The Art Genome Project's Roots

We cast our scholarly net wide. Our research brought us to the 18th century and back again. We looked at thesauri, encyclopedias, and some of the first art history books ever printed. We were consistently amazed at the variety and volume of attempts that have been made recording, cataloging, classifying, and sharing art throughout history. We developed a new appreciation for engravers who trekked to Greece and Italy to painstakingly document antiquity’s artistic treasures, well before the invention of photography. We thought about polyhierarchies a lot; probably too much. It turns out that The Art Genome is related to a lot of things (and yet is exactly like none of them) all at once… 

Yes, in very broad strokes. Both are combinatorial frameworks of attributes.No, more specifically. The Art Genome has roughly 800 unique genes. The Human Genome has between 20,000 and 25,000 genes, though it is comprised of roughly 97% junk or redundant/repetitive DNA. The Art Genome is also descriptive and inherently incomplete while the Human Genome is generative and complete. 

Yes, since both aim to create comprehensive (though by no means exhaustive) analyses of types of art by identifying a set of criteria (which both call “genes”). These are both broadly applicable to their respective art forms, as well as useful for generating interesting connections for users. No, because while The Art Genome is currently one extensive list of genes for all works of art, Pandora has separate genomes (lists of genes) for each genre of music. Also, musicology generally studies quite different attributes than art and art history.  

Yes, in part, because The Art Genome, like a taxonomy, classifies things to show how they are related to each other. Biological classification (i.e. genus, species) is probably the best-known form of taxonomy. Library catalogues are also taxonomic. Taxonomies are useful to us because they help people browse, and browsing is a significant goal for The Art Genome. No, because taxonomies are monohierarchical or pyramidal, like the structure of a tree (where there can be many branches but they all flow back into the trunk), whereas Art.sy’s genome is polyhierarchical. What does polyhierarchical mean? In short, there are many relationships: many trunks and many branches. What this means in terms of the genome is that the characteristics of art can never be simplified into a basic hierarchical structure. Some characteristics relate so superficially or almost not at all (such asRomanticism and Kinetic Sculpture) and that’s OK. 

Yes, because it is a compendium of terms with verifiable definitions agreed upon by committee.No, because encyclopedias include mostly facts - knowledge about specific people and places. The Art Genome on the other hand contains almost entirely “generic” terms, or terms which are not proper nouns, people, or places, such as “Painting” or “War,” and these can potentially be applied to any artwork on the site.   

Yes, because like a thesaurus, The Art Genome includes only generic terms.No, for two reasons. First, The Art Genome provides definitions of its terms; and second, thesauri group words together based on similarity of meaning to help the user retrieve the specific term they need for a particular context while The Art Genome is not focused on retrieving similar terms but on browsing and making connections between artworks. Stepping back now from this series of questions and comparisons, which show how The Art Genome is related to many preceding classification systems but is not identifical to any of them, the next question is, so what? Put simply, this project allowed us to understand the limits and capacities of the genome. Or put another way, it provided those of us working on The Art Genome with a well of historical inspiration and locations for advice.

-Matthew Israel, Director of The Art Genome Project, and Jessica Backus, Researcher