ANA (ANA is Not Aaron) is a computer program I am developing to draw and paint autonomously. The eventual goal is for ANA to be able to accept some basic parameters and to create original drawings and paintings within some genre and with a defined set of subject matter possibilities, but without any other direction of what exactly to draw or paint.
TBR Research presents insights and excerpts from peer-reviewed scholarship.
At present, ANA is able to use a palette of 40 colors based on the characteristics of artist’s oil paint colors. However, the colors are not just the color (the characteristic artists call hue) — I have constructed a database to define each of the colors for use by ANA such that each of them also approximates the transparency/opacity and subjective color temperature of real oil paint color as it would be used by an artist using conventional materials. This allows ANA to do unique things. For example, when ANA uses Prussian blue (a potent pigment that is also a very transparent and warm blue) it “behaves” differently than does Cerulean blue (a cool and very opaque pigment) when applied alone or mixed with other colors. In my tests with ANA, these and the other colors in ANA’s palette have similar effects to what an artist would expect if mixing real oil paints. ANA isn’t fully autonomous at this point. I still have to make programming tweaks by hand to get ANA to make certain types of variations in the images it produces.
Making the Familiar Strange
The author’s “Making the Familiar Strange” exhibition opens April 18, 2016 at the SUB Gallery on the Boise State University campus and runs through May 22. It includes art generated by a software program named ANA that Winiecki is writing as “an art-making device and a platform for researching both the physical and intellectual process of making art” as he lives it. A small grant from Boise City Department of Arts & History will pay for the materials and professional quality printing of images for the exhibition.
This is the first of three articles that will appear in The Blue Review to describe the ANA project. In this first article, Winiecki gives some background on the project and its components. The second article will describe the software itself in more detail. The third article will describe aspects of the philosophical component of the research that underpins the ANA project.
For example, the middle area of the painting displayed above (“Meanwhile, in Iceland 2”) has Cerulean blue (the light blue) as a base and Prussian blue (the darker, purplish blue) applied irregularly into it. The transparency of Prussian blue prevents it from covering the Cerulean blue but the Prussian blue alters the underlying color by making it slightly darker and warmer in its appearance in areas across the middle part of the painting. The addition of transparent Prussian blue produces what might be described as a sort of “dreamy” or “washy” and warming effect on the Cerulean blue.
When viewing this painting in real life, the Prussian blue is noticeably translucent against the Cerulean blue. Near the top of the painting, ANA has used Yellow ochre (a very dense and very opaque pigment). The opacity of Yellow ochre contrasts with Prussian blue and Cerulean blue to approximate my own aesthetic toward color. Procedurally, Yellow ochre was added before the blues. The transparency of the blues on top of this opaque color is apparent here: where transparent Prussian blue is applied over Yellow ochre, the color turns to an earthy green, which is just what an oil painter would expect when a transparent blue like Prussian blue is applied in a glaze-like manner over an opaque yellow like Yellow ochre.
Designing and programming ANA to include relevant elements of what I know about painting, and my aesthetics toward things like color has been one of the main focal points of the work up to this point. I’d like to note that while the output of ANA (the paintings) are the visible product of the program, it is not always what I consider to be the most important part of the project.
Aside from producing images that can be printed for display, a main part of the ANA project has involved an ongoing effort to translate what I know about drawing and painting manually and my own personal aesthetics for color, form, subject matter and other things, such that they can be performed or demonstrated by the computer programming code. The idea here is to build a program that approximates what I do when I draw and paint by hand, with physical brushes and paint. My goal in this effort is not to produce an artificially-intelligent program, but to learn more about my own ways of knowing and understanding painting and its components.
A SERPENTINE PATH: PROGRAMMING A COMPUTER TO DRAW AND PAINT?
For as long as I can remember, I have had aptitude at drawing more-or-less realistically by hand. I know this is not especially unique and I make no claim to be a master or a great technician, though in my young adult life and before the introduction of personal computers, I actually made a living as a technical illustrator — a mix between a draftsperson and a commercial artist. Most of my work in that field was focused on creating technical documents and advertising illustrations to help laypersons understand the technical workings of mechanical equipment. I did work for major American corporations, all branches of the U.S. armed services and on one occasion even developed animations for NASA Skylab. I also freelanced medical illustrations to accompany journal articles and textbook instruction on a particular type of cardiac surgery for a faculty member in the medical school of a local university.
Notes for technophiles or those otherwise curious:
- ANA is written in the programming language Common Lisp on computers running the Ubuntu Linux operating system. The cl-cairo2 library is used to create the graphic output to PDF. Common Lisp is the second-oldest programming language (FORTRAN is the oldest) and is frequently used for research in the field of artificial intelligence.
- Programmers like to argue or tease each other about the tools they use to create software. I use EMACS as my tool for writing programming code because it is especially friendly and supportive of some of the unique features of Common Lisp. Within EMACS, I use SLIME and Quicklisp in order to keep everything running with minimal drama.
- ANA stands for “ANA is Not Aaron,” an homage and reference to a computer program written by artist Harold Cohen. Aaron is the first computer program to be able to draw and paint autonomously — albeit in different ways than ANA. Cohen has been writing Aaron since the 1970s and the scope of Aaron’s capability far exceeds anything ANA is presently capable of.
- Cohen has written about Aaron and his work and he and Aaron have been the subject of many articles, chapters and books (for example; Cohen, Cohen, & Nii, 1984; McCorduck, 1991), but Cohen has never shown the Aaron code to anyone. He has always maintained that Aaron is the product of how he thinks about drawing and painting, and that if another artist is interested in using programming to draw and paint, then that other artist should build a system around how he or she thinks about drawing and painting. (I agree with him on this point, and as described in more detail below this is what I have tried to do with ANA.) He been very gracious in talking with me about his work, and has offered advice when I have asked.
I left that field of work in the late 1980s when computer software started to become capable of producing accurate technical illustrations based only on computer-aided drafting (CAD). In effect, computer software replaced my innate and learned skill as a commercially-viable profession. Perhaps surprisingly, I found that computers did not totally replace what I had done. In the year 2005, a coworker from one of my many previous jobs emailed me and attached a technical document he was rewriting, in which a set of illustrations I made in 1985 were to be used. He said the company CEO insisted that they reuse my drawings because “…nobody has made anything better, since…” Nice to hear, though I’m not going back!
While I could have joined the crowd and learned how to create CAD drawings, I chose not to do this because — for me — a major part of the challenge and enjoyment of interpreting blueprints (or photos and sketches I made in the field) and then translating them into illustrations that allowed a non-expert to understand very technical processes was the intellectual and physical work. When software started to replace that process, I went back to graduate school. Oddly enough, that’s where I first encountered personal computers and where I acquired some knowledge and skill for writing software.
From that point until 2008, I paid almost no attention to art or drawing or anything of the sort. I was focused on graduate school and then starting in 1996, on building my career as a faculty member at Boise State University.
At some point in the middle part of the 2000s I started to look back at my life and what I had let slip away. Drawing and painting was near the top of the list. At about that time I noticed an advert in an internal Boise State newsletter that the Art Department was holding an open advisement session for individuals interested in art courses. I dug out some of my sketchbooks from the 1980s and went to the advisement session, curious about drawing and painting courses.
In a still-important stroke of luck, the faculty member who was assigned to this open-advisement session was Dan Scott, then a new assistant professor of painting and drawing. I described my interest and showed him some of my sketchbooks from the 1980s. To my surprise he suggested that I enroll in his entry-level oil-painting course for the coming semester. Since then, Dan has been an important touchstone for me in art, aesthetics in general, and based on my experience as a student in his courses, even as an influence in my own teaching. Below is my third painting from that first class — the third piece of intentional art I had produced since 1989.
Perhaps it doesn’t show, but the most challenging and enjoyable part of this painting — and most other paintings I have done since — is the shadows. I found that shadows are the home of what I consider the most seductive part of a painting, containing subtle and beautiful color blending. They are also an example of something I think I can best describe as an abstraction — a suggestion of some idea or thing, but not that idea or thing itself — that is nonetheless firmly grounded in reality.
I completed all of the undergraduate painting and drawing courses offered by the Boise State Art Department and during that time went from a total infatuation with making realistic images to a more introspective orientation toward finding ways to communicate what I thought were important critical messages through visual art, and the abstract use of color and form became a major part of that. Many art viewers may be surprised to learn that abstractions are almost always much more than subjective expressions of emotion or aesthetics, and that they are often a reflection of very intricate intellectual ideas and theories that rest on reality. This aspect of abstraction became very interesting to me as an aspiring painter.
This period of my progress in learning how to produce art culminated in 2013 with an exhibition of my paintings and drawings at the Boise State Student Union Gallery — a show titled “Heterotropias” in which nearly 50 of my paintings and drawings explored and illustrated some of the many ways we, in society, create spaces and places that have multiple meanings to us, and which assert multiple forms of subjectivity upon us.
This more intellectual turn to communicating things through art took its toll. I had no intention to become a full-time artist and the time and intellectual commitment to continuing on this path was exhausting. Dedicated and full-time artists are indeed serious intellectuals who deserve much morethan they typically receive.
TOWARD A PHOENOMENOLOGY OF ART AND WORK
But I wasn’t ready to step away from art again. I started looking for ways to blend the creation of visual art with my work as a faculty member in a department focused on teaching and supporting students whose professional goals are to help individuals and organizations improve workplace performance — a job that requires a continuous effort to learn new ways of studying and new ways of understanding the world of work and the ways work is changing in the 21st century.
I challenge my students to make use of a small number of models, theories and loose collections of psychology, sociology and anthropology, and other subjects to help organizations improve worker performance and organizational effectiveness without reducing work and workers to robot factories and humans-as-robots. To help them learn how to accomplish this, I have found it effective to immerse them into real world projects and help them learn to apply the available tools and theories, but especially to sort out the messes that they find and learn how to learn, as they go.
This process is akin to the philosophical orientation known as phenomenology.
With respect to art, it just so happened that I was also beginning to develop an interest in focusing inwardly to analyze what it is that I am paying attention to (with the subject and content of a painting, and the paint, and the brush, and the canvas, and the palette, and…) when I create a painting — how I translate “the world” into a painting that is really only an abstraction of aspects of that world.
I imagined that if I could capture that translation process with respect to what I do in painting and drawing, I could develop an intellectual tool-set to help my students make a similar journey in their own learning and eventual professional practice.
The philosophical orientation of phenomenology is just this — learning how to make accessible one’s own pre-conscious experience in doing something that allows one to be an active member of the part of the world in which he or she is operating, and then communicating that in a manner that describes how it has come to occur.
My principal example of this has been David Sudnow’s book Ways of the Hand, in which he — an academic sociologist — learned to be a largely-self-taught player of improvisational jazz piano. Sudnow made oblique references to the philosophical work of Edmond Husserl, but his descriptions seemed to me to be more consistent with the way a student of Husserl’s, named Martin Heidegger, approached phenomenology and those who have extended Heidegger’s work to include the participation of non-human actors in technological processes — people including Bruno Latour, Don Ihde and Graham Harmon — stuff that has sometimes been called post-phenomenology.
Odd as this sounds, this is conceptually not so difficult, but tricky in practice. This phenomenological part of the ANA project this will be the focus of the third article in this series.
But back to the issue of how I came to incorporate writing software as a process of (a) making art and (b) making that composing/coding process the intellectual focus of the project.
In my many years of schooling and professional practice, I have learned that one of the more effective ways for me to learn something new and to develop fluency in that something, is to introduce some sort of device that prevents me from making easy use of my current knowledge and skills and instead forces me to distance myself from things-I-can-do-without-thinking. This forces me into a situation where I have to learn new ways of thinking and using what I know and can do. In social science this has been called “making the familiar strange,” as Harvey Sacks put it in his 1963 essay, “Sociological Description.” With this connection, I began looking for ways to “strange-ify” drawing and painting so I would be forced to find a new way of thinking about it, rather than simply channeling through my innate skill. I sought a process that would allow me to continue to be connected to art and art-making, but in a different way than I had attempted in the past, and in a way that I could bring back to my courses.
And this is where I remembered the work of artist Harold Cohen and Aaron. I had first become aware of Cohen’s work in the mid 1990s when my graduate school research interests leaned toward the use of graphics as an aid to learning and guiding workplace performance. Searches through related literature hit upon some of the articles he had written up to that point.
Cohen had achieved high status in the 1960s for his conventionally-produced abstract painting. In the 1970s he had found his way to San Diego and the University of California, and as a side effect of acquaintances with computer scientists at Berkeley, started developing software he named Aaron, that produced abstract drawings and paintings.
I started to wonder if I could follow Cohen’s lead and “strange-ify” painting and drawing by trying to accomplish it through software code. If he could draw and paint with his own programming — I thought — so could I.
At that point I retrieved and read all of Cohen’s published writing and what I could find in secondary sources written about him and Aaron. While the articles I read never reported it, just from looking at his art and the way he described his process, I reasoned that he must be using Common Lisp — a programming language I had tinkered with in graduate school.
The next nine months was spent learning enough of Common Lisp to begin writing a program that could aspire to approach my goals, and reading up on how to do and report phenomenological research. I made lots of fieldnotes to try and describe the process of becoming familiar and then nominally competent at programming in Common Lisp.
The first images created by ANA were made in 2013 and development has continued (on and off) since. Several academic articles are — as they say — “in review” at peer reviewed journals as I write this, and more are underway. I use what I am learning about learning-as-you-go in my own teaching. It’s all coming together.
Here is the very first printable image created in my efforts — simple color, lines, a square, and a capital “A” signature for the program I had already named ANA (for “ANA is Not Aaron”).
As is I hope to be apparent after comparing this image with the one that appears at the start of this article, there has been substantive developmental progress toward my stated goal!
Well beyond what is shown immediately above, ANA is now capable of producing a range of abstractions that often appear as skyscapes. These vary from placid blue skies with fluffy clouds, to irreal, other-worldly scenes, to ominous-looking atmospheres with arrays of color that nonetheless still evoke a sky-like affect. Here are a few more examples of ANA’s work that may appear in the exhibition “Making the Familiar Strange” in April-May 2016.
The title of each painting includes the date (down to the second!) it was created, along with a name that I assigned after inspection. Perhaps curiously, the painting named “X-ray Clouds” was actually something I considered at the time to be a throwaway example of an algorithm that didn’t do what I expected. I have since come to pursue the creation of images that demonstrate unexpected features. I do this by experimenting with what programmers call “edge cases” and “corner cases” where I force ANA to run with inputs that are near to or exceed the boundaries of what the program is supposed to be able to deal with. In the next article in this series about the ANA project, I will describe in more detail the software that comprises ANA —including some of the code itself — how paintings are converted from digital files to artifacts that can hang on walls, and a little bit about the things I am learning about programming an artmaking machine that uses my own knowledge, skill and aesthetics as a painter.
This process of experimenting or tinkering (what an anthropologist would call bricolage) is an essential component of my work with and on ANA. In a very real sense it is a way of having a conversation with ANA, the code that comprises it (and which I am writing to add features), and the algorithms I am writing in an attempt to capture what I know and can do with real art materials and convert those things to the formal language of Common Lisp. As noted above, this conversation with things in the world is a big part of what makes phenomenology possible.
The views and opinions expressed here are those of the writer and do not necessarily reflect those of Boise State University or the School of Public Service.