Mary Flanagan: Hope Is The Thing With Feathers

May 13 - July 09, 2022

Hope is the thing with feathers

Some projects open a window into the past. Others draw us towards the future. 

In “Hope is the Thing With Feathers,” Mary Flanagan presents three new projects that speak to our collective imagination: [Grace: AI] Daydream (2021), [the Mirror Book] (2022), and [Colors of Remembrance] (2020-ongoing). Through magical collaborations and algorithmic daydreams, these projects together make visible the humans in our technology, re-imagining the present and suggesting more empathetic possibilities for the future. 


[Grace:AI] Daydream

What if there were an algorithm biased towards, instead of against, women artists? This is the question that underpins Flanagan’s long-term research-based collaboration, [Grace: AI] [2019-ongoing]. As a Deep Convolutional General Adversarial Network (G.A.N.), [Grace: AI] uses a deep learning model to generate new data from training data, which could either be selected by humans or scraped from the internet. The Daydream series is the second installment in that project.

During the past few years of intense techno-isolation, explains Mary Flanagan, the simple act of staring at the sky felt like freedom. One day, she began to imagine, what do machines dream? Specifically, she imagined how [Grace: AI], trained on images from female artists, would render that cloud-filled sky.

Flanagan was inspired by both physical and virtual experiences. 

At a residency at the home of French Surrealist Dora Maar, Flanagan was intrigued to find only two of Maar’s paintings exhibited. Where were her life’s work (an impressive oeuvre including thousands of paintings, photographs, collages and collaborations)? A quick Google Search for, “Where is Dora Maar’s work?” returns the artist’s “spouse” of eight years (Picasso) above her profession¹. Even Maar’s sketchbooks are in the Picasso collection, rather than her own. 

Poet Marie Ponsot observed, “This dawns on me: no cloud is measurable.” 2 Searching for artwork by women seems to turn up the same problem. If you search “Painting by a woman,” the top 20 results are: two De Koonings, a Vermeer, two Leonardos, a Whistler, a Manet, a Monet, a Kahlo, two Klimts, a Sargent, a Picasso, a Boticelli, a Hopper, a Degas, a Klimt, two paintings by a Pre-Raphaelite named John William Waterhouse, Matisse, and two Rafaels.

As a recognized scholar in the field of game theory, Flanagan set out to collaborate with historical archives to create a new, intentionally biased smart machine—one that would flip the gender bias of our current technology. 

She began contacting global museums and archives, only to find that much of the work by women hadn’t been prioritized for digitization, or was confusing to access. What Flanagan discovered was even bigger than her initial vision: an entire missing database, made up of the artwork of women. Historical paintings by women of color in the West were especially underpreserved (despite many commendable new initiatives to address this).

For search algorithms, if something isn’t properly digitized, it doesn’t exist. The historical archive of women’s art is largely, in effect, invisible. Flanagan began to assemble that database herself, contacting global museums, archives and foundations to begin to put together the missing archive.

In the first series, [Grace:AI] Origin Story (Frankenstein) rendered new images, inspired by female painters and Mary Shelley, resembling Frankenstein portraits.3 Today, in [Grace:AI] Daydream, Flanagan turns her attention to tropes of creativity. She’s letting the machine daydream.

Each piece begins with a set of potentials. [Grace: AI] is trained four times, using different datasets: the first three involve works by female-identifying artists from global archives, pre-1950. The last set includes images of clouds.

In the arts, clouds carry a rich symbolic history (as does daydreaming). They can suggest isolation or ecstasy, a romantic reverie, or the wrath of the heavens. For Joy Harjo, a cloud can lift souls to heaven. In “A Daydream,” Emily Bronte laments the fleeting joys of seasonal rebirth: “When winter comes again, Where will these bright things be?”

Do machines, indeed, dream of electric sheep?4 It depends on the dataset they’ve been trained on. 

There are invisible histories all around us. Our current archives don’t show them to us, yet, but [Grace:AI] Daydream can give us one glimpse. This is a speculative project, bridging the bias of our present with a possible vision for the future. To get there, today, you still have to dream. 


[Colors of Remembrance]

May 13, 2022, Johns Hopkins University reported a tragic milestone: 1 million Covid deaths in the US. The enormous tragedy of that single number is impossible to process.

[Colors of Remembrance], a computational drawing, is a solemn monument to those individuals. The work is executed using the artistic language of minimalism and geometric abstraction, with a nod to the ancient tradition of representing numbers directly with hashes. By rendering the data into a visualization of those who perished, the memorial creates space to consider the tragic immeasurability of the pandemic.

Each drawing represents one day, and each unique colored hash represents one life lost, as recorded that day by public data. The drawings begin on the first recorded day of the Covid pandemic—for the US, that was February 29, 2020. There are a total of 2161 Pantone colors, enough to create a unique base color for 2161 days. Each hash a unique saturation and value, derived from the day’s Pantone base. There are 16,777,216 possible colors total. Flanagan initially hoped the project might end after a few months; today, she reports, it seems unending.5

Attempting to fully come to terms with this tragedy, at the scale of recorded losses, is an impossible task. [Colors of Remembrance] is therefore an compassionate, poetic, yet still impossible challenge, an attempt to embody the dehumanizing data of grief.


[the Mirror Book]

In this installation of [the Mirror Book], Flanagan collaborates with Emily Dickinson to create a dynamic, hybrid poet across time and consciousness. To the left are poems by Flanagan, 2006-present. To the right, Dickinson, 1858-65. Words flow between, creating surprising flashes of language.

Flanagan’s poems were written around and in the decade and a half following the financial crisis, a moment notable for the collapse of traditional infrastructures, the rise of digital identities and cryptocurrencies, and political polarization. Dickinson’s were written in a time of fractured national identity, Emancipation, and hyperinflation, the years leading up to and during the American Civil War.

It’s an exercise in juxtaposition and projection, a program playing out through language and time. Mary Flanagan through the headset of Emily Dickinson, Dickinson through the headset of Mary Flanagan. Meaning occurs as a relational flows through the living texts, one floating word at a time. 

Pay attention to position, momentum, a trading of context and consciousness. The book itself is mesmerizing. Look out for gems like, “Hope is the lot with feathers / That perches in the corner.”

This is a magical notebook, full of fire and potential. It’s also a tender and humbling experience. “If I read a book and it makes my whole body so cold no fire can warm me,” Emily Dickinson wrote once to her editor, “I know that is poetry. If I feel physically as if the top of my head were taken off, I know that is poetry. These are the only way I know it. Is there any other way?”6


Katie Peyton Hofstadter

May, 2022


¹ Maar was integral to the early Surrealist movement, both as an individual artist and collaborator; the “d” in Andre Breton’s Gravida gallery even stood for “Dora.” The popular The Art Story website reveals Maar’s intentional erasure from the popular narrative, and the weird, misogynist justification often attached—because, simply, she had it coming: “Maar’s career in its later stages was unfortunately tainted due to this tumultuous affair and as such stands as a warning for others.” “Dora Maar Artist Overview and Analysis,” 2022. Content compiled and written by Kristen Osborne-Bartucca.

² Marie Ponsot, “This Bridge, Like Poetry, Is Vertigo,” 2005. The epigraph, a quote from William Blake, also speaks to this project: “In a time of dearth bring forth number, weight, & measure.”

³ Mary Flanagan, [Grace:AI] Origin Story, 2019.

4 Paraphrased from Phillip K. Dick’s 1968 novel, Do Androids Dream of Electric Sheep, used as the basis for the 1982 film Blade Runner

5 Mary Flanagan, [Colors of Remembrance], 2020.

6 From a note written to editor Thomas Wentworth Higginson, 1870.