The revision of vision

I have persistently written about the limitations of vision and how language can affect our perceptions. In “Why naming the ‘thing’ can be a problem” I pointed out how language defines our visual interpretations. Another post, “Intuition is just another form of pattern recognition”, was also about the limitations of language description and the importance of trying to find new patterns by defining information in a different way.

When I posted “Shades of Truth” at the beginning of August this year, I was reminded, yet again, of the similarities to an essay I came across in the Disney Imagineering Library in Glendale, CA many years ago.  The essay was “The Revision of Vision”  by S.I. Hayakawa (1906-1992), a linguist, psychologist, and teacher. It was written as an introduction to the book "Language of Vision" by Gyorgy Kepes. There are many writings that have impacted my interpretations of painting and art, and this essay is certainly at the top of my list.

The Revision of Vision

by S.I. Hayakawa, Illinois Institute of Technology 

Introductory essay for "Language of Vision" by Gyorgy Kepes, originally published in 1944

Whatever may be the language one happens to inherit, it is at once a tool and a trap. It is a tool because with it we order our experience, matching the data abstracted from the flux about us with linguistic units: words, phrases, sentences. What is true of verbal languages is also true of visual "languages": we match the data from the flux of visual experience with image-clichés, with stereotypes of one kind or another, according to the way we have been taught to see.

And having matched the data of experience with our abstractions, visual or verbal, we manipulate those abstractions, with or without further reference to the data, and make systems with them. Those systems of abstractions, artifacts of the mind, when verbal, we call "explanations," or "philosophies"; when visual, we call them our "picture of the world."

With these little systems in our heads we look upon the dynamism of the events around us, and we find, or persuade ourselves that we find, correspondences between the pictures inside our heads and the world without. Believing those correspondences to be real, we feel at home in what we regard as a "known" world.

languages select . . . they leave out what they do not select.

In saying why our abstractions, verbal or visual, are a tool, I have already intimated why they are also a trap. If the abstractions, the words, the phrases, the sentences, the visual clichés, the interpretative stereotypes, that we have inherited from our cultural environment are adequate to their task, no problem is presented. But like other instruments, languages select, and in selecting what they select, they leave out what they do not select. The thermometer, which speaks one kind of limited language, knows nothing of weight. If only temperature matters and weight does not, what the thermometer "says" is adequate. But if weight, or color, or odor, or factors other than temperature matter, then those factors that the thermometer cannot speak about are the teeth of the trap. Every language, like the language of the thermometer, leaves work undone for other languages to do.

Visually, the majority of us are still "object-minded" and not "relation-minded". 

. . . Revisions of language are needed. Every day we are, all of us, as persons, as groups, as societies, caught in the teeth of what the older languages leave completely out of account. We talk of a new, shrunken, interdependent world in the primitive smoke-signals of "nationality," "race" and "sovereignty". We talk of the problems of an age of international cartels and patent monopolies in the economic baby-talk of Poor Richard's Almanack. We attempt to visualize the eventfulness of a universe that is an electro-dynamic plenum in the representational clichés evolved at a time when statically-conceived, isolable "objects" were regarded as occupying positions in an empty and absolute "space". Visually, the majority of us are still "object-minded" and not "relation-minded". We are the prisoners of ancient orientations imbedded in the languages we have inherited.

The language of vision determines, perhaps even more subtly and thoroughly than verbal language, the structure of consciousness. To see in limited modes of vision is not to see at all - to be bounded by the narrowest parochialisms of feeling.

. . . Purposely depriving us of the easy comfort of all aesthetic stereotypes and interpretative clichés, Mr. Kepes would have us experience vision as vision. (His) endeavor may perhaps best be characterized by the following analogy. To a Chinese scholar, the pleasure to be derived from an inscription is only partly due to the sentiments it may express. He may take delight in the calligraphy even when the inscription is meaningless to him as text. Suppose now a singularly obtuse Chinese scholar existed who was solely preoccupied with the literary or moral content of inscriptions, and totally blind to their calligraphy, How would one ever get him to see the calligraphic qualities of an inscription, if he persisted, every time the inscription was brought up for examination, in discussing its literary content, it accuracy or inaccuracy as statement of fact, his approval or disapproval of its moral injunctions?

Something of the quality of a child's delight in playing with colors and shapes has to be restored to us before we learn to see again . . .

It is just such a problem that faces the contemporary artist, confronted with a public to whom the literary, sentimental, moral, etc., content of art is art - to whom visual experience as such is an almost completely ignored dimension. . . . We have all been taught, in looking at pictures, to look for too much. Something of the quality of a child's delight in playing with colors and shapes has to be restored to us before we learn to see again, before we unlearn the terms in which we ordinarily see.

...How we deal with reality is determined at the moment of impact by the way in which we grasp it. Vision shares with speech the distinction of being the most important of the means by which we apprehend reality.

When we structuralize the primary impacts of experience differently,we shall structuralize the world differently.

To cease looking at things atomistically in visual experience and to see relatedness means, among other things, to lose in our social experience... the deluded self-importance of absolute "individualism" in favor of social relatedness and interdependence. When we structuralize the primary impacts of experience differently, we shall structuralize the world differently.

The reorganization of our visual habits so that we perceive not  isolated "things" in "space" but structure, order, and the relatedness of events in space-time, is perhaps the most profound kind of revolution possible - a revolution that is long overdue not only in art, but in all our experience.

Shades of truth

A while back I posted an article titled “Do facts matter or is truth just another possibility?”. I wrote about our outdated, misleading primary color system and the confusion it can cause. I also mentioned the inaccuracies of the most commonly used world map (Mercator) and how its distortions affect our perceptions. In a follow-up article “A win for visual truth” I covered a new design for a world map, called the Authagraph Map – it looks strange, but is far more accurate than the Mercator map. I don’t know that we will soon be using a more accurate color primary system, or a more accurate world map, but I do hope that we can learn new information and resist the temptation to treat our version of reality as some kind of worn-out shoe that we keep around just because it’s comfortable.

Now along comes an article from Wired.com – another great example of misleading perceptions, how more information is better information, and how nuance can be as important or even more important than simplicity.

Is the US Leaning Red or Blue? It All Depends on Your Map

by Issie Lapowsky, Wired.com
For the complete article click here.

On May 11, 2017, a reporter named Trey Yingst, who covers the White House for the conservative news network OANN, tweeted a photo of a framed map of the United States being carried into the West Wing. The map depicted the 2016 election results county-by-county, as a blanket of red, marked with flecks of blue and peachy pink along the West Coast and a thin snake of blue extending from the northeast to Louisiana. It was a portrait of the country on election night, but on Twitter, it was also a Rorschach test.

Conservatives replying to Yingst's tweet interpreted the expanse of red as proof of their party's dominance throughout all levels of government. Liberals viewed the map as a distortion, masking the fact that most of that redness covers sparsely populated land, with relatively few voters.

In reality, both sides are right, says Ken Field. A self-proclaimed "cartonerd," Field is a product engineer at the mapping software company Esri and author of a guidebook for mapmakers called Cartography. The problem, he says, isn't with people's partisan interpretation of the map. The problem is believing that any single map can ever tell the whole story. "People see maps of any type, and particularly election maps, as the result, the outcome, but there are so many different types of maps available that can portray results in shades of the truth," Field says. "It’s a question of the level of detail that people are interested in understanding."

It stands to reason that President Trump would want that particular map hung in the West Wing. There is an awful lot of red on it. But focusing on that map alone could lead Republicans to overestimate their advantage, and lead Democrats to misunderstand the best ways to catch up. That's one reason why Field recently published an extensive gallery of more than 30 alternative maps designed to tell markedly different stories about what happened on election night 2016. 

"All of these maps show different versions of the truth," he says. "None are right, and none are wrong, but they all allow you to interpret the results differently."

Take the map Yingst shared, for example. In the language of mapmakers, it’s a “choropleth diverging hue map.” The term “choropleth” refers to maps that use color or shading to visualize a given measurement. In this case, the map uses either the color red or blue to indicate which party won a given county. It’s accurate, and it’s familiar. These colored county-level or state-level maps are some of the most commonly used to illustrate the results of an election. But, Field says, they also lack nuance. There’s nothing on that map to indicate to the viewer, for instance, that fewer votes were cast in the rural mountainous regions of Montana than in Manhattan.

Understanding that nuance—or lack thereof—is key heading into the 2018 midterms, when amateur cartographers will no doubt climb out of Twitter’s recesses to proclaim their definitive readings of electoral maps. Here’s what we can learn from just a few of Field’s examples:

The Pointillism Approach

Presidential election 2016: dasymetric dot density KEN FIELD

To Field, there's no such thing as a totally comprehensive map, but he says, "Some are more truthful than others." The so-called dasymetric dot density map is one of them. The term “dasymetric” refers to a map that accounts for population density in a given area. Instead of filling an entire state or county with the color red or blue to indicate which party won, Field uses red and blue dots to represent every vote that was cast. On this particular map from 2016, there are roughly 135 million dots. Then, rather than distributing the dots evenly around a county, he distributes them proportionally according to where people actually live, based on the US government's National Land Cover Database. That’s to avoid placing lots of dots in, say, the middle of a forest, and to account for dense population in cities.

Taken together, Field says, these methods offer a far more detailed illustration of voter turnout than, say, the map in Yingst’s tweet. That map uses different shades of red and blue to indicate whether candidates won by a wide or slim margin. But by completely coloring in all the counties, it gives counties where only a few hundred votes were cast the same visual weight as counties where hundreds of thousands of votes were cast. So, the map looks red. But on the dasymetric dot density map, it’s the blue that stands out, conveying the difference between the popular vote, which Clinton won, and the electoral college vote, which Trump won.

Shades of Red and Blue

Presidential election 2016: Value-by-alpha KEN FIELD

The value-by-alpha map is similar to the dasymetric dot density map, and in some ways, even simpler. It doesn’t account for where votes were most likely cast within a county. Instead, it uses color to indicate the party’s vote share in each county, and opacity (in mapmaking, it’s called the “alpha channel,” hence, value-by-alpha) to indicate the population in a given area of the county. A bright, vibrant blue indicates a high Democratic vote share in a densely populated area. A light pink indicates a high Republican vote share in a sparsely populated area. Purples portray areas where one party or another won by a narrow margin.

What you notice first when you look at the map is that the solid red wall extending from North Dakota to Texas on the map Yingst shared is almost white in this rendering. What you notice second is just how much purple there is everywhere else. It’s a good reminder of what people often forget about the 2016 election: “It was very close,” Field says. President Trump won Michigan, Wisconsin, and Pennsylvania, the three states that clinched his victory, by about one percentage point or less.

The View from Above

See link to article above for the complete interactive map.

What Field likes most about the 3D prism map is how people react to it. “It’s just cool. People like 3D stuff,” he says. But it also illustrates an important point. Counties are colored red or blue, based on which party won, but the vote totals are portrayed in three dimensions, where the height is equal to the number of votes cast for the winning party. Because Clinton predominantly won big cities, where more votes are cast, it creates a map that looks a bit like a city itself, with dozens of mile-high blue skyscrapers jutting out from between red row-homes and strip malls.

Click around the map and you’ll see that viewed from above, it looks not unlike Trump's map—all in red. But click to tilt the map and it’s mostly blue spikes. It demonstrates perhaps more effectively than any of the other maps how President Trump won in 2016, Field says. “You had a Republican who was very successful in getting the smaller areas to vote Republican, while the larger populated major cities went Democrat,” he says.