Algorithms, Narrative and the Uncanny



Donald McKay
Nagle Hartray Architecture

August 2013



 “On the Web, powerful algorithms are sizing you up, based on myriad data points: what you Google, the sites you visit, the ads you click. Then, in real time, the chance to show you an ad is auctioned to the highest bidder.”

This quote from a New York Times article seems prosaic despite the relative infancy of using computer algorithms for commercial purposes. Few internet users are unaware that they are being ‘sized up.’ We make a sort of peace with this idea in exchange for the convenience, knowledge and connectivity that the internet affords. Our increasing reliance on algorithms to choose potential mates, to decide stock transactions in the millisecond when some advantage exists, to identify sympathetic voters for political campaigns has fostered a laissez-faire attitude towards their use. Since computer algorithms thrive on data and more data is now collected than can be processed, computer algorithms are and will continue to be a growth industry.

A 2012 article in The Atlantic titled “Can the Computers at Narrative Science Replace Paid Writers” offers a glimpse of what that growth may look like. Narrative Science “transforms data into stories and insight.” Th e company’s roots are in Northwestern University’s Medill School of Journalism. Narrative Science uses computer algorithms to convert raw data into narratives in the belief that stories are more valuable than numbers for gleaning insights. Th e company’s early work focused on clients swimming in numeric data to create local sports stories and financial reports, work that journalists typically did not engage in. But as their system evolves, it will become able to draw its own conclusions from non-numeric data and this development threatens traditional journalism. The Atlantic writer concludes on a hopeful note for her profession:

 Figure:  The Uncanny Valley

“…even our simplest moments are awash in data that machines will never quantify—the way it feels to take a breath, a step, the way the sun cuts through the trees. How, then, could any machine begin to understand the ways we love and hunger and hurt? The net contributions of science and art, history and philosophy, can’t parse the full complexity of a human instant, let alone a life. For as long as this is true, we’ll still have a role in writing.”

The developers of computer-generated narrative systems believe it will be the limitations of their systems that will define future roles for writers, not the other way around. The extent to which writers and other similarly creative professionals maintain meaningful roles that resemble the ones that attracted them to their professions in the first place may depend on the answer to an almost existential question—is the human mind uniquely capable of imagining profound creations?

Commercial interests have forced the fields of robotics and 3D computer animation to address this question.  Robot designers discovered that an industrial robot with some human likeness triggers a slightly positive emotional response in human observers and that our emotional responses become increasingly positive as human likeness improves. The most positive response results from a likeness that is indistinguishable from a human being. But when human replicas look and act almost but not exactly like human beings, then our emotional response changes to one of revulsion. In the graph that maps this progression, the shift from positive to negative emotional response is illustrated as a large dip that is popularly known as the “uncanny valley.” Understanding the uncanny valley is important to gaming designers and filmmakers that rely on 3D computer animation because, as in the field of robotics, commercial success depends on avoiding it. The animated film The Polar Express is an oft-cited example of the revulsion that results from straying into the uncanny valley.









A similar phenomenon has developed with architectural illustrations. Computer drawing and modeling programs come equipped with a rendering function that will generate illustrations of a design based on relatively simple material selections. Illustrations with a conceptual or abstract quality tend to generate a positive response. And when an illustration is indistinguishable from a photograph of a real building, similar to when a human replica is indistinguishable from a human being, the illustration can generate genuine excitement for a proposed project. But when a computer-generated illustration attempts
but does not achieve photorealism, as often happens with rendering programs, it tends to trigger an uncanny valley-type negative response.

In these examples, the uncanny valley is an unintended consequence of trying to render real something that is not. But the uncanny is not necessarily accidental. Alfred Hitchcock employed the uncanny to create anxiety in his audiences—birds that betray expectations of docility, a crop duster as killing  machine, the mistaken and assumed identities of his characters. Other filmmakers, including many  Hitchcock imitators, have employed the uncanny to similar effect. The renewed popularity of zombies attests to the enduring entertainment value of the uncanny.

Architects have also employed the uncanny intentionally. In his book of essays, The Architectural Uncanny, Essays in the Modern Unhomely, Anthony Vidler uses Sigmund Freud’s essay, The Uncanny, to link the psychology of the uncanny with modern architecture’s alienating eff ects, some of which were unintended and others that were. Some deconstructivist architects employ the uncanny as a design strategy aimed at shaking loose humanist-idealist traditions that they perceive as having  informed all architecture since the Renaissance, including Modernism. Peter Eisenman explains this point of view:

1  The scarred surfaces of the Jewish Museum convey a sense of a horror and loss.

“For five centuries the human body’s proportions have been a datum for architecture. But due to developments and changes in modern technology, philosophy and psychoanalysis, the grand abstraction of man as the measure of all things, as an originary presence, can no longer be sustained, even as it persists in the architecture of today.”

In the Deconstructivist Architecture exhibit that gave definition to the term, co-curator Mark Quigley’s abstract for the exhibit captures the Freudian “that which ought to have remained secret and hidden, but has come to light” essence of the uncanny in deconstructivist architecture:

“The nightmare of deconstructivist architecture inhabits the unconscious of the architect. The architect merely countermands traditional formal inhibitions in order to release the suppressed alien.”

Built deconstructivist architecture has proven a mixed bag. Ironically, given Deconstructivism’s goal of autonomous architecture, it has been most successful when it expresses shared cultural repressions. The best example may be one of Deconstructivism’s earliest realized projects, the Jewish Museum in Berlin designed by Daniel Liebiskind. The museum’s gashed exterior skin and its interior spatial voids express an appropriate sense of scarring and absence. The craftsmanship with which the construction is executed only serves to heighten a communal sense of anxiety. By contrast, Deconstructivist architecture has been least successful when it employs its characteristic splintered and

threatening forms simply to destabilize and disorient us as a critique of perceived humanist-idealist traditions. In these examples, the sense of alienation that is successfully cultivated does little to reveal  new values and little more than extend negative aspects of the Modern architecture that  deconstructivists intend to critique.

The idea of autonomous architecture owes something to literary theory that seeks to shift focus from the author to the reader. Roland Barthes captures the essence of this reconsideration in his essay, The Death of the Author:

“Thus is revealed the total existence of writing: a text is made of multiple writings, drawn from many  cultures and entering into mutual relations of dialogue, parody, contestation, but there is one place where this multiplicity is focused and that place is the reader, not as was hitherto said, the author. The reader is the space on which all the quotations that make up a writing are inscribed without any of them being lost: a text’s unity lies not in its origin but in its destination.”

There is a notable similarity between architectural theory that seeks to shift emphasis from ‘author’ to ‘reader’ and real life enterprises like Narrative Science that actually dislocate authors. The computer algorithm has become an important tool that oddly connects architectural theory and commercial  interests. Similar to the way that algorithms are being used to turn raw data into meaningful narratives,  architects are creating computer algorithms towards realizing an objective, and therefore universal,  architecture that defies authorship.

Proponents of this approach credit Postmodern and Deconstructivist thought with preparing the way for their work.

A problem with the subject (author) versus object (reader) thinking that underlies much current architectural theory is that it assumes a genius-creator role for the architect similar to the role of an author writing a text. For many architects and educators, this concept is a myth that gained prominence in Modernism and may have reached its zenith with the recent popularity of the “starchitect” model. Historians, the media, and architects themselves have conspired over time to foster this concept. The privileged position that this gives an architect is an understandable basis for criticism that includes calls for “objective” architecture.

But most architects understand that architecture is social by nature, involving a client, a broader community, and a team of professionals and constructors that are necessary to realize even modest buildings. For many, architecture’s communal nature underlies meaningful architecture and for them the idea of abandoning collaboration to ‘authorless’ creation is anathema. If architecture results from collaboration, then the question of objective architecture seems less relevant.

The idea of an objective architecture achieved prominence a century ago when profound advancements in science and industrial design threatened to make architecture appear little more than a decorative art by comparison. Architects developed Modernism (Functionalism) in an attempt to root design in a rationalism as unassailable as science. Postmodernism followed as a critique of Modernism’s failures, particularly at the urban scale. Postmodernist urban design successfully addressed many of these ailments, but Postmodern architecture was less successful, tending towards a decorative approach like a repressed memory from a century earlier.

Recent interest in an objective architecture is partially a reaction to Postmodernism similar to the way in which Modernism was a reaction to the historicism that preceded it. But recent calls for an  objective architecture are also a reaction to the challenge of multiculturalism. An objective  architecture, particularly one that achieves universality, could avoid the potential offenses of otherwise “subjective” architecture.

This admittedly reductive history suggests that architecture embraces the idea of an objective  architecture when it is threatened with cultural insignifi cance. The same history also suggests that the  consequences of pursuing an objective architecture can result in the type of irrelevance that it attempts  to avoid. Considered together with the potential uncanny consequences of computer-generated design, a suspicious view of algorithm-generated objective architecture is justified.

Perhaps the most surprising aspect of Narrative Science’s success is that it demonstrates that  narrative itself remains relevant and compelling. Architects have long used narrative similar to the way  that Narrative Science does—to render otherwise inscrutable information, design in the case of  architecture and numeric data in the case of Narrative Science, understandable and meaningful. The uneven success of narrative as an architectural design strategy subjects it to criticism and claims of  obsolescence in favor of emerging strategies like computer-generated design. Distinguishing between  two forms of architectural narrative, imitative and associative, will help us to understand narrative’s  past failures and its continuing relevance to architectural design.

Imitative narrative is characterized by the imitation or mimicry of established, sometimes universal, forms. The stylistic historicism that predominated both pre-Modern and Postmodern architecture  exemplifies the skin-deep

character common to imitative architectural narratives. These examples illustrate how periods in which imitative narrative strategies flourish are often followed by periods that call for ‘objective’ architecture.  Imitative narratives ring hollow because they are divorced from both their original and contemporary value systems. This cognitive dissonance often reveals itself in uncanny aesthetics that are unintended and evoke a reaction similar to that of the animation that approaches but does not achieve human likeness, of the architectural illustration that attempts but falls short of realism.

Associative narratives are rooted in something shared between us—histories, memories, culture and, most importantly, our human condition. The communal nature of associative narrative assures a relevance that eludes imitative narrative. The most durable narratives have a ring of authenticity that derives from something essential about the human condition. There seems to be a proportional relationship between the durability of an associative narrative and the depth to which it addresses our shared human condition—the more essential, the more authentic and durable. One example is the Jewish Museum cited earlier. Its design convincingly conveys the absence and scarring of the historical events that it documents in a way that transcends specific historical events to address something fundamental about the human condition, in this case man’s inhumanity to man. The Museum’s architecture and exhibits move me not because I am a victim of the Holocaust, but because I am human. Another example is the Stockholm Public Library in which its architect, Gunnar Asplund, uses the biblical allegory of the Garden of Eden to reflect on the nature of knowledge. Standing in the Library’s central book-lined drum, its Eden, I understand knowledge as a kind of heaven on earth.


2 Adam and Eve door pulls introduce the Garden of Eden allegory to Stockholm Library visitors. The  Library’s book-lined drum equates knowledge to heaven.

Differences between these imitative and associative forms of narrative may be understood through a series of corollary characteristics:

imitative                        associative
literal                             abstract
superficial                      psychological
unintentional uncanny   intentional uncanny
cognitively dissonant      authentic
universal                        communal
homogenous                  various

We understand the profundity of the Jewish Museum and the Stockholm Public Library as works of architecture at least in part through their associative design narratives. As a practicing architect, I am less interested in knowing whether an associative design narrative is necessary to realize profound architecture than I am in simply knowing that it can. Associative narrative has proven an effective means of addressing something essential about our shared human condition, which is, I think, necessary for profound architecture.

Are computer-generated creations capable of similar profundity? This returns us to the question of whether the human mind is uniquely capable of profound creations. If humans are simply incredibly rich and complex machines, as a founder of Narrative Science believes, then it is only a matter of developing complex enough technologies before computers imagine profound creations.2 But if there is a spiritual dimension to humanity that makes us something more than machines, then it is hard to imagine computers ever achieving more than sophisticated mimicry. Even if this proves true, as we continue to adapt to an algorithmic future, what we may at first consider uncanny we may eventually come to accept as a near enough substitute. Architecture provides evidence of this phenomena with every building that disappoints the knowing eye in its

 mimicry of past architecture, but which is widely accepted nonetheless. There may be a tipping point at  which this sort of uncanny becomes so normalized that the notion of profound creations becomes  irrelevant. If this happens, then we will lose something that distinguishes us as human beings. The most  serious question for computergenerated design may be whether it moves us towards such a tipping point  or helps us to resist it.






3, 4, 5 Three public libraries designed by Nagle Hartray Architecture that use narrative as a  design strategy (clockwise): Oak Park Public Library in which the exposed heavy-timber roof  structure recalls similar wood construction in the Harry Weese designed Oak Park Village Hall of  a generation earlier; Warren-Newport Public Library in which the new red panel clad addition  transforms the iconic white modern existing building into a collaborator in a narrative about the  site’s and community’s agrarian roots and relates the Library to other nearby red brick-clad civic buildings; and Fountaindale Public Library in which custom patterned glass casts forest-like  shadows that fulfi ll a community daydream for a library in the park despite a substantially  treeless environment.


1 Natasha Singer, “Your Attention, Bought in an Instant,” New York Times, November 18, 2012.
2 “Humans are unbelievably rich and complex, but they are machines,” Hammond says. “In 20 years,  there will be no area in which Narrative Science doesn’t write stories.”

1 By author
2 By author
3 Scott McDonald©Hedrich Blessing
4-5 ©

200 OK


The server encountered an internal error or misconfiguration and was unable to complete your request.

Please contact the server administrator, [no address given] and inform them of the time the error occurred, and anything you might have done that may have caused the error.

More information about this error may be available in the server error log.