William R J Harrison

The possibility and probability of image duality when viewing contemporary art. Specifically seeing landscapes within paintings; how AI vision could redefine the fundamental processes for creating a landscape painting.

​

Defining landscape, the sublime, and some broader views on our differences when reading images.

​

Primarily I must define the landscape: an area of land holding physical elements or defined landforms which can included vegetation and/or human elements present within. Anne Whiston Spirn said ‘experiences of partnership with the land’ and its ‘physical shaping,’ and she noted that ‘land means both the physical features of a place and its population.’ Landscapes also hold values beyond the physical and visual, effecting our emotions having a greater presence through an aura. Largely Referenced within this area is the sublime. Therefore, I must also define this idea; Most dictionary definitions of the sublime still hold outdated ideas around godly presence or other worldly phenomenon. However, I will be using the more philosophical ideas around the sublime as a feeling. Burke wrote, ‘The passion caused by the great and sublime in nature, when those causes operate most powerfully is astonishment, and astonishment is that state of the soul in which all its motions are suspended, with some degree of horror.’ Although Burke seems to always reference a sense of fear or horror, I see the sublime as an many emotions or feelings that can be accessed in an environment which is specific to each of us; but some of the more common ones are being greatly overwhelmed, feeling insignificant in scale or fear as Burke suggests. I believe therefore the sublime is commonly linked with landscapes and nature as these places can be common environments where these feelings are fulfilled. Within ‘Modernism and the Sublime’, Philip Shaw when talking on abstract expressionism and the sublime states, ‘sublime is on the side of the mind rather than nature’. Although Environmental factors seem to be the catalysts for sensing the sublime, effectively it is a feeling or emotion that is very human and thus this feeling should be possible to achieve while viewing imagery. So why and how do we view or read images differently, James Elkins states within ‘The object stares back: on the nature of seeing’ that ‘Seeing is metamorphosis, not mechanism.’ Suggesting that humans learn to see in different ways based on past experiences or current feelings. using the word metamorphosis showing seeing is subject to change and not purely a mechanical process of the eye viewing the object, but the brains interpretation of it can be the major factor. This has been tested in the past within psychology with the Rorschach Test. Where the viewer is subject to ink blots on a page and is asked to interpret or describe what they see within them. Richard Taylor researched the difference in answers for the test and stated. ‘Your eyes are amazing pattern detectors; they are seeing things that aren’t there’ suggesting answers given were often linked with what that person had seen or experienced before. Ian Sample studied Taylor’s research and found ‘the less complex the blots, the more images people tended to see’ so the more open ended the image was the greater the differences in answer that occurred. Therefore, in Art which in particular is Abstracted the viewer tends to look for or see objects, imagery or meaning which others with varied personal experiences do not. However, patterns occur where people of similar origins or regions in the world will commonly answer in similar ways.

​

How can using Compositional Rules turn an abstract or minimalistic painting into something representational (specifically of landscapes) to the human viewer.

​

As we are looking at Abstract and Minimalistic paintings that look like Landscapes, we do not need to look at how traditional landscape painters used the paint as a medium but instead how they composed the image and there use of light and dark. Looking at the defining points within landscape painting. I have found that the key rules/ ideas around painting / depicting a “Landscape” can be categorised under four headings. Focal Point, speaking on the topic Lisa Marder said ‘Focal points are based on the physiology of vision, the process by which humans actually see, which allows us to focus on only one thing visually at a time. Everything else beyond the centre of our cone of vision is out of focus, with soft edges, and only partially discernible.’ Artists use this to guide the viewers eye around the work which is key to telling a story of the landscape. But it can be used in all forms of painting to create key points of interest within a work. Using this along side Compositional Techniques can create a sense of depth within a work even if purely abstract. This leads on to the next heading, which is Composition, Edgar Degas once said, ‘Even nature has to be composed.’ This is reference to landscape paintings however it shows the power of composition and how effective it is to an image. Using the rule of thirds, golden ratios and other such techniques can completely alter an image changing an abstract work into a landscape purely with positioning of marks, changes in texture or by using the next rule. The Horizon line, Helen South writes ‘A horizon line is essential in art because it allows you to control the height of a viewer's eye as they look at the picture.’ This holds similar effects to using a focal point, but it also adds a perspective and can affect the emotion of the image, large landmasses or large skies effect the viewer differently. Finally, the last key aspect which also adjusts the mood or emotion behind a piece is Value and Form. On Value Dan Scott suggests ‘It is widely considered to be one of the most important variables to the success of a painting, even more so than your selection of colour’. Light and Dark can shape objects create focal points and effect the whole composition of the artwork, giving a large sense of depth and creating a mood to a painting even if there is no physical imagery. Using these Key Aspects or rules can completely transform Abstract and Minimal work in the mind of the viewer, and they all have the power to change the reading of the work on their own.

​

Tools, Methods, Colours and Techniques. How Creating Paintings in these styles can contribute to our understanding.

​

Paint as a medium is extremely malleable changing the way you apply it can have drastic effects on how it reads. "Paint records the most delicate gesture and the most tense. It tells whether the painter sat or stood or crouched in front of the canvas. Paint is a cast made of the painter's movements, a portrait of the painter's body and thoughts." States James Elkins this is a key aspect of why contemporary abstract paintings can hold so much emotion and each mark can hold a sense of direction or movement. Often mark making methods and application of paint are the focus when painting abstract or minimal works. It was Matisse who said, “what I am after, above all, is expression”. Showing that the artist is the motion behind the paint and the paint itself is the attraction. James Elkins said ‘Painter’s love paint itself: so much that they spend years trying to get paint to behave the way they want it to...’ this backs up Matisse’s thoughts on how the artist merely enables the paint, I believe this is vital understanding to hold when looking at expressionistic works. The only thing between the artist and the paint is the tool, using different tools and techniques can change the reading of the paint dramatically. Within Abstract Expressionism some of the main practices are dry brushing, pallet knifes, rollers, squeegee pulls, marbling, pouring/flicking, and colour fields. Although this is a broad spectrum of techniques all with different tools in play Sam Abell states ‘It matters little how much equipment we use; it matters much that we be masters of all we do use’. Suggesting that using one tool effectively is stronger than a work full of different techniques and can give the viewer a clear understanding of a work even though completely abstracted. Donna Baspaly adds ‘Being familiar with art materials and what you can do technically, will only heighten your ability to express emotion.’ Suggesting how once material and tool are mastered marks can create a sense of emotion and give a work newfound meaning. Using this could be the answer to accessing the essence of the landscape. Use of scale can also attempt to invoke the sublime or this emotion. Pierre Schneider an art historian comments ‘to represent the infinite, as perspective claims to do, is to comprehend it in both meanings of the word: to understand but also to contain, and to contain is to limit’. Therefore, to paint the infinite or the impossible the artist must take away the perspective or the “limit” to unlimit the landscape; using the flatness of ground shows the truth of the infinite which is normally not revealed by the illusion of the receding horizon. I believe when artist use scale they attempt to remove this limit by having the work engulf the viewer as they would be in a natural environment. 

​

Can using AI taught purely with Past Paintings find what and how a landscape has been constituted, without seeing a real landscape.

​

After distinguishing what a landscape is, what the sublime is and the key ideas around how we view art and how we use paint, tools, and composition to create emotion, understanding and meaning. I had all the information I needed to start collaborating these genres in different ways to find the point at which these paintings became a landscape. However, I was bias, or at least I had my own personal ideas and as seen in the Rorschach Test these differ between viewers. John Bergers said “seeing comes before words. The child looks and recognises before it can speak.” For me this shows how deeply ingrained human vison is with our preconceptions of what is and is not. Therefore, using AI to define the landscape I have created I take away social constructions based on the idea of what a landscape is. Hopefully through doing this I will find a more truthful definition of a landscape. I decided to train some Image Recognition software with around 250 images of Representational Landscape, Abstract Expressionist and Minimalistic Paintings. The images I used to teach the program all contained at least one of the key rules I have outlined previously. Steve Nouri the head of data science and AI at the Australian Computer Society suggests that ‘AI systems learn to make decisions based on training data, which may contain skewed human decisions or represent historical or social inequities.’ Even though I have programmed the software with a selection of paintings I have chosen. I hope by using the key rules and then the recognition software itself I have diluted my personal definition of a landscape enough for the software to give me an interesting interpretation of my personal works and that of other artists. After Creating the program, I wondered how my software was understanding these images. Looking at Trevor Paglens work I could see how forms within the image were read by the AI and how it searched for similar patterns in other images in order to categorise them Trevor said about his project ‘The works in this exhibition seek to provide a small glimpse into the workings of these platforms, and into the underlying data that structures how machines ‘perceive’ images, language, landscapes and people’ seeing this work massively helped me comprehend how my software was working and made me aware of the possibility creating works to confusing it or make it believe it was all of my categorises at once. I did stumble across one problem however as previously stated a landscape is inextricably tied to the sublime. So, in a sense I am trying to get a computer to interpret the essence of a feeling so while my software will be adept at seeing correlations between composition, value, and texture for each category it may not have outcomes we expect when we see the data. Could this change our own interpretations.

​

My Painting and Collaboration Project and how or if humans can relate to the data outputted and acquired in the process. (and if we can use it).

​

I collaborated my ideas a decided to put my software to the test. Firstly, I showed it previous works I had made to see how it faired with those and to get a better understanding of how it was processing the images. Some of the outcomes surprised me as I thought they would however I noticed it was really affected by some of the key rules which showed to me how prevalent they are within the classical landscape paintings. I then split my line of inquiry in two, can humans relate to this data and if so, can we use it. The next was to create works specifically with key fundamentals from all three of my Categories at once and see how the software interpreted them and hopefully take me closer to the boundary where they become one. So, what would I do to create these paintings? From Landscape painting I looked to simply take the rules of old, such as rule of thirds and golden ratios which seemed to hold such an important part within the fundamentals of the practice. For Abstraction I sort after the techniques and application styles for painting, such as dry brushing, throwing, squeegee pulls, and Colour Fields. I wanted to use 3 different shapes for my colour fields (black area) as Abstraction and Minimalism were heavily derived from cubism and other similar movements which relied greatly on shapes. The background layers I had decided on was a primed layer of black, followed by white and black squeegee pulls mixed on the surface. Then using a scrapper to apply thick layers of black over the top leaving some areas from the previous layer visible. After this I drybrush white over this whole layer to bring out the highlights and depth of the paints surface. The next step is the colour field which was a squeegee pull of black in a selected section. Finally, I would apply a large white minimalist mark over the work to create the final piece.  Most of the design process for the works were done in photoshop laying ratios and areas over textures I had made during painting the background layers. Testing different marks, I had taken from my photos and past works by using the wand tool to steal sections. I had already taken colour out of my practice within different projects however I thought it would be good to continue this as I did not want my software to use a bias based on the images of landscape, I had taught it with.

​

The next inquiry was a matter of how to use my data and what could it teach us. I sought to identify if artists could interpret or create images based on the percentage outcomes my software was returning. So, without showing them paintings that had be categorised or explaining how my software worked I asked various Artist, Students and Friends to create a painting that was 60% abstract 20% landscape and 20% minimal. I look to show these entries within a body of work but also re use the data collected to gain more works from other artists. Hopefully in the near future my question of can humans related to this data will be answered and my search for the boundary in which landscape becomes abstract and vice versa will be more coherent.