In Parts I and II of Gestalt Principles, Bang Wong describes core elements of Gestalt psychology, a 1920s German theory of “how people organize visual information.” The German term Gestalt means shape or form. As Wong summarizes in Part II, “our visual system attempts to structure what we see into patterns to make sense of information.” In other words, we naturally and reflexively process visual input by attempting to group objects into “unified wholes.”

In Part I, Wong explores the principles of similarity, proximity, connection and enclosure. “The fundamental concept behind these principles is grouping;” he argues. “We tend to perceive objects that look alike, are placed close together, connected by lines or enclosed in a common space as belonging together.” Color schemes, visual clustering, and lines on a graph are all tools which can differentiate datasets.

In Part II, he examines the principles of visual completion and continuity:  “Because we have a strong tendency to see shapes as continuous to the greatest degree possible, we fill in voids with visual cues found elsewhere on the page.” This principle has an important implication: “every element on a page affects how we perceive every other element.”

Wong presents all these principles as helpful design tools which can leverage human mental processing in order to present data clearly.

What’s missing from these short essays, however, is any discussion of possible misuse of these design principles. Presumably, an altruistic designer would solely use these tools to “let the data speak for itself;” using Gestalt principles to highlight and clarify the ground truth which is already there.

But this seems to gloss over an important detail: all design choices are choices. Even putting aside the occasional malicious designer, who deliberately presents a warped visualization in order to leave viewers with an erroneous impression; it seems entirely possible that a lazy designer could accidentally imply something unintended, or that a researcher could be mislead by the Gestalt of their own visualization.

Furthermore, while these principles may be the simplest way to communicate data, there is no discussion of whether they are the right way to communicate data.

Last semester, Lauren Klein of Georgia Tech gave a talk at Northeastern in which she highlighted the visualization work of Elizabeth Peabody. Remembered primarily as an educator, Peabody created of elaborate mural charts of history, intended to provide historic “outlines to the eye.” Her work was intentionally complex and difficult to engage with; people had to interact with it to understand it. In the mid-1800s, this approach pushed the question who is authorized to produce knowledge? And subversively answered: everyone.

So Gestalt principles may make it easier to process information, but it should also be acknowledged that this may diminish the agency of the viewer – whose brain reflexively interprets visual stimuli in a given way, even if it’s not accurate and even if they know it’s not accurate.

At the beginning of the two articles, Wong quotes founding Gestalt scholar Kurt Koffka, in saying “The whole is ‘other’ than the sum of its parts.” While this is sometimes translated as “greater than the sum of its parts,” Wong is clear that this was not Koffka’s meaning: “the emergent entity is ‘other’ (not greater or lesser) than the sum of the parts.”

This quote highlights the need to think more robustly of the experience of the viewer. The design that is created, the visualization that expresses some aspect of the data, is a new thing, other than what existed before. Peabody’s visualizations were exhaustingly interactive, but they did invite the viewer to become an active participant in the act of creating this other.