“Tool Without a Handle” – Tools For Art and Politics, Part 1

“Tool Without a Handle” – Tools For Art and Politics, Part 1

This blog has, to date, primarily focused on the qualities of networked information technologies and regulatory responses to them – in particular qualities that raise issues of privacy and free expression.  This installment of “Tool Without a Handle” looks at the qualities that render these tools influential on artistic and political discourse.  In this first part, I will look at one particular quality: searchable text.

Several recent pieces have looked at digital tools through the lens of the arts:  Virginia Heffernan’s Magic and Loss: The Internet as Art flags many features of the Internet and their relationship to creativity, and the impact of that creativity on culture.  A New York Times article on the “Dark Web”[1] describes how artistic techniques are being used to illustrate online content moderation (a topic I’ve covered extensively in earlier blogs)[2] and to draw attention to parts of the Internet often unseen.  In this blog, I’ll note qualities of Internet tools that have changed the nature of art and political discourse it in ways that make discourse more subjective, more textual, and more fluid.

One reason to link artistic and political expression is that both combine the concrete and the abstract.  Art works with one foot in colors, images, textures, and various media – and another in perspective, emotion, or experience.  Even a realistic portrait represents something about humanity in general, and even a Jackson Pollock painting uses particular colors in a particular medium.

Art uses materials to express an opinion, an argument, an idea, or perhaps simply to evoke interest; to “let the viewer make of it what s/he will.”  Portraits can be literal or evoke a person (or, more accurately, a persona) without being a portrait (e.g., Rodin’s statue of Balzac).[3]  Art can be an abstracted image which evokes the idea of a specific being (e.g., Alexander Calder’s Eagle).[4]  And it can be expressly political or religious; in doing so, be as specific as a Guernica or as universal as Yoko Ono’s “Wish Tree.”[5]

Similarly, a political argument will build with the concrete:  economic figures, crime statistics or specific events, and link them to a narrative, an argument, an emotion.  It may attempt to craft a detailed picture of an issue (as is commonly done in regulatory filings and occasionally in Congressional testimony), to rouse action by highlighting selected facts, or simply to create emotion associated with an abstract concept (e.g., slogans such as “The Audacity of Hope” or “Make America Great Again”).

Digital technologies have been used to create art works for some time. For example, in 1974, Canadian artists incorporated telecommunications capabilities into an art exhibition called String Games (evocative of children’s cat’s cradle games).[6]   “Internet art” incorporates a variety of techniques, including web design, interactivity and other methods – many of which aim to alter boundaries between viewers and art objects (or, as technologists would put it, between users and code).[7]  Some Internet artists utilize not only the technologies but the cultural referents of the Internet as well – the wide variety of memes, .gifs, and design elements, to form artistic expressions.[8]

The use of digital information tools in politics extends back at least as far as 1952, when computers were used to calculate election outcomes[9] – but for examples in my lifetime I look to the 1992 campaign (on which I volunteered), with two examples I remember clearly:  the use of electronic systems (such as Lexis/Nexis) for research, and distribution of the Clinton Administration’s 1993 economic plan on diskette.

Even before widespread availability of the commercial Internet, digital tools allowed access to news of government activities, positions, and votes. Researchers who might otherwise have gone to the library for printed newspapers or copies of the Congressional Record could use a PC and a modem to acquire the same information from their own apartment (though superior air conditioning in the library still made visits appealing in summer).

In February 1993 (well before the launch of www.whitehouse.gov in September 1994), a searchable, easily portable copy of what was a voluminous printed text was made widely available – even individuals on the periphery of policymaking in Washington ended up with a copy.  This disk included complete text and tables of President Clinton's economic plan, "A Vision of Change for America," selected historical budget information, and details of spending and revenue options.[10]  A top issue of the day - the debate over taxes, spending and deficits - was being shaped by digital tools.

The distribution of the economic plan on diskette incorporates the concrete (figures, tables and projections) and also argument (position statements on spending priorities) and even atmospherics and emotion (use of the latest digital tools helped support efforts to portray the Clinton/Gore administration as smart, competent, and engaged in ways that – rightly or wrongly – the previous Bush administration was seen to lack).  The role of the Internet in politics today is, of course, extensive:  news and information sharing, fundraising, organizing, event planning – as well as media (in particular social media).  How digital technologies have affected politics is derived in part from important qualities of the tools themselves.  In this post, I start with one such quality:  searchable text.

Searchable Text

Prior to digital tools, most collections and libraries were organized by a code:  common codes being alphabetical by title, or by author, or by subject, or organized by subject and then alphabetical within that subject.  Once the code was understood, it became a matter of searching systematically per the code to locate a given work.  More detailed codes, such as the Dewey decimal system,[11] allowed more efficient searching for a specific work, though these required much organizational work up front to label the works (and re-stacking by librarians to maintain the catalogue according to the code).

None of these systems, though, allowed for instantaneous, targeted searching of specific words or phrases.  And the degree of searchability depended on the individual work itself; some works (such as the Bible) came with their own internal codes (title, chapter, verse and line).  And so, for these works, scholarly memorization would allow rapid location of a specific word or phrase.  Scholars also prepared concordances, linking important subjects and phrases to tables of chapter and verse:  a primitive form of hyperlinking.  Even so, none of these had the power of digital tools, including tools such as automated indices,[12] “engines” to search those indices,[13] Boolean search queries,[14] the organization of the “library” into linked files,[15] and the World Wide Web client-server architecture.[16]

In terms of consequences, it’s reasonable to hypothesize that when text and files are searchable, the reader or researcher’s agenda has far fewer obstacles.  If you know what you are looking for, digital tools improve the chances that you will find exactly that; there will be no hike through subject>card index>titles>chapters>book index that creates inefficiencies or digressions.  This is, on one hand, tremendously helpful.  Material can be found to match the artist’s mental vision.  Facts and quotes can be more easily verified when the precise source attribution or reference can be quickly found.  In the political context, this greatly speeds fact-checking, improves the quality of discussion, and enables journalists to better keep the public alert to exaggerations.

Yet it may also contribute to confirmation bias[17] and the “filter bubble” phenomena. [18]  A pre-conceived preference as to what news or information is desired can be easily bolstered with references found via targeted searches.  If one believes, for examples, that the majority of immigrants are prone to commit crimes, search tools can quickly find references that support that proposition (however well-reasoned or accurate those references may be).

Fortunately, most well-designed search tools are ruthlessly democratic and thus the top results for any one query are likely to be not only the most relevant but the most often referenced, and searchers looking for validation of erroneous beliefs are most likely to first encounter search results pages pointing out those errors.[19]  At least one study has shown that confirmation bias can be attenuated when challenging information is presented in search results, though the study also surmises that indirect methods (such as tag clouds) are more effective than direct methods.[20]

Nonetheless, searchable text still yields rapid accessibility to written materials, much of it framed as journalism or academic studies, supporting wildly inaccurate or improbable propositions.  A post by Ethan Zuckerman[21] refers, in turn, to the “three spheres” theory of Daniel Hallin:  1) areas of discussion that are widely accepted (“slavery is bad,”), 2) areas that are legitimately disputed (tax policy, abortion, gun control), and 3) areas of “deviance.” [22]

It is this third sphere into which I would put fantastical conspiracy theories such as “Donald Trump is an Alien” and “Hillary Clinton Murdered Vince Foster.”  In part due to searchability of web pages (and text on web pages), these theories are afforded more exposure, traction and political impact than would ever be the case in an entirely curated media.  Figure 1 below shows the number of search results for three fantastical propositions, for two major search indices for searches conducted August 14, 2016:

This is an unrepresentative sample, to be sure, but it illustrates what I believe is a reasonable hypothesis:  that the ability to query vast libraries of text has given greater purchase to those who traffic in what are fairly called “deviant” political theories,[24] which would otherwise have been less discussed and less trafficked where information was only available through curation and catalogues.

In addition to the volume of results, it’s also notable that different indices yield very different results, meaning that a searchable Web is, of course, subject to the preferences of the search tool.  The topic of search algorithms is beyond the scope of this post, but it is notable that Google returns far fewer results for these queries than does Bing.  This does raise a discussion point as to whether search should be more selective or more inclusive, given what we know about confirmation bias, particularly with respect to content falling within the “deviant” of the 3 Hallin “discussion spheres.”[25]

This sort of design choice is, in part, what makes this quality of digital tools relevant to discussion of the arts as well:  both artistic expression and the art (design) of the Web itself.  In Heffernan’s Magic and Loss, (the text of which, sadly, is not searchable on Amazon’s Cloud Reader), she makes the observation that the Web has become a “teeming, sprawling, commercial metropolis…so crammed with links, graphics, ads, and tarty bids for attention that they’re frightening to behold.”[26]  In part because searchability gave control to the user to select the desired content, the aesthetic incentives of the Web are now all skewed towards capturing and recording the user’s choice.  To modify Andy Warhol’s proposition, it is the future and online we are all now famous, with crowds clamoring for our attention.

At the same time, the impact of search on creative expression, then, is in many respects to open opportunities for non-conventional forms of expression.  “Design” is almost an inapposite term for the collaborative, anarchic, unplanned nature of the Web, but this has allowed for myriad positive innovations.  These include outlets for literary styles such as “fan fiction” who have niche audiences, the design form of the tweet or status post, photo collages (e.g., Tumblr, Pinterest), and the ubiquitous video monograph on every imaginable topic (including useful how-to videos on topics from makeup to plumbing).  One of the distinct values of search technology is that it has enabled a vast number of entertainment and creative works to find audiences.[27]  A library of programming the size of Netflix or YouTube is simply not navigable with the rotary dials that graced the television set of my childhood.

In a future installment, I’ll look at additional qualities of digital tools that have helped make political and artistic expression more subjective, accessible and fluid:  including portability of expressive content and the rise of visual content relative to the written word.  The example of searchable text, though, serves reasonably well to illustrate the core observation:  as expressed by a McKinsey report, search tools “shift the balance to empower individuals or small organizations with something to share that would otherwise reach only a small audience.”

[1]http://www.nytimes.com/2016/06/26/arts/design/illuminating-the-dark-web-and-content-monitoring.html?_r=0

[2]See, e.g., “Tool Without a Handle:  Kittens, Cities, and Creepshots,” http://cyberlaw.stanford.edu/blog/2012/12/tool-without-handle-%E2%80%9Ckittens-cities-and-creepshots%E2%80%9D and “Tool Without a Handle:  Tools for Terror, Tools for Peace,” http://cyberlaw.stanford.edu/blog/2016/03/tool-without-handle-tools-terror-tools-peace

[3]See http://www.musee-rodin.fr/en/collections/sculptures/monument-balzac

[4]http://bit.ly/2a12ozC

[5]https://en.wikipedia.org/wiki/Wish_Tree_for_Washington,_DC

[6]http://www.fondation-langlois.org/html/e/page.php?NumPage=1974

[7]/content/files/2012/11/cox-and-juliff_the-post-display-condition-of-contemporary-computer-art.pdf

[8]http://www.complex.com/style/2013/04/25-internet-artists-you-need-to-know/michael-manning

[9]http://www.npr.org/sections/alltechconsidered/2012/10/31/163951263/the-night-a-computer-predicted-the-next-president

[10]This being digital information, after all, it’s now available online at: http://webapp1.dlib.indiana.edu/virtual_disk_library/index.cgi/1249316/FID3039/CLINTON

[11]Online at: http://www.gutenberg.org/files/12513/12513-h/12513-h.htm

[12]See “Crawling & Indexing,” an online tutorial from Google: https://www.google.com/insidesearch/howsearchworks/crawling-indexing.html

[13]“Archie,” considered the first search engine, was described as “a collection of resource discovery tools that together provide an electronic directory service for locating information in an Internet environment.”  http://www1.chapman.edu/gopher-data/archives/Internet%20Information/whatis.archie

[14]Nearly everyone is familiar with this approach to search, which is a type of search allowing users to combine keywords with operators such as AND, NOT and OR to further produce more relevant results.  It is named for the mathematician George Boole, who made significant contributions to algebraic logic and laid important foundations for modern information services.  Learning to make efficient Boolean queries turned out to be the most important research skill I learned in the law library of the early 90’s.

[15]https://en.wikipedia.org/wiki/Timeline_of_hypertext_technology

[16]http://webfoundation.org/about/vision/history-of-the-web/

[17]See, e.g., “Confirmation Bias Shapes How We Read Online,” The Atlantic, Jan. 2012, http://theatln.tc/2bv68j8 (citing: Yoon, Y., Sarial-Abi, G., & Gürhan-Canli, Z. (2012), “Effect of Regulatory Focus on Selective Information Processing,” Journal of Consumer Research, 39(1), 93-110. doi:1; online at: http://www.jstor.org/stable/10.1086/661935); “Bing Your Brain: Confirmation Bias and Branding,” Bing Blog, 23 April 2013, online at: http://blogs.bing.com/search/2013/04/23/bing-your-brain-confirmation-bias-and-branding/

[18]https://www.ted.com/talks/eli_pariser_beware_online_filter_bubbles

[19]See, e.g., Google search results for the query “most immigrants are criminals” http://bit.ly/2bfQfLV

[20]Schweiger S, Oeberst A, Cress U, “Confirmation Bias in Web-Based Search: A Randomized Online Study on the Effects of Expert Information and Social Tags on Information Search and Evaluation,” J Med Internet Res 2014, online at: http://www.jmir.org/2014/3/e94/

[21]http://www.ethanzuckerman.com/blog/2011/04/25/overcoming-political-polarization-but-not-through-facts/  Hallin gives the example of “pro-pedophilia” content to indicate the category of “deviant” content; I know from first-hand conversations with staff of the National Center for Missing and Exploited Children (“NCMEC”) that indeed, digital technologies have made such deviant content (including illegal abuse content) far more available.

[22]Daniel C. Hallin, The Uncensored War: The Media and the Vietnam (1986), p.117; online at:  http://bit.ly/2aWJ8JU

[23]On the first page of Google results (as of August 14, 2016), 7 of the 10 results displayed link to a story that, in turn, claims half of Georgia voters believe Hillary Clinton did murder Vince Foster.  http://bit.ly/2bsnagx. This extreme figure (50% of Georgia voters) is itself subject to some skepticism, and highlights in turn the way certain “data” can be fueled by loaded questions and biased polling techniques, such as may have been applied to this survey.  See http://www.thedailybeast.com/articles/2016/06/01/half-of-georgia-trump-voters-think-hillary-murdered-vince-foster-poll-trolls-find.html.  Searchable text can thus give prominence to both dubious theories and dubious accusations as to the extent to which particular groups support such theories; in effect unfairly stereotyping certain groups as “deviant.”

[24]I take to heart Zuckerman’s diagnosis that a challenge for contemporary discourse is the shifting boundaries between spheres of legitimate discourse and ‘deviancy.’ See n.18, supra.  For these 3 examples, though, both the facts (in the case of Foster’s death, it was reviewed extensively by several independent fact-finders) and the preponderance of public opinion are still such that these propositions are fairly beyond legitimate discourse.

[25]Google has introduced algorithmic changes to favor “high-value” content:  pages with greater numbers of links to ‘credible’ sources. See http://searchengineland.com/library/google/google-panda-update.  To the extent this simply limits “scraper sites” – who have little or no original content – this is all to the good as it reduces clutter without reducing discoverability of any given content; to the extent it involves Google decision-making as to what constitutes a ‘credible’ source or “quality” page there is room for much discussion.  See http://www.seo-theory.com/google-panda/ for a good overview of Panda-related issues.

[26]Magic and Loss, Chapter 1 (“Design”), Section on “Sprawl.”

[27]Search technology is thus not only a driver of expressive opportunity but economic opportunity – including for artistic and creative producers.  A McKinsey study on the value of search calculated that, in 2009, search capabilities had a measurable impact approaching gross annual value of $780 billion. See /content/files/2bgmimd.pdf