Rethinking The Art Genome Project

Mia Gao

Abstract: The Art Genome Project is a classification system of the online art auction as well as education platform, Artsy. This paper critiques the design of The Art Genome Project in terms of its semiotics rationale and its affordances as an education as well as business tool. The analysis shows that The Art Genome Project fails to provide information to users efficiently and it is an inadequate imitation of real education or auction platforms.


The Art Genome Project is the search technology and classification system that powers Artsy, which is a website claims to “make all the world’s art accessible to anyone with an Internet connection as a resource for art collecting and education” (“About | Artsy,” 2015). The Art Genome Project maps the characteristics that connect artists, artworks, architecture, and design objects across history and will recommend artworks to users that share similar ‘genes’ with artworks they like (“About | Artsy,” 2015). The generative feature is an essence embedded in human meaning making process; however, The Art Genome Project lacks a visible classificatory referential, so all descriptive, emotional and art historical categories are rendered in a flat network, which made the Art Genome Project an irrational discourse of meaning making process and a poor imitation of museums.

Although promoted as an educational online art resource and has an ambitious mission to make art accessible to general public, Artsy’s business model is commission based. The Art Genome Program serves as an art dealer to potential buyers, it recommends artworks to buyers and connects them to sellers. However, Artsy’s affordances leave the audience an impression that the artworks they are recommended via Art Genome Project are not scarce.

Therefore, Art Genome Program is the backbone of Artsy but its designs made the website an incompetent education resource and an unprofessional online auction platform.

C. S. Peirce’s Model of Semiosis and The Art Genome Project Rationale

The Art Genome Project’s database is composed at present of more than 1000 characteristics or “genes” that range from subject matter (including conventional art history categories like portraiture as well as descriptive characteristics like “dark”) to art movement. Led by Carter Cleveland, computer science graduate from Princeton and son of art historian and collector David Cleveland and Sebastian Cwilich, former executive of Christie’s and Haunch of Venison director, Artsy recruited a team of art historians who spent years compiling the taxonomy (Miller, 2011). Notably, Artsy’s taxonomy is not function through binary tags. Artsy disclosed a main feature of their algorithm saying:

“Genes are applied with values ranging from 0 to 100. While not seen by users, such gene values account for the strength of a relationship between artists and artworks. It also enables similarity to be expressed in a more nuanced way than it might be with just tags because one can weigh various attributes of an artist or work of art to establish which might be the most or less important. Furthermore, such nuance allows for matching potential collectors with artworks based on their tastes and preferences” (Mufti, 2011).

The technical complexity and academic support allow Artsy’s recommendation engine to evaluate a artwork at a click and provide endless associations among different artists and artworks. In an attempt to describe the association between infinite characteristics , I will employ C. S. Peirce’s model of semiosis as an explanatory model for The Art Genome Project.

The model of semiosis as a cognitive-generative-social process was developed by C.S. Peirce from 1860s, but it is one, I would argue, that can be applied to varying degrees of any meaning making process. According to Peirce, “meaning is an open triadic process with one element of the structure, the interpretant, always unfolding new meaning.” We can apply two of Pierce’s major discoveries to further understand Art Genome Project:

1.  Human thought is based on signs in symbol systems which have a structure  of material/perceptible and cognitive relations that unfold dynamically in human-experienced time as shared cognitive processes.

2.  Meanings, learning, and knowledge are produced through symbolic-cognitive transformations (signs yielding interpretants expressible in further signs in unlimited and open-ended chains or networks) (Irvine, 2015).

Art Genome Project users’ understandings of the associations between different artists are based on their art history knowledge and interests and the dynamic associations can be expanded as an endless network. The symbolic-cognitive transformations enable users to associate or relate Egon Schiele’s Self Portrait with Raised Bared Shoulder, 1912 to “Close Up”, which is a conceptual component of Schiele’s painting.

Screen Shot 2015-04-25 at 4.42.41 PM

“Egon Schiele | Self Portrait with Raised Bared Shoulder (1912) | Artsy.” N.p., n.d. Web. 25 Apr. 2015.

Egon Schiele is known for his erotic and deeply psychological portraits, particularly self-portraits. Schiele’s Self Portrait depicts the artist himself half naked, looking cynically at the viewers making the portrait confrontational and intense. Starting from Self Portrait with Raised Bared Shoulder, 1912, the software generates and presents other “genetically” related artworks below the painting.

Screen Shot 2015-04-18 at 11.00.39 AM “Egon Schiele | Self Portrait with Raised Bared Shoulder (1912) | Artsy.” N.p., n.d. Web. 18 Apr. 2015.

By using “Close Up”, a descriptive characteristic, The Art Genome Project links Egon Schiele’s Self Portrait (1912) with  Lina Scheynius: Untitled (Diary), 2004.

Screen Shot 2015-04-18 at 11.19.00 AM

“Lina Scheynius | Untitled (Diary) (2004), Available for Sale | Artsy.” N.p., n.d. Web. 18 Apr. 2015.

Scheynius is a young Swedish photographer, noted for her raw and unabashed depiction of intimacy and sexuality. The signs embedded in Lina Scheynius’ photograph including “love”, which directs the chain to new “interpretamen”, such as Keith Haring’s Untitled, 1982.

Screen Shot 2015-04-18 at 2.14.45 PM

Keith Haring is an American artist and social activist, whose works are great representations of “New York Street Art”. Haring directs the chain to “East Village Art”, which further leads to tremendous number of artists and artworks. Artsy listed American artists Jeff Koons (1955) in “East Village Art”, and the American Contemporary Pop Artist is connected to Japanese artist Takashi Murakami. Jeff Koons is famous for his giant reproduction of ordinary objects such as enlarged balloon animals in stainless steel. Takashi Murakami, on the other hand, is a contemporary Japanese artist who marries traditional Japanese art with contemporary Otaku culture. The “gene” Koons shares with Murakami is “provocative”, which made Artsy recommend the Chinese artist, Ai Weiwei, who shares both the “Asian” and “provocative” genes with Murakami. The meaning chain starts from Egon Schiele opens an endless Pandora’s Box, which leads the Austrian Expressionist to a Chinese Conceptual artist.

Art Genome Project’s design shows the generative as well as open-ended features Peirce highlights in his Semiosis Model.

Three Phases of Interpretants and Artsy’s Flat Juxtaposition

The Art Genetic program seems provides a tangible model for natural human meaning-making process. However, Peirce emphasizes human symbolic cognitive process is a generative thus not standard model. The characteristics one associates with an artwork highly rely on her personal experience, knowledge set as well as motivation. Therefore, an efficient and pragmatic classification system should provide a clear hierarchical structure to guide users.

Peirce distinguished three phases of “Interpretants”: Immediate (our first impression that something perceived as symbolic has meaning), Dynamic (the meanings that emerge and develop with new associations in uses and contexts), and Final (the motivation for completed meaning experienced as future-directed unlimited possibility) (Irvine, 2015). The Art Genome Project juxtaposes different “phases of Interpretants” to the users simultaneously without explanation, which makes the meaning network flat and difficult to understand.

The Immediate Interpretant is “the schema in [our] imagination…all that is explicit in the sign apart from its context and circumstances of utterance” as Peirce described (Atkin, 2013). For Egon Schiele’ Self Portrait with Raised Bared Shoulder (1912), the Immediate Interpretant could be “close-up”. “Close-up” is a general and explicit feature people without any pre-requisite professional knowledge will associate with the painting because the subject in the painting is tightly framed. Notably, “close-up” at here is not a photography or filmmaking term but a descriptive trait people naturally use to describe a close view.

The second type of interpretant is the Dynamic Interpretant. Peirce describes the Dynamic Interpretant as the “effect actually produced on the mind” (Atkin, 2013). For Egon Schiele’ Self Portrait (1912), the Dynamic Interpretant could be “provocative” because the subject stares at viewers directly displays an unnatural eyebrow flash, which is an universal facial expression “will often be perceived as a foe signal” (Schafer & Karlins, 2015). “Provocative” consists of the interpretations made previously, that is, it consists of the “close-up” in the sign chain. Often a close-up scene will generate a confrontation feeling; therefore, people will spontaneously reach the “provocative” understanding.

Peirce describes the Final Interpretant as, “effect that would be produced on the mind by the sign after sufficient development of thought”(Atkin, 2013). At the end of inquiry for the Egon Schiele painting, people will reach “self-portrait” because of the self-reflective nature and “close-up” composition artist often adopted when creating their self-portraits. As David Savan puts it, “Peirce’s intention was to identify the third type of interpretant as providing a norm or standard by which particular stages (Dynamical Interpretants) of an historical process may be judged”(Savan, 1988). Self-portrait became a classical main subject since Jan van Eyck’s panel self-portraits in fifteen century, and set up a “close-up” composition tradition for this genre of art. Therefore, for art historians, “self-portrait” is a term, which directs the future use.

Artsy lists “close-up”, “provocative” as well as “self-portrait” together under Egon Schiele’ Self Portrait with Raised Bared Shoulder (1912) without any rationalized description among those “genes”; therefore, the juxtaposition breaks the linear phrases of interpretant. Art Genome Project’s design drives users to focus on trivial “preferences” while ignoring the depth of structure in meaning making process. The irrational juxtapositions are overwhelming for general users and useless for knowledgeable collectors. Although The Art Genome Program is a reasonable application of Peirce’s Model of Semiosis, it is not a pragmatic tool for education or collection.

A Museum Without Wall

The Art Genome Project has an ambitious mission, which claims it offers digital images housed in partners’ museums as well as galleries and it will be an online resource for education. Lacking a visible classificatory referential, all descriptive, emotional and art historical categories are rendered in a flat network, which makes the Art Genome Project an irrational discourse of the real artworks in museum environment.

Prevailing debates on the relationships between virtual “historical” objects and physical historical collections used to be bounded by established material culture and “aura” paradigms. A liberating force of scholars, including André Malraux and media theorist Lev Manovich, took up Walter Benjamin’s concern argued the digital historical objects are subjects to cultural politics and were not attempt to identify with physical reality. Therefore, a virtual museum is an abstract organizational system, which will ideally create a more condensed and well-structured “cultural encyclopedia”. Malraux describes museum’s function is more and more intellectualized and viewers will have the expectation that curators will present well-organized prototypes as art historical exemplars to them. Malraux writes,

For over a century our approach to art has been growing more and more intellectualized. The art museum invites criticism of each of the expressions of the world it brings together; and a query as to what they have in common. To the “delight of the eye” there has been added-owing to the sequence of conflicting styles and seemingly antagonistic schools-an awareness of art’s impassioned quest, its age-old struggle to remold the scheme of things (Malraux & Malraux, 1978, p.15).

Art Genome Project’s mission is same with the notion of virtual museum but it fails to provide a clear structure. Without a rational discourse of the history behind the objects, viewers will find it hard to understand why certain art works are masterpieces thus find the“cultural encyclopedia” difficult to read. Malraux explains, “an artist’s supreme work is not the one in best accord any tradition-nor even his most complete and ‘finished’ work-but his most personal work, the one from which he had stripped all that is not his very own, and in which his style reaches its climax. In short, the most significant work by the inventor of a style” (Malraux, 1978, p.15). Therefore, people will still long for a clear referential structure.The Art Genome Project categorized artworks according to artists, subject matter, medium etc. without any explanation of the categorization, hidden the complexity as well as the process. The multiple levels of abstraction generates infinite hyperlinks, which render it difficult to delve into each characteristic. Taking Egon Schiele’s Self Portrait (1912) as an example, people would be easily distracted by the series of hyperlinks ending with Ai Weiwei’s artworks and forget about the clue linked from the Austrian impressionist to the Chinese conceptual artist. The computer-generated imagery museum offered the viewer unlimited opportunities to explore, but as Fiona Cameron descripted “It represents a social desire for rational meanings by users in order to give them a clear statement of the visual surrogate’s value as a signifier, as an analogy of its physical counterpart (Cameron & Kenderdine, 2007). Therefore, viewers need a rational structure to explain why certain exemplars were chosen and what’s their inner connection.

Compare to The Art Genome Program, Google Art Project is also an online platform, which provides digital images of artworks housed in the initiative’s partner museums and gives a much clearer structure to users. Museum curators handpicked representative masterpieces to digitalize and the Virtual Gallery Simulations offered the viewer a sense context of the original structure and curatorial design embedded in the exhibition space. Different from Artsy’s Art Genome Project, it does not provide further curatorial explanation besides the description provided by museums. Notably, Google Art Project allows curators from their partners’ institutions to create digital exhibitions. The online exhibition feature is a digital story telling process organized by professional curators with images, videos and word descriptions. Users are welcome to explore and search collections by medium, curator and date etc.. Each exhibition is a complete story articulated by professional curators, which ensured its credibility so users won’t feel the artworld is accessible.

Screen Shot 2015-04-20 at 11.06.47 AM

“Search – Google Cultural Institute.” Explore Exhibitions. N.p., n.d. Web. 20 Apr. 2015.

The Art Genome Program is designed to build a concrete and visible model to guide viewers to explore the artworld but it fails to construct a clear structure. Compare to Google Art Project, The Art Genome Program provides redundant characteristics under each artwork without any legitimate explanation or hierarchy of categorization among the associations. Therefore, The Art Genome Project is not a well-organized digital museum.

Art Genome Program Website’s Affordances

The art industry is a relationship-driven business funded not only by auction houses or galleries, but more importantly the merits are embedded in “cultural capitals” generated by art fairs, museum exhibitions and more.

Art Genome Project claims it’s an online resource for galleries and art lovers, but Artsy’s business model reveals it is actually a middleman between galleries and collectors. Artsy relies on two major funding sources: first, it charges the galleries a monthly subscription fee to access to the platform; second, it takes a small commission on benefit auctions (Griffith, 2015). Artsy is not taking a combative approach attempting to democratize the artworld by excluding galleries from the industry, while it adopts a collaborate approach tried to thrive in the old-school business.

In the art ecosystem, the artistic merits cannot be derived directly from the economic transaction. The artistic capitals people do receive from the art they pay for are not only a tangible good, but social capitals such as inspiration, pride and enlightenment. As Pierre Bourdieu describes, the conversion between social capitals and economic capitals can be obtained only by virtue of a social capital of relationships (Bourdieu, 1986). Museums, auction houses and galleries are in a symbiosis system, namely an artwork’s market value relies on not only the authenticity and condition of the work, but also depends on “whether the painting is hung in the museum, where it is hung and how often, where it was before it came into its present ownership, in whose private collections the painter is represented, whether he is being spoken of in the popular as distinct from the professional art journals and if dealers are promoting or disregarding him”(Grampp, 1989, p.27). Therefore, art dealers are eager to leverage the “social capitals” of an artwork by carefully manage where and how it was presented to viewers.

As an online platform, the social capitals offered by The Art Genome Project solely rely on its interface. Gibson introduced a concept, “affordance”, which means “action possibilities” latent in the environment (Zhang & Patel, 2006). I will apply the concept here to analyze The Art Genome Project’s interface. One of the important properties of Gibson’s affordance is that he describes affordance provides values and meanings complementarily to the perceiver, so affordance is neither an objective property nor a subjective property”(Zhang & Patel, 2006). In visual representation layout like website interface, the structures and information showcased on the platform provides the allowable actions to the viewers. Users of The Art Genome Project capture the external representations such as the layout, structure and functions on the website and internalize the cognitive signs simultaneously. However, The Art Genome Project’s interface design gives users an impression that the artworks they are present on the website are not scarce, which is withering its “social capitals”.

The Art Genome Project’s interface has three distinct features compare to other online fine art auction websites.The Art Genome Project lists artworks share similar “genes” with an “infinite scroll” embedded in each characteristic page; also, users can “follow and like” artists and artworks; and finally, they can sort artworks displayed by price and size.

When Art Genome Project’s users encounter the interface, their perception and activity will be constrained by its physical and cognitive affordances. Compare to Artsy, Invaluable as the world’s largest online auction marketplace has a different image display design. When users click the categories, the total number of artwork will be displayed. Users can personalize how many artworks they prefer to see on each page and they are offered a clear vision of page number with a drop-down list design. Invaluable’s interface gives users a sense of scarcity and each piece of art is distinct from others. By one click of the “Humor” category for example, 4929 pieces of artwork are represented on a vertical plane of identity. When scrolling down to the bottom of the page, more images will show up and there seems will be infinite images, which all share the same characteristic. This design gives user the illusion of equality.

In addition, users can choose to “Follow” a category, artist and museum to collect their favorite genres in their personal page. By following them, users can share their collections via Facebook and Twitter, also they can browse and follow the content of other users’ main pages. By doing so, the users’ “art feed” displays personalized images. Pinterest is a photo sharing website which share the very similar design with The Art Genome Project’s. Users can upload and manage images—known as pins—through collections known as pinboards. What’s more, these two websites’ “like” and “follow” bottoms share similar styles. The design of “follow and like” function reminds users of photo sharing websites like Pinterest, and the inefficient design renders artworks on the websites to“delight of the eye”.The Art Genome Project’s interface creates the distraction that art works can are images, which can live without content. Users can appreciate and collect them base on their visual figures. Malraux describes the art museum saying, “The art museum, born when the easel-picture was the one living form of art, came to be a pageant not of color but of pictures; not of sculpture but of statues” (Malraux & Malraux, 1978). However,The Art Genome Project’s interface makes the artworks deteriorates to color.

In addition,The Art Genome Project’s users can sort the artwork displayed by price and size, which creates the impression that the products and services users can receive on the website are consumer goods. For consumer goods, buyers’ primary concerns are their budgets and they assume the products listed are interchangeable. is the world’s largest online retailer of posters for wall déco. Consumers are also allowed to sort posters listed by price and size, which perfectly fulfills their needs: find a poster to fill their empty wall with limited budget. The Art Genome Project’s sorting tool’s affordance indicates that the artworks are also interchangeable and are used for home decoration.

In conclusion, I have attempted here to critique the Art Genome Project, which is a generative classification system of Artsy. The system’s design philosophy seems reasonable based on C. S. Peirce’s Model of Semiosis, which highlights meaning making is a generative as well as infinite process. Notwithstanding, Art Genome Project’s design drives the users to focus on trivial “preferences” while ignoring the depth of structure in meaning making process, which makes it a poor well-organized digital museum for education. In addition, The Art Genome Project’s affordances make it an untrustworthy online auction platform.


“About | Artsy.” N.p., 2015. Web. 17 Apr. 2015.

Adam, Georgina. Big Bucks the Explosion of the Art Market in the 21st Century. Burlington: Lund Humphries, 2014. Library Catalog. Web. 16 Apr. 2015.

Atkin, Albert. “Peirce’s Theory of Signs.” The Stanford Encyclopedia of Philosophy. Ed. Edward N. Zalta. Summer 2013. N.p., 2013. Stanford Encyclopedia of Philosophy. Web. 18 Apr. 2015.

Boucher, Brian. “Assessing The Art World’s Pandora Goes Beta.” Art in America (2012): n. pag. Web.

Bourdieu, Pierre. “The Forms of Capital.” Handbook of Theory and Research for the Sociology of Education (1985): 242–258. Print.

Cameron, Fiona, and Sarah Kenderdine, eds. Theorizing Digital Cultural Heritage: A Critical Discourse. The MIT Press, 2007. CrossRef. Web. 20 Apr. 2015.

“Egon Schiele | Self Portrait with Raised Bared Shoulder (1912) | Artsy.” N.p., n.d. Web. 18 Apr. 2015.

Grampp, William D. Pricing the Priceless: Art, Artists and Economics. First Edition. New York: Basic Books, 1989. Print.

Griffith, Erin. “A Bold, New Direction for the Fine Art Industry: Online Sales.” Fortune (2015): n. pag. Web.

Irvine, Martin. “Semiosis and Cognitive Semiotics.” Google Docs. N.p., 2015. Web. 17 Apr. 2015.

“Keith Haring | Untitled (1982), Available for Sale | Artsy.” N.p., n.d. Web. 18 Apr. 2015.

“Lina Scheynius | Untitled (Diary) (2004), Available for Sale | Artsy.” N.p., n.d. Web. 18 Apr. 2015.

Malraux, André, and André Malraux. The Voices of Silence. Princeton, N.J: Princeton University Press, 1978. Print. Bollingen Series 24A.

Miller, M.H. “Point, Click, Collect: Brings the Art World Online… Again.” Observer. N.p., 4 June 2011. Web. 17 Apr. 2015.

Mufti, Shahan. “’s ‘Genome’ Predicts What Paintings You Will Like.” WIRED. N.p., 23 Nov. 2011. Web. 17 Apr. 2015.

Savan, David. An Introduction to C.S. Peirce’s Full System of Semeiotic. Toronto Semiotic Circle, 1988. Print.

Schafer, Jack, and Marvin Karlins. The Like Switch: An Ex-FBI Agent’s Guide to Influencing, Attracting, and Winning People Over. Simon and Schuster, 2015. Print.

Zhang, Jiajie, and Vimla L. Patel. “Distributed Cognition, Representation, and Affordance.” Pragmatics & Cognition 14 2 (2006): 333–341. Print.