometimes not (like the recent observations of gravity waves). Both of these are unusual cases (on opposite ends of the spectrum) in physics.

  4. Thanks all for providing so many points of clarity. I liked reading about Ellen Levy’s work and her relationship to the work of individual scientists. Elaine, it sounds like your view is that artists and scientists can both inspire each other and interpret what they are learning from the other. Question: should science with its quest for precise statements (least ambiguity) and art with its penchant for layered meaning (including ambivalence) change their respective modes of communication? Lorrie’s question about visualization and Werner’s reminder that each is a process makes me wonder about the different steps in reporting results: artists exhibit a body of work and scientists write papers. Each must satisfy their field’s respective gatekeepers: gallery curators and peer review editors respectively. Do these institutions need to change in order to promote the possibility of more dialogue? And if so, does anyone have suggestions about how?

    1. A provocative comment, Chris. I think the markets would have to shift drastically. In academia, there has been an upsurge in interdisciplinary initiatives. But these initiatives tend to insist on superficial “deliverables”. There is meager incentive for (and high barriers to) engaging in the unseen hard work of honestly exchanging ideas. Changing that reward structure would take (institutional) courage to fund the process even when there are no tangible outcomes.

  5. I would like to stress the idea that the fact of interpretation in Science is particularly relevant to this discussion. In order for a person to ‘notice’ something it has to be perceived as salient from the background of all other things being sensed. Scientists’ biases (their previous knowledge, their methods, etc.) render them able to notice just a fraction of all that could be noticed. Artists, with their different set of expectations may be able to perceive and express another fraction. The collaboration of Art and Science, in part, could take the form of a reciprocal re-focusing of attention if we let it. That is, if a vigorous field of Art/Science practitioners reaches a critical mass.

    1. This is an excellent point, Luis. When analyzing data, scientists must decide what is signal and what is noise. The signal is the star of any published result. In contrast, artists might be able to find beauty in the noise.

      More generally, the act of measuring any object must promote some of its qualities over others — if one only measures the height of a box, one ignores its length and width, color and mass, material and…. Hence postmodernism’s rejection of reductionism. But this winnowing of reality occurs even with information we gather involuntarily through our senses — sensing is measuring, after all. So, one can never completely know an object through measurements, but one can imagine it through art.

    2. This is an excellent point, Luis. When analyzing data, scientists must decide what is signal and what is noise. The signal is the star of any published result. In contrast, artists might be able to find beauty in the noise.

      More generally, the act of measuring any object must elevate one of its qualities over all others — if one only measures the height of a box, one ignores its length and width, color and mass, material and…. Hence postmodernism’s rejection of reductionism. But this winnowing of reality occurs even with information we gather involuntarily through our senses — sensing is measuring, after all. So, one can never completely know an object through measurements, but one can imagine it through art.

Leave a Reply

Your email address will not be published. Required fields are marked *