The phenomenon: Errant commenters on blogs In October of 2003, I posted some notes to my personal blog about how I was settling in to my first few months of Teachers College, along with a few pieces of cultural detritus that had recently captured my attention. One was a Glamour Magazine call for men to participate in a hair-makeover article. I had passed the call along to an angry punk friend of mine who was bald, but never heard back as to whether Glamour had called on him to participate. “Maybe his hair isn't ‘floppy’ enough,” I wrote in the blog post. “Oh, lord, all that suggests to me is more Ashton Kutchers.” Two months later, the first comment was posted to those notes: hey Ashton we think you r sooo hott!!! hehe!! We r from Tennessee but Maggie moved to North Carolina. we want to tell you are number but we dont know if you are really ashton. but we still love you. We love your show punk'd and we think your a great actor. We want you to talk to us online except for shybaby might not be on that much because her Dad took it away. We cant wait to see your movie cheaoer by the dozen! remember we love you and we think your $exi. love, maggie, and MaryAnne It was quickly followed with: is this ashon we dont know so if its not and u know him tell him we love him give him our names and our email addresses . So we didnt know because we read some other peples after we wrote it so we r writing this!! Buhbye Posted by: same people that rote the hott thing to u at December 28, 2003 12:02 PM It appeared that two young girls, ostensibly from North Carolina and Tennessee, thought that my website was actually Ashton Kutcher’s – or else were amusing themselves by pretending it was. These comments were far from unique. The post garnered a number of other comments from people who also apparently thought the site belonged to Ashton Kutcher. A few of my other posts also began to gather comments; a post where I’d mocked Mary Kate and Ashley Olsen gathered demands for nude pictures of them, and a post in which I had briefly described a day on which my third-grade students talked about a dance called the Crip Walk collected dozens if not hundreds of requests for lessons on how to do the dance – as well as comments from self-proclaimed Crip gang members who warned that non-gangsters should not do the dance. A friend’s site gathered requests for answers to riddles, and angry defenses of singer Celine Dion. As I began to ask around, it grew clear we were not alone. The vaunted community weblog MetaFilter has at least three threads (one of them quite long) featuring instances where blogs, consumer review sites, guestbooks, and other sites which permit feedback have accumulated comments unrelated to their original posts, many of them from searchers looking for celebrities or answers to questions. Many bloggers I speak to can identify a page on their site or someone else’s where comments like these have been posted. Software developers and webmasters I have spoken to confirm that many text boxes, not just blog comment fields, elicit similar responses. One developer who worked for Blogger summed it up thus: “Give people a text box, and they’ll put absolutely anything in it.” In December of 2007, I began a blog to keep track of the instances of these happenings, which, in shorthand, I’ve been calling the Gumbaby phenomenon.1 What is the source of this phenomenon? At what point do these commenters fail to understand the websites they are reading? Or is it rather a misunderstanding of how the Internet works? How do people rationalize commenting on web pages they do not appear to have read? These are the questions I hope to answer in this study. Introduction: Problematizing search Judging from the long threads on MetaFilter, and the exclamations of recognition uttered by other bloggers and software developers when I bring up this topic, there appear to be a large number of Internet users who do not read search results and web pages the way one might expect them to. Regular readers of MetaFilter visiting the errant comment threads asked a number of questions which might be voiced by concerned educators: Can’t these people read? Why are they commenting when this is clearly not the site they’re looking for? Do they actually think they’re writing to celebrities? How did they get to this website? (In a few cases, readers answered the last question using referral logs and experiments with Google: a webcomic about human-computer interaction turned out to be the #1 result for “cancel Google,” with the algorithm paying attention to the site’s title, OK/Cancel.) The entire picture of why these sites are garnering such off-base comments is not clear yet; this study aims to shed more light on the phenomenon. 1 In the legends of West Africa, the Gum Baby was a device used by the trickster hero Anansi to catch a fairy, Mwatiya. Anansi carved the figure of a child, covered it in sap, left it by the side of the road — and let the fairy’s misunderstanding of the sticky figure’s silence do the rest of the work. The fairy, assuming the gumbaby was alive, greeted it. Receiving no response, the fairy got angry. Did nobody teach you manners? the fairy said. Do you think you’re too good to talk to me? Still getting no response, the fairy hit the gumbaby, and stuck to the sap. How dare you grab me! said the fairy. She hit the gumbaby again — then kicked it — and was firmly stuck. Anansi took the captured fairy to the Sky God and claimed his reward: the title of King of All Stories, and a golden box containing every story known to humankind. This legend is an ancestor of the story of Br’er Rabbit and the Tar Baby, in which roles are reversed and the trickster Rabbit is trapped by Br’er Fox. The Disney movie Song of the South is perhaps where most Americans know of the Br’er Rabbit tale. Through that portrayal, the term “Tar Baby” has unfortunately taken on a negative racial tone. Seeking to avoid the twisting of this tale that occurred through the legacies of slavery, which tends to be associated with the term “tar baby,” the name of this site returns to the roots of the story to retrieve the original gist of the metaphor. It could be argued that users’ misreading or misunderstanding of search results is largely responsible for large quantities of spurious information on the Internet. Advertisers looking to profit off prescription drugs, travel deals, insurance, pornography, and other big sellers deluge blogs and other feedback mechanisms with spam links, looking to change their standing in Google’s PageRank and other engines’ rankings so that their clients’ sites will be listed more highly in search results. An entire industry -- search engine optimization -- has arisen to provide this service. While the success of the optimizing industry relies on “gaming” search engines and developing networks of pages, links, and spam which will drive traffic to the desired pages, it also depends on searchers who click through high ranking results which do not have the information they are seeking. Guiding searchers through redirects, dummy search engines, and business websites ostensibly helps drum up sales for businesses which choose to advertise this way. Driving this traffic is far easier when users do not stop to assess the quality of the links and pages they click through. For the most part, search engine users do not understand that this is going on.The Pew Internet and American Life Project found that only 62% of its national sample were aware of the difference between search results which appeared because of advertising placement and those which were just ranked by algorithm. Only 18% of their overall respondants thought they could tell the difference between advertising results and unpaid results. (“Search Engine Users,” 2005) It should be noted that simple pay-for-play search manipulations are much more obvious in their appearance than optimized results, as search engine companies are under obligation to visually distinguish between paid and unpaid results. If roughly two thirds of searchers are not aware of pay-for-play, one wonders how many are aware of manipulations (either benign or malign) at the algorithm level, like optimization and Google PageRank. Speaking to this issue, 68% of the Pew study’s respondants thought “search engines are a fair and unbiased source of information.” 57% of their respondants were not aware that a number of search engine companies had considered or experimented with tracking user behavior for the sake of improving search results. (“Search Engine Users,” Fallows, 2005) The Pew study found differences among populations of searchers based on whether or not they thought search engines were “fair and unbalanced.” Those who did believe in the fairness and balance of search engines were found to be more likely than those who didn’t to use just one search engine, were much (about 20%) less likely to know about paid placement in search results, and were also more likely to feel satisfied with the results their searches generally returned. (“Search Engine Users,” Fallows, 2005) The Pew study found that only 7% of their population reported making use of more than 3 search engines; about half (48%) used two or three, while almost as many (44%) used only one search engine. (Fallows 2005) Similar findings about monolithic use of search engines has been echoed elsewhere. (Graham and Metaxas, 2003) An interesting and largely unnoted finding of the Pew study was that some 35% of those who have had no more than one to two years using the Internet claim that they have not made use of search engines. This drops off as they gain more experience online, with 27% of those between 2-3 years online, 21% of those with 4-5 years online, and 10% of those having spent 6 or more years online claiming they have not made use of search engines. (Fallows 2005) It might be worth investigating how many of these Internet users have used search engines without realizing it – indicating a lack of basic comprehension of web elements. Pew does not look into this question. To date, educators and librarians have generally been less concerned about the roles of search engine optimization, spam, and paid placement than they have been concerned that searchers will place faith in parody websites (Leu et al 2007), political or racist propaganda (Graham and Metaxas, 2003), misinformation on Wikipedia (NEED CITE), and other less ephemeral, more concerted web-based efforts to disseminate information they deem “wrong.” These warning calls do share the same concern with searchers’ poor development of search queries and haphazard approach to selecting search results. Henry (2007) found that both low- and high-income middle-school students – and the teachers who taught them -- performed poorly on tests of reading for “accuracy and bias” on the Internet, though high-income students and teachers performed better on tests of critical evaluation of information they found there. A number of studies have found that users, particularly younger users, settle on the first result returned by a search engine and either do not bother to look at the rest of their results or go through them dogmatically without a thoughtful plan (Graham and Metaxas 2003; Guinee, Eagleton, & Hall, 2003, cited in Henry, 2006; OTHER CITES) This poor use of search results seems to be evident in another Internet artifact: the user search records that AOL released to the public (and subsequently withdrew) in 2006. In records of the queries made by an AOL user assigned the number 711391, one finds that the sites where this user ended up frequently seem unrelated or only peripherally related to the words in the query: Query friends online can be different in person women with curvy bodies anti anxiety drugs tips for healing blisters top alaska attractions what is middle aged spread Landed on site www.salon.com, an Internet magazine groups.teenhelp.org, an advice site for teenagers www.newyorkmetro.com, which redirects to the website for New York Magazine www.beliefnet.com, a Christian website www.funnewjersey.com, a tourism website for New Jersey www.cartoonstock.com, a database of cartoons Ultimately, though, we do not know what AOL user 711391 did with his/her results. Did (s)he ultimately feel (s)he found what (s)he was looking for? (S)he often clicked through multiple results for the same search. Was (s)he merely enjoying the serendipity of finding a trove of cartoons when (s)he started out looking for “middle aged spread?”2 While the form of gumbaby comments seems to confirm that a superficial, uncomprehending search style is common to many Internet users, the subject matter of many gumbabies – celebrities -- also follows common patterns of search. In a 2001 analysis of search queries made to Excite.com, the word “celebrity” was the 50th most commonly entered search term (and 11 of the top 50 terms were prepositions, conjunctions, modifying symbols, and other very common words). (Jansen et al 2000; Spink et al 2001) In a follow up study, the word “star” was number 75, and terms related to “princess diana” (who had died shortly before the date the data was collected) were also in the top 75 (again, with a number of small words preceding it in the listing). (Spink et al 2001) This paper notes that another study of Alta Vista search data turned up two celebrity phrases (“pamela anderson” and “spice girls”) in their top 25 queries. (Silverstein et al 1999, cited in Spink et al 2001) Celebrity names routinely rank among the most popular search queries; the Pew study notes that Pamela Anderson and Britney Spears had been in Lycos’s top 50 searches for 277 weeks running. (Fallows, 2005) Google, meanwhile, has documented close to 600 misspellings of Britney Spears’s name in queries made to their search engine (with the top three misspellings being entered over 100,000 times)3, so it is possible that estimations of query popularity which only take well-spelled queries into account drastically underestimate how popular these searches are. Which celebrities appear among the most popular searches changes over time, but on any given day it is a relative certainty that some celebrity name will appear in a given search engine’s top-ranked phrases. The Gumbaby phenomenon lends further credence to researchers’ findings that users unthinkingly click on links, and that they maintain faith in results which may be biased by spam or payola. Watching these comment threads unfold, one might feel inclined to embark on an emergency campaign to develop better digital literacy instruction. However, some lines of thought in the field of literacy education suggest it is better not to go off half-cocked. The problem with problematizing digital literacies It would be easy to look at the gumbaby phenomenon and conclude that the commenters are misguided illiterates who need to be trained to read the Internet appropriately. But 2 AOL took down the records it erroneously posted, but they remain in circulation on the web and are still retrievable by way of the Wayback Machine: for this user, http://web.archive.org/web/20060815121427/http://czern.homeip.net/aolsearch/search2.p hp?text=711391 3 http://www.google.com/jobs/britney.html these people are still navigating the ‘net. They pass through search engines, typing queries which find the names of celebrities and software programs; they click links; they successfully post comments using blog software. These are, ultimately, still literacy practices. And B.V. Street, a formative thinker in the field of New Literacies, has demonstrated that to be so quick to dismiss a group’s literacy practices interferes with formal attempts to change those practices. (Street, 1995) Street has delineated problems which arise when those aiming to foster literacy in an "illiterate" population ignore the reading and writing practices in which the "illiterates" already engage. Street describes top-down literacy instructional programs as essentially colonialist, enforcing the use of particular reading and writing skills to support the development of particular kinds of labor, social or government organization, or religion. The effect of such programs in practice is generally not what its developers intend. Rather than adopting the practices taught in these literacy programs whole-cloth, the population being taught is more likely to pick and choose which kinds of reading and writing best fit its existing literacy practices, abandoning those which do not fit their existing structure. Street gives numerous historical examples of this selective adoption, including Fijians’ adaptation of Christian missionary reading and writing practices into the development of cargo cults; urban Iranians’ use of bland textbook facts as a shibboleth to distinguish themselves as more rational than “backwards” villagers; and the adoption of writing as a means for previously forbidden emotional expression on an island in the South Pacific. Street notes that frequently these repurposings of literacy involved manipulation of the “paralinguistic and pragmatic features” of texts, “their formal appearance, decoration, covers, etc.,” as well as working with syntax and word meaning. (p 92, Street, 1995) The result is practices which do not match the ones intended by the literacy teachers. Scholars following in the footsteps of Street similarly find ways in which educators’ ignorance of “native” literacy practices proves to be the undoing of school-based literacy activities. Margaret Finders documents a similar conflict in a junior high writing classroom in the US. Unlike Street, the “native” literacy practices she focuses on are not institutionalized as formally as they would be in masjids or cargo cults; they consist of junior high girls’ passing notes, reading magazines together, signing yearbooks, and writing grafitti on bathroom walls. Finders makes a convincing case that girls develop identities in these unrecognized literacy practices which they use to undermine the goals of the teacher as she invites students to share their writing in a “safe” space. (Finders, 1997) Gee (CITING HALLIDAY AND MARTIN, 1993 – FIND) has developed some hypotheses about why, on an individual level, students sometimes refuse to pick up particular literacy practices. He describes modern school-related literacy practices as demanding that participants sacrifice subjective aesthetic appreciation, empathy, and a willingness to deal with ambiguity, and posits that this sacrifice may not seem sustainable to many learners: The crucial question then is this: “Why would anyone – most especially a child in school – accept this loss?” My view is that people will accept this loss only if they see the gain as a gain.... People can only see a new specialist language as a gain if: (a) they recognize and understand the sorts of socially situated identities and activities that recruit the specialist language; (b) they value these identities and activities, or at least understand why they are valued; and (c) they believe they (will) have access to these identities and activities, or at least (will) have access to meaningful (perhaps simulated) versions of them.... Thus acquisition is heavily tied at the outset to identity issues. It is tied to the learner’s willingness and trust to leave (for a time and place) the “everyday” world and participate in another identity – one that for everyone represents a certain loss. For some people it represents a more significant loss in terms of a disassociation from, and even opposition to, their home- and community-based cultures. [emphases in original, Gee 2004 p 93-94] If existing literacy practices -- even marginalized or stigmatized ones such as passing notes and writing grafitti – are implicated in the failings of literacy instruction; if they are closely tied in enough with students’ identities that giving them up represents an unsustainable sacrifice, it behooves scholars interested in teaching literacy to understand why. It behooves digital literacy scholars in particular to understand students’ existing practices in the terra incognita these scholars are currently approaching. (WORDS FROM STREET TO THAT EFFECT) If we follow Street, Gee, and Finders’s reasoning, developing a description of users’ native, everyday Internet literacy practices is necessary as a foundation for information literacy programs, if we want them to have a concrete effect on the populations we wish to reach. To proceed without such an understanding, to proceed under the assumption that gumbaby commenters are simply “illiterate” -- or even to proceed as if the information literacy practices of educators align perfectly with the infrastructure of the Internet-- is likely to result in the development of information literacy programs which the target population might regard as alien, condescending, or useless, or from which they pick and choose skills as they fit into what they already do. My project at hand is to observe the literacy practices of those who comment on gumbabies – and, possibly, the literacy practices implied by the software (search engines and blog software) with which they come into contact. The actual development of literacy curricula which might be based on these findings is another project, beyond the scope of this study. Studies on Internet literacies Leu et al (2007) at the TICA (Teaching Internet Comprehension to Adolescents) project ran a study which tackled similar concerns to the present study’s, though with an approach much more oriented to developing effective school curricula. The study sought to identify “which skills and strategies appear to be important for successful online reading comprehension.” (Leu et al, 2007) To investigate this, they had 53 seventh-grade students perform web searches on topics selected by the students as well as by the researchers. Students alternately talked aloud, were prompted to talk aloud by researchers, or reviewed recordings of their searches to explain their thinking afterwards. The most striking finding was student interpretations of a website on the researcherdefined search task, a joke website which claimed that a fictitious cephalopod, the Pacific Northwest Tree Octopus, was endangered, and called for assistance in saving it. Only six of their students took this site as anything but literally true; the rest of the students diligently researched the octopus as if it existed, and reported back as requested by the researchers. The six students who questioned the validity of the site had specifically been shown this site by a school librarian, who used it to demonstrate the untrustworthiness of information on the Internet. Leu et al concluded that locating and “critically evaluating” information were two “circuitbreaker” skills, whose absence from a student’s search process could hamper their success in finding the right information online. Like (SO AND SO’S MODEL) in information retrieval, they concluded that “critical evaluation” could take place at any of a number of steps in the process – choosing a search string, looking at pages, summarizing their information for others, and so forth. Leu et al labelled students who rejected the validity of the Tree Octopus website “more successful with the evaluation of online information,” but ultimately concluded these students did not employ any strategy to critically evaluate the reliability of this site nor did [they] provide evidence from the Internet to support his evaluation[; they] only relied upon a previous classroom experience, where [they] had been told that the information was false. (Leu et al 2007) Saying these students “did not employ any strategy” is perhaps a bit facile. Research on information retrieval suggests that relying on one’s own past experience is in fact the most common informational strategy, and around half of the adult population will rely on a friend or neighbor, or a professional, when looking for information. (Julien 1998, 1999 in Case 2002; Chen and Hernan 1982, Dervin et al DATE in Marchionini 1991) Leu et al are thus left with a problem which they do not delve into in this paper. The six students who rejected the Tree Octopus site as spurious produced an evaluation of the site which the researchers appear to approve of; and yet, they did not perform the behavior the researchers wanted lead to this evaluation. This raises a question also left unanswered by the paper: What do Leu et al mean when they say “critical evaluation skills”? (DID THEY SAY IT ELSEWHERE) What do these encompass? Are they just specific to the Internet, or are they supposed to be generic, domain- and Discourse-agnostic, like traditional conceptions of literacy skills? In a separate article, Coiro, one of the researchers involved in TICA, described skills that would ostensibly fall into the category of critical evaluation. (Coiro, 2003) She mentions a couple of websites which offer teachers resources to help students “validate online information and to recognize commercial propaganda and bias,” as well as “key habits of mind” described by Brunner and Tally (1999). These include asking the following questions: What particular perspective of reality is represented? What explicit or hidden values underlie this text? What media conventions are used in this text and how do they shape the way the information is interpreted? Who is the intended audience and how might different audiences interpret the text? Who owns the text and who benefits from it? (Brunner 1999, via Coiro 2003) These are no doubt useful questions to ask. There is, in fact, something approaching consensus in the media literacy education community – a community which is not generally prone to consensus (Hobbs 1998) – that questions similar to these are a bare minimum for students to ask as they approach media texts. (Hobbs 1998) (ADD: FROM BUCKINGHAM) These questions are, however, not simply matters of skill. Many of them appeal to a body of factual, contextual knowledge. “Who owns the text?” and “what particular perspective is represented?” suggest that the viewer has some political, economic, or social knowledge which can be applied to the text. To ask about “explicit or hidden values” and “perspectives on reality” again suggests political, religious, cultural, or even scientific ways of thinking about different information. And questions about “intended audience” – an implied monster-in-the-closet in a number of popular press articles on youth merrily relinquishing their privacy on YouTube or MySpace – might even imply a worldliness encompassing other cultures and age groups. In order to understand texts through these lenses, students will have to be exposed to what Jim Gee calls Discourses or domains, and will need to have mentally run through “simulations” or models of the text content and how it fits in among other texts. (Gee, 2004) A problem with “internet literacy:” domain knowledge, or skills? Street emphasizes that a group’s acquisition of technical “literacy skills” does not inherently imply that the group has taken on the attitudes which a teacher or other instructor wishes them to learn. A major failing of some of the research on literacy practices, Street writes, is its tendency to conflate technical skills with ways of thinking or being social: Exponents of the ‘autonomous’ model of literacy, as we have seen, have attempted to treat literacy as an independent variable, supposedly detached from its social context, and then to ‘read off’ its consequences.... The major ‘consequence’ of literacy has, then, been that it ‘facilitates’ logic, rationality, objectivity and rational thinking. (cf. Goody, 1968, 1977; Ong, 1977, 1982a and b; Olson, 1977) (Street 1995, p 76) Street’s examples of Fijian, Iranian, and Melanesian repurposing of reading and writing demonstrate the ease with which the techniques of literacy can be separated from the ways of thinking and being with which they are often associated, and the content they are often used to convey. In the field of Library and Information Studies, concerns similar to Street’s are voiced by Tuominen, Savolainen, and Talja (2005). These researchers share Street’s attention to the ways that technical skills and content knowledge are being mistaken for each other, and the sensitivity to top-down prescriptions for literacy uses. Reviewing literature in their field Tuominen et al point out that, like the literature on literacy in schools, much of the literature in LIS has been prescriptive, and does not sufficiently consider information-seeking which does not fit a library model – particularly interpersonal information seeking or interaction with technology. Inspecting standards for information literacy, Tuominen et al find them written as if such literacy skills are “a cluster of abilities that 'are common to all disciplines, to all learning environments and to all levels of education.'" (p 333) Additionally, they find in these standards a suggestion that learners should be able to construct meaning entirely on their own – without interaction with others, without input from media. A key concern of Tuominen et al is that such constructions of “reading” information ignore the production of knowledge by readers. Tuominen et al conclude by calling for instruction in information literacy which is based on research into daily tasks and situations within specific communities of practice. The goal, they say, should be “enabling groups and communities to cultivate existing information strategies and about supporting them in their interactions with information technologies." (p 340-341) They suggest support should be given so that users can not only evaluate information in the specific domains they are engaging with, but also evaluate the domain of the technology being used to search, considering how it is constructed, its underlying values, and its context. A major proponent of the distinction between literacy practices in different domains is James Gee. In Situated Language and Learning, Gee distinguishes between “vernacular varieties” and “specialist varieties” of a given language. (Gee, 2004) Everyday language falls under the category of vernacular dialects, while specialist varieties are used by specific communities in working with particular subject matter in depth. Gee emphasizes that both kinds of dialects are closely related to distinct ways of seeing, relating to, and interacting with the world; elsewhere, he has called these ways of being or “playing the game” “Discourses.” (Gee 1996, Gee 2003, Gee 2004) He claims that “the vast majority of texts in the modern world are not written in the vernacular, but in some specialist variety of language,” requiring familiarity with these ways of being in order to succeed in a given field. (Gee 2004, p 17) This further supports the idea that “Internet literacy” as posited by the TICA group scholars would need to involve familiarity with a range of knowledge domains. If we consider that Internet knowledge is its own domain, some of the findings of Henry (2007) about correlation of Internet literacy skill and time on task stand to reason. Her study demonstrated that amount of Internet use appeared to correlate with Internet literacy skill level. From this finding, we might guess that the additional time successful Internet spend on task is spent adding to their knowledge of the domain of Internet technology. She suggests that this contributes to the Digital Divide, with both students and teachers who have less Internet access both in and out of school less proficient at reading online. It follows from these calls to attend to domain-related literacy skills that if we want to know about the apparent failures of readers online, it might behoove us to look at their Internet knowledge as a domain unto itself, while also looking at their familiarity with the domain implied or sought by a particular search. It may also be important to attend to the possibility that there are both vernacular and specialist varieties of Internet knowledge, and these may be different and possibly even conflicting. To better understand Internet fluency, this study will pay attention to Internet knowledge as its own domain, as well as content knowledge, so as to avoid conflating the two. It will be adviseable to focus the study on search queries within a particular domain of knowledge, in order to reduce the noise introduced by a range of different domain practices. It will also be necessary to ask subjects about their knowledge within distinct domains. Because a preponderance of gumbabies found to date seem to be related to searches for television shows, movies, and celebrities, this may be a fruitful domain of content knowledge to consider, perhaps bringing in knowledge of the entertainment industry, film and video production, or fan communities surrounding these people and properties. Such an investigation should be of interest to scholars in the field which has called itself “media literacy” since well before the advent of the Internet, as well as “new literacies” scholars and others interested in newer communications technologies. Media and celebrity as a domain of knowledge The domain of fan behavior has been studied extensively in departments of cultural studies, giving some insight which may be of use to this study. A survey of the literature yields a few models of the ways celebrities may be used by those who love them from a distance. Some of these are social, while others tend toward the psychological. Social models of fandom and celebrity Recent scholars have emphasized the importance of not treating fan activity as pathological. To treat fandom fairly may require that scholars step outside of their own socialization outside academia; popular discourses about fans are generally pejorative. In her contribution to The Adoring Audience, “Fandom as Pathology,” Joli Jenson analyzes these discussions themselves. She discusses two common characterizations of fans: fans as “obsessed loners,” and fans as members of an irrational mob. Jenson finds the roots of both of these stereotypes in anxieties about modernity, and dismisses these origins as more emotional than they are based in actuality: Each fan type mobilizes related assumptions about modern individuals: the obsessed loner invokes the image of the alienated, atomized ‘mass man;’ the frenzied crowd member invokes the image of the vulnerable, irrational victim of mass persuasion.... the present is seen as materially advanced but spiritually threatened. Modernity has brought technological progress but social, cultural, and moral decay. The modernity critique is both nostalgic and romantic, because it locates lost virtues in the past, and believes in the possibility of their return. (p 14) The “obsessed loner,” she writes, supposedly builds their fictional relationship to a celebrity to replace the lost, stabilizing forces of community and family. Meanwhile, screaming female concertgoers and male sports fans have supposedly built “irrational loyalties” to the crowds they take part in, which exert an irresistible emotional pull on them. These images crop up in the media in response to seemingly incomprehensible events like John Hinckley’s attempted assassination of Ronald Reagan to impress Jodie Foster or violent soccer riots. (p 11) Jenson suggests the themes of loners and screaming fans serve more of a purpose for the self-identification of those writing and reading these stories than they do as explanations of how media affect us. Comparing the obsessive behaviors of fans with the equally obsessive activities of academics (!), Jenson points out that the only real distinction between these behaviors is social sanction. She notes that the quality of the fixation on a famous person – is it openly emotional? “excessive”? -- and the status of that person in society, are used by the educated upper class to determine whether a fan is an acceptable “aficionado” or a questionable “fanatic.” “By conceiving of fans as members of a lunatic fringe which cracks under the pressure of modernity,” she writes, ... we tell ourselves a reassuring story – yes, modernity is dangerous, and some people become victims of it by succumbing to media influence or mob psychology, but we do not.... Unlike obsessed and frenzied fans, we are in touch with reality. We have not crossed that line between what is real and what is imaginary. (p 24) Anthropologists no longer need to explain why they are not treating their respondants as irrational participants in inferior cultures – they have a longstanding theoretical tradition to draw on -- but because of the history of popular and academic discourses about the media, communications scholars like Jenson still do. Jenson closes her article with a case for studying “the experiences of others in their own terms” (emphasis in the original, p 26). A study of gumbabies will need to be careful of pejorative characterizations of celebrities’ fans and of apparently irrational Internet users. More arguments for taking fans on their own terms are to be found in the work of Henry Jenkins, and in John Fiske’s work on the “cultural economy of fandom” in The Adoring Audience. (Fiske, 1992) Drawing off of Bourdieu’s ideas about cultural economy, Fiske talks about the forms of cultural capital, of “cultural tastes and competences,” which are enacted in the use of pop-culture texts by self-identified fans. (Fiske, 1992) This attention to skills and means of representation can be seen as parallel to literacies, particularly Jim Gee’s work on academic ways of being. (Gee, 2003 2004) Fiske, following Bourdieu, also outright connects cultural competence to education, linking the teaching of “official culture” (high art, music, dance, theater, etc) in schools – and the exclusion of popular culture therefrom – to the construction of sought-after social status. (Fiske, 1992) Fiske is ultimately more concerned with pop culture than high culture in this chapter. He investigates the seperate cultural production sphere of pop culture, describing its particular practices. He outlines a number of uses of pop culture texts in fandoms. “Selection of texts or stars,” he writes, ... offer fans opportunities to make meanings of their social identities and social experiences that are self-interested and functional. Those may at times be translated into empowered social behavior... but at other times may remain at the level of a compensatory fantasy that actually precludes any social action. (Fiske, p 35) The latter distinction sounds as if it may prove useful in looking at gumbaby behavior. So many gumbaby requests seek assistance of one sort or another from celebrities; it will be important to determine whether commenters believe they are fantasizing or actually acting by reaching out to celebrities. Other possibilities suggested by Fiske are that representing as a fan of a particular celebrity may actively issue challenges to social norms (he mentions young girls dressing up like Madonna in the 1980s), or that claims to “ownership” of characters, shows, or even actors may signify defiance of one-way media channels owned by unseen powers at a distance. (Fiske, 1992) Meanwhile, Jenkins (1992) writes about fans of science fiction writing. He noted that these fans pick and choose representations of their favorite characters to “[articulate] concerns which often go unvoiced within the dominant media.” Jenkins documents the art, videos, stories, and other media texts generated by fans themselves in response to their favorite commercial media. Sometimes these challenge common portrayals of masculinity; other times, they define the fan community as distinct from other media users; at still other times, they directly confront commercial media producers who threaten to take a favorite show off the air. In general, characters and actors serve as a set of reference points available to all viewers Prandstraller (2003) builds off of Jenkins’s claims, looking at an online community of fans dedicated to discussing U2 bassist Adam Clayton. Dyer, quoted in Prandstraller, writes that celebrities “’enact ways of making sense of the experience of being a person’ in a society that separates life into public and private spheres.’” (Prandstraller p 5) Prandstraller continues: The star image promises what mass society denies: intimate access to the authentic self. Through stars, mass media offer their fans consumable images, ideological values, and personal relationships. Thus, we can say that fantasies about stars are cultural products of mass media commoditization, which are used by fans to close the gap between what they need or want and what they can have. (p 5) Prandstraller describes ways in which fans post messages, photos, and so forth as a means of making themselves part of a community, using messages, photos, and other artifacts to “reify” (in the sense of Wenger, 1998) the identity of the celebrity in question and fans’ relationships to that celebrity. This reification even happens as fans select online nicknames, which often establish their feelings about the celebrity, or repurpose certain words and phrases, establishing an “in group” of those who understand what it means to be a fan of this celebrity. At the same time, she finds that fans want to identify as “not crazy” in their feelings about Clayton, identifying as “otherwise fairly normal” aside from their crushes on him. (p 23) The community Prandstraller studies, in particular, leaves a good deal of room for fan development of the identity of the celebrity in question, as Prandstraller notes Clayton is a private figure, leaving not many “clues” as to his actual personality in the public eye, and lots of room for fans to fill in. While they negotiate his identity, they also negotiate the appropriate complimentary female identity and models of romantic relationships. Like Hinermann, Prandstraller also notes that fans’ love of Clayton provides a sense of escape from their everyday lives, while not actually confronting the things they would like to escape. Prandstraller suggests that one appeal of being part of a community like this is availability of information; alone, individual fans would miss out on some information about a celebrity, but together they are more able to amass a large body of texts to share with each other. Prandstraller suggests that Internet fan sites in particular make it much easier to support smaller, more specific fandoms and a feeling of greater closeness among distant fans around the world. (Prandstraller CITE) On an emotive level, Prandstraller’s understanding of fans’ desire to relate to celebrities resonates with gumbaby comments. The use of celebrities to fill an absence in fans’ lives appears to hold true across many investigations of fan behavior. However, while some of the practices Prandstraller notes appear in gumbabies – particularly the identification with celebrities through handles, construction of celebrity identity, normative descriptions of how one ought to approach the celebrity or be a good member of one’s gender – there are many ways in which gumbabies involve a lot of “maladaptive” fan behavior, which does not appear to actually contribute to creating a fan community. Gumbaby commenters frequently do not respond to each other or refer to each others’ posts; there is not much sharing of information. They more often address the celebrity directly. The blog comment thread is not ideal as a space to share media; beyond the fact that they’ve chosen the “wrong” threads to begin with, blog comments often do not support sharing the graphics, links, or movie clips which fans use to cement their group identity. In fact, there is little reference to a fan group in these forums at all; fandom is more often expressed as a solitary individual affair. One wonders whether these commenters are really seeking out a community of fellow fans, or whether they are modelling their relationship to the celebrity in a different way. Reading Fiske, Jenkins, and Prandstraller, it begins to feel appropriate to make a distinction between “high fandom” and “low fandom,” as between highbrow and pop culture. Both researchers are specific that self-identified fans make distinctions between themselves and “everyday viewers.” Their cultural production – of videos, fiction, poetry, images, and other texts – sets them apart from other viewers. Neither Fiske, Jenkins, nor any other writer of this perspective on fandom seems to have a sense of exactly how large communities of “fans” like these are in proportion to “regular viewers.” My own research on video game players suggests the community of “high fans” may be quite small. Out of some 118 students at two schools of very different socioeconomic status, I found only about half a dozen were engaged in viewing or creating movies, art, screenshots, walkthroughs, music, or other texts related to but outside of the games themselves, the way Fiske and Jenkins describe fans as doing. (Andrews, 2007) Another troubling aspect of research on self-identified fans is that even though it refers to “everyday viewers,” references to literature on these viewers is not presented. Who are the “normal” viewers of shows like Star Trek or fans of Madonna or Elvis? What about casual viewers or listeners who do not actively seek out the company of others in viewing these shows, or viewers of shows which are less marginal and get higher Nielsen ratings? How do these viewers interact with their texts? It is not clear what an investigation of “low fans” might reveal – or, in fact, whether there is any real difference between fans and their imagined others at all. Fiske, however, hints that there may be a continuum, or at least multiple fandoms, exercising their pop-cultural capital in different ways. He suggests that the more intensive participation of “serious” fans could be “understood as elaborated and public versions of the interior, semiotic productions of more normal viewers.” (p 46) Fiske identifies the distinctions “fandoms” make between themselves and other groups of viewers as part of the cultural capital they wield. He also notes gender distinctions between fandoms, with older and male groups more likely to employ the “official or aesthetic” criteria of high culture to their favorite pop cultural texts than younger female groups, who will more often draw on criteria developed within pop culture which may fly in the face of highculture norms. (Fiske, 1992) Considering that in the initial review of gumbabies, the vast majority of commenters seem to be female, it will be very important to follow commenters as they speak and enact distinctions between “everyday” and “serious” fans, high and low culture, and genders. Psychological models of fandom and celebrity Writing elsewhere in The Adoring Audience, Hinerman (1992) recounts stories of fan activity which are very reminiscent of gumbaby comments, and may thus shed some light on the phenomenon. Hinerman draws themes from Elvis fans’ narratives about “the King’s” participation in their lives – since his death in 1977. Drawing on Lacanian psychology, Hinerman suggests that “fantasy” experiences, from the paranormal experiences described by some Elvis fans on down to anyone’s idle daydream about their favorite celebrity, satisfy desires which are otherwise unattainable or prohibited. He links this to Freud’s “pleasure principle,” which states that humans seek pleasure to minimize tension. Hinerman is specific that tension-reducing fantasy behavior is universally human; like Jenson, he does not cast judgments on Elvis fans for the expression of their specific fantasies, even though others might characterize these fans’ claims to have had paranormal experiences as “irrational.” (Hinerman 1992) Hinerman describes Elvis fans as drawing off what what they know about Elvis from various media sources to fantasize about him as a “conversational partner.” Frequently, they have conversations with Elvis during a time of trauma; he speaks words which the fan interprets to make sense of emotional pain. Hinerman ultimately notes that many of these traumas are caused by “institutional and social systems of power and control,” specifically gender inequality, in the case of Elvis fans. Hinerman notes that while fantasies fail to solve these systematic problems, they do give individuals relief from the distress they cause. (Hinerman 1992) Judging by the tremendous number of gumbaby comments which allude to family or personal hardship (many of which also implicate structural inequalities in the healthcare system, the job market, and domestic violence), and the intimate requests and statements many fans make on gumbabies (“i was wondering if u could come to my birthday;” “if i ever do meet u i promise i won't fall all over u ill be a regular like a sister!”) it seems plausible that celebrity fantasies play a role of sense-making and healing trauma in their lives. It also seems likely that the comments they leave might be a continuation of fantasies about conversation with celebrities. As such, it would be problematic to treat these comments as simply irrational. The practice of commenting on a gumbaby may serve an emotional purpose in commenters’ lives. Fan activity as healing; as a mark of belonging; as a sign of rebellion; as a form of capital; as enjoyment: clearly, there are many possibilities of what we might find as we begin to speak with gumbaby commenters. It will be important not to go in assuming that one particular model of the fan-celebrity relationship applies. Reaching out to celebrities celebrities The literature on fandom discussed above mostly concerns fan activity in society at large, among each other, or fans as they explain themselves. But what can the literature on fan behavior tell us about fans when they actually try to reach out and speak to celebrities? Beginning some time ago, with high-level theories, one finds C. Wright Mills’s explanation of the power of celebrity. (Mills, 1956) Mills frames the power of elites in terms of public rhetoric adopted by politicians: they often describe their enemies or foreign powers as “omnipotent.” “According to such notions of the omnipotent elite as historical cause,” Mills writes, “the elite is never an entirely visible agency. It is, in fact, a secular substitute for the will of God, being realized in a sort of providential design...” (p 16) This description of elite power resonates somehow with the comments left on gumbabies. Oftentimes, a celebrity or television show is called upon to intervene in the life of the commenter, to make a spouse happy or a child behave, or to help the commenter find a job or pay for medication. Viewed on television, Maury Povich may intervene to send a recalcitrant teenager to boot camp; Bill Gates may pay for schoolchildren’s medications in Africa; Overhaulin’ may pick a janitor’s car to be revamped on camera. What appears to remain obscure to the commenters is the process through which this happens: the selections made not by the celebrity themselves, but by program managers and other staff of the show or organization. One wonders whether the lack of visibility contributes to the sense that appealing to these elites will bring about a sort of divine intervention. One other contribution from Mills is the idea that more traditional elites – the wealthy, businesspeople, politicians and military figures – now garner some of their status from their association with media celebrities. He also posits media celebrity as a means for non-elites to attain equal status with and connections to the powerful; specifically, he mentions the role of modelling schools in grooming non-elite women for matches with elite men. (p 81) This suggests two more reasons for non-elites to seek connections with celebrities. One, celebrity in and of itself is a means to attain parity with the powerful. Two, elites such as politicians and businesspeople reciprocate and reinforce this status by seeking out the company of media celebrities, modelling and affirming the value of media celebrities to those reading tabloids and gossip columns. Mills stands apart from current scholarship on media and celebrity: he looks not at the influence of celebrity over everyday people, but the way star power works over (and with) that of the Power Elite. The work of Mills and that of contemporary popular culture scholars are somewhat at odds; the suggestion that “mass man” engages in any kind of magical thinking about celebrities – or that in fact that audiences consist of uncritical “masses” to begin with -- is more or less anathema in contemporary studies of popular culture. However, Mills’s focus on celebrities’ interaction with power structures is worth keeping in mind. Some explanation must be made as to why celebrities attract the attention and energy of fans to begin with, and Mills begins to lay a foundation for such an explanation. There are also some more concrete examples in the literature of fans trying to communicate with celebrities. An impressionistic piece towards the end of The Adoring Audience sheds a bright shaft of light on the gumbaby phenomenon, as it presents requests digested from fans’ letters to a British newspaper prior to 1984: A 16-year-old boy requests Paul McCartney’s address. He knows McCartney has a house in Kent and ‘once caught a train towards that direction but ended up lost.’ A housewife wants Michael Jackson’s address immediately so she can ‘fly to the States, find him and I would be a friend to him.’ A born-again Christian wants the same: ‘I believe very strongly that I can help him.’ A boy complains he can’t get near Adam Ant: ‘I have forked out so much money for concerts all over Great Britain to get his name in my [autograph] book...’ A girl wants the Sun to arrange for her to spend the day with Duran Duran. Another girl, who has been a David Essex fan for eight years, begs for ‘just two minutes’ with him to say: ‘Hello, how are you?’ (Fred and Judy Vermorel, 1992) The similarity in tone and content between these requests and the gumbabies is so strong as to leave no doubt that the gumbaby phenomenon – at least insofar as the content is concerned -- is not at all new. The requests being made by celebrity-seekers on the wholly inappropriate channel of blogs are not specific to the Internet. People have been making requests like these for years. For how many years? It could be argued there is evidence that people have been confusedly seeking contact with those they see in images or read about for at least a century and a half. Wilkie Collins, whose Victorian “sensation” novel The Woman in White was published in 1860, received letters from men who wanted to know the identity of the woman on whom the fictional heroine was based. Many of them sent marriage proposals. (Collins, 1999) The tomb of Juliet Capulet in Verona, Italy, has received letters to the star-crossed heroine since the late 1930s, when the first film version of Romeo and Juliet was released; these letters are often addressed simply to “Juliet, Verona, Italy.”4 (Most sources pin responsibility on the movie, but one wonders if letters arrived before that.) Consider, also, children’s letters to Santa Claus. While Santa’s existence may or may not be known to writers, the Juliet letters present a clearer case, considering anyone who has finished Shakespeare’s play ought to be aware that the heroine is really in no shape to be answering her mail, even if the stabbing only happened yesterday. Juliet and Santa Letters and many of the letters written to the Sun share something with gumbabies which letters received by Wilkie Collins do not: they will end up being read, if at all, by the “wrong” people. To triangulate the gumbaby phenomenon, it might be 4 http://www.shaksper.net/archives/2004/0441.html, which comes from “Shakespeare's Juliet is an agony uncle in Verona” Deutsche Presse-Agentur (DPA), at Khaleej Times Online, UAE, February 13, 2004; http://www.julietclub.com/ ; Google’s cache of http://www.europeanjournal.de/index.php?p=news&newsid=25&area=1 . Tip of the hat to Elvis Costello and the Brodsky Quartet for bringing these letters to broader attention. worth looking into the literacy activity associated with sending messages which are not received by their intended receivers, or contain “inappropriate” content. Doubtless, movie studios and agents as well as newspapers and postal services could provide a great deal of information on the subject. Of course, Juliet and Santa letters involve sending messages out into the wilds of the postal system with addresses which are fictional (the house in Verona known as Juliet’s is only a best guess at where the girl assumed to be Shakespeare’s model lived) as well as vague. This begins to suggest that the literacy practice these writers engage in is in fact a knowing engagement in perpetuating fiction. It remains to be seen whether gumbaby commenters address themselves to stars as writers address themselves to the ostensiblyno-longer-able-to-read Juliet, or whether they actually believe their message will reach their target. Library and lab as a domain of knowledge: Problems with classical models of information seeking As Gee’s work suggests, each domain of knowledge has its own means of organizing information to make it legible. The literature on “information seeking” which comes from departments of library and information science (LIS) has recently tried to expand its domain to everyday information seeking; however, its strength lies where it began, in libraries and academic search. Here I will briefly describe the field and the theories it has suggested to make sense of information-seeking behavior, but this will mostly serve to explain why I do not plan to rely heavily on this body of literature. I will then briefly discuss some of the more promising recent literature arising from this field. Were he subject to journalistic Discourse, Donald Case -- the author of a major review of the literature in the field of what is variously called “information behavior” or “information retrieval,” often lumped in with “library science” – would be accused of the informational faux pas of “burying the lede.” His most insightful summation of his field is buried deep in his section on research about information-seeking performed by scientists and engineers: this is where the research in this vein really got started. The “Big Science” (Price, 1963) sparked largely by World War II and afterward by the Cold War resulted in an explosion of research material. There were simply too many findings being published for individual scientists and engineers to monitor effectively. The outcome was frustration and sometimes outright duplication of research efforts, because researchers did not always know that others were gathering or even publishing findings of interest to their work. As a result, money and attention became available to address problems in the dissemination of scientific information and to study communication among scientists and engineers. A 1984 comment by Tom Wilson accurately characterizes the nature of the literature at that time: “the study of information-seeking behavior can be said to be the study of scientists’ information-seeking behavior” (p 199). From the 1940s through the 1970s, investigations of scientists (and to some extent, engineers) dominated all others. (Case 2007, p 252) These paragraphs are immensely telling about the historical assumptions and structure of studies of “information.” The field long took for granted that a search for information is 1) textual, 2) performed for work, 3) accomplished in a formal setting, 4) involves one seeker with a high level of sophistication in the domain in question and probably also 5) an intermediary such as a librarian; it assumed 6) formal organization of such information (card catalogs, citations, etc.), 7) articulable and practically-applicable reasons for performing a search, and finally, by extension, that 8) empirical facts could be found in a pot of gold at the end of the search rainbow. These traditional assumptions left a mark on the field for a long time after the focus had shifted away from scientific knowledge; the influence of these beginnings is still visible in many of the models of information behavior which Case (2007) puts forth, and is even more so in older summaries of the field (see e.g. Marchionini, 1995) As recently as 2001, summaries of the literature still found it necessary to remark on the limited number of studies which dealt specifically with Internet searching. (Jansen and Pooch, 2001) Older literature on information seeking is thus not particularly applicable to current searching on the Internet, which involves searchers and situations who in the majority of cases are different from the searches enumerated above. The theoretical offerings of this field are also only recently coming around to a point at which they can be useful lenses for viewing Internet search. I explore these theoretical approaches below. Within (what I will somewhat idiosyncratically call) information studies, there have been a number of theories developed about the general reasons why people search for information. Case (2007) describes these as drawing predominantly on sociological, psychological, and mass communications theories. It may be because it draws so much on psychology, Case speculates, that information literature has often posited information “needs” which arise from desire, conflicting ideas, or other sources of emotional tension – a focus which he suggests derives obliquely from the work of Sigmund Freud (1922, cited in Case 2007 p 149). A couple of lines of thought derive from this way of looking at information: the Uses and Gratifications paradigm and Play Theory/Entertainment Theory. (Case, 2007) Uses and Gratifications is a tradition most often applied to media in general, not just library-based or general information seeking. It posits that audiences actively choose the media they watch, read, or otherwise consume in order to fulfill internal needs, resulting in a feeling of gratification. (Case, 2007) Play/Entertainment Theory – which Case suggests can be viewed as a subset of Uses and Gratifications -- posits that much information-seeking that people do arises from a desire to be entertained, not just from a desire to be informed. Studies in this vein have looked at reading newspapers as a form of “intellectual puzzle” and catharsis, researching hobbies, and the emotional states of people watching television. This theory’s strength, Case suggests, is being able to do away with the “artificial distinction between information and entertainment.” (Case, 2007) Case notes that these traditions have unfortunately not developed many solid operationalizations of their terms (“needs” and “gratifications,” among others), and do not make many connections to other traditions of research. He does note that the field has recently been grounded somewhat by a sub-field identified as Media Use as Social Action, which pays some attention to how media viewers shape their interpretations depending on their personal characteristics and context. (Case, 2007) Because of the weaknesses Case cites in the Uses and Gratifications tradition, this study will not generally deal with models of information seeking which are based on “information needs” or the internal state of the speaker, except as they come up in the specific case of fans’ relationships to a celebrity. As Case notes elsewhere, “most information needs could be said to be accounted for by more general needs, and in any event they cannot be observed.” (Wilson 1981, Poole 1985, cited in Case 2007) It may be useful, however, to follow the lead of Play/Entertainment Theory and not draw distinctions between “playful” and “serious” information uses. Following traditions of grounded research, whether this distinction appears should ultimately depend on whether it is a distinction made by the population being studied. (DO I NEED TO DEFINE “GROUNDED”?) According to Case, it is only relatively recently that information studies have come around to social constructionist theories which treat the information which searchers encounter as co-constructed by all participants in the context of their current situation. A major proponent of this line of thought in information studies is Brenda Dervin, who has developed what she calls a “sense-making methodology.” (1992, 2005, cited in Case 2007) Tuominen, Savolainen, and Talja, who have been cited elsewhere in this paper, have also worked on projects with a similarly constructivist focus. (CITES – Case says 2005, T&S 1997) (WAS THERE MORE HERE?) Because of the mismatch of older information studies work and its theories with the kind of searching done on the web, this study will draw more heavily on literacies research than it will on the older corpus of information studies. Certain parties in information studies have begun to turn in the direction of literacy research anyway, with Tuominen et al recently citing James Gee, Jean Lave, Etienne Wenger, and other stalwarts of literacies research. (Tuominen et al 2005) There are, however, recent efforts in information studies literature which have begun to overcome the traditional assumptions of the field and have come up with findings with direct impacts on this study; I will explore these below. More recent information studies work Pioneering research on large bodies of data about everyday Internet search behavior highlighted the problems with applying the assumptions of earlier information studies literature to this new set of behaviors. Jansen, Spink, and Saracevic (2000) found striking differences between the literature on experts’ formal searching of article databases and patterns they themselves found in 51,473 queries made to the Excite search engine. While older studies of information retrieval had found query length to be on average between 7 and 15 words long, Jansen et al’s Excite searchers averaged 2.21 words per query. Fifty eight percent of searchers did not go any further than their first page of results (which, on Excite, displayed 10 results), They also found that Excite searchers made much less use of Boolean operators, quotation marks, and other modifiers than formal searchers did. Less than 1% of all queries in their dataset used the Boolean operators OR or AND NOT. Only 8% of queries and 5% of users used the operator AND, though the authors acknowledged they did not and could not distinguish between standard grammatical and Boolean uses of the word. (Jansen et al, 2000) Jansen et al were able to look at data on the level of a “session,” meaning that Excite kept unique IDs for users for a length of a few minutes up to a few hours. Their study thus does not yield any information about searchers’ longer-term information seeking behavior, but it did turn up some results about sessions over the short term. On average, these users generated 2.84 queries per session, though 67% only entered one query in a given session. 22% of all queries were found to modify an earlier query. Those who did modify their queries added and subtracted terms from their query strings in roughly equal numbers (one in five added terms compared to one in six subtracting). About a third of queries used the same number of terms as the user’s preceding search. About 43% of all queries (from about 33% of users) were assumed to indicate that the first ten results users were presented with did not satisfy their search. A later study on a larger body of Excite data turned up similar findings about queries per session, though distributions revealed that about half of searchers only entered one query in a given session. In this study, a larger number of users added terms when modifying a search (41.6%) than subtracted terms (25.9%) In general, this study confirmed the findings of the authors’ previous study. (Spink et al 2001) Recent responses to information studies oversights: Focus on context Information scholars have recently called certain of their field’s historic assumptions and oversights into question. One failing of the literature has traditionally been its treatment of information-seeking out of context. This has more recently been addressed. Rieh (2004) notes that [s]ince 1996, information seeking in context (ISIC) has been the theme of a series of international conferences. Research on context highlights an approach to the study of information seeking that emphasizes real users with actual information needs prompted by situations arising in daily life (Kuhlthau & Vakkari, 1999) (Rieh, 2004) The dependence of search strategies on the subject area being searched has been indicated by a range of studies focusing specifically on searches for information on particular topics; Case notes, drawing on Bryce Allen (1996), that “Being a member of a group, such as abused spouses, cancer patients, senior citizens, or janitors, is seen as sufficient to influence individual information-seeking behaviors and patterns.” (CHECK CITATION ON THIS) A number of studies of information-seeking found that outside of the formal context of libraries, information-seeking was conducted very differently. In at least three different studies (Julien 1999; Chen and Hernan 1982 and Dervin et al 1984 as reported in Case 2002), participants cited relying on their own existing knowledge and experience more often than any other source (74% of respondants in Chen and Hernan; 89% in Dervin et al). In Chen and Hernan, the next most popular source of information was interpersonal: 57% of respondants said they went to a friend, neighbor, or relative to seek information; 43% would go to a co-worker, 41% to a professional. (Chen and Hernan, 1982, via Case 2002) Within traditional library searches, Gross has acknowledged the importance of context. In the qualitative component of her study, she found that requests for research made by a teacher tended to be accompanied by the provision of texts the teacher intended students to use and by specific language shaping the question. When the language provided by the teacher did not employ the same language as the texts students found, students tended to be confused. Synthesis of many sources proved tricky for them, though older students knew to compare sources to other sources or “known facts” when assessing relevance. (Gross 1999) Interestingly, Gross notes that “In general, the idea of a correct answer made sense to them mainly in terms of their mathematics lessons” (p 513). This is a rather reductive approach to knowledge when approaching more qualitative topics like history, or ongoing theoretical scientific explorations like the process of evolution. Gross does not extensively explore what “context” means in the classroom environment, but her finding that students seek empirical “correct answers” suggests that beyond the context of an individual class domain (English, history, biology, etc), the context of “classroom” itself may have the strongest formative influence on queries imposed by a teacher. Some of Rieh’s work has explored the specific situation of information seeking in the home. (2004) Rieh employed semistructured interviews in review of participants’ Internet search diaries to explore how users’ Internet information-seeking at home built off that environment and its affordances. Making use of Xie’s (2000) breakdown of user search goals into four categories including short-term and long-term concerns, Rieh found that her participants often felt satisfied about their searches even when they did not end up on sites which had the information they sought. She saw users conducting searches on the same topic over a long period, explaining that these searches supported a long-term interest of theirs. (Rieh, 2004) This finding undermines the reliability of studies where users’ landing on particular pages is taken as a measure of “success.” At the same time, it supports suggestions that Internet reading can be a very personal, dialectical process (Bertram Bruce, 2000), and suggests the utility of longitudinal studies of users’ search practices. Rieh found interesting indications that searchers’ limits of domain knowledge have an impact on how they search: general search engines such as Google, Altavista, or Excite were not the first place that subjects went to in looking for information. Rather, for most subjects a search engine was the last site to turn (sic) when they could not think of any topicspecific sites.... Subject 01 said “I only start at Google if I don’t know anything about what I’m doing.” (Rieh 2004, p 749) Rieh also found that even when using general search engines, participants calculated their queries to seek out topic-based sites or sources of information first, then looked around these sites for more specific answers. (Rieh 2004) It should be noted that her population was not based on a random sample, and their reported use of different search engines puts them in a specific bracket of the Pew study findings: they did report using two or three different search engines, which in the Pew study was correlated with greater savviness about paid placement and less trust in the objective truth of search results. (Fallows 2005) Possible sample bias aside, Rieh’s work and other studies on context support the case that critical digital literacy cannot be separated from specific domain knowledge. Rieh in particular notes that her relatively sophisticated users employ specific search practices when it comes to subjects like weather, maps, and images. (Rieh 2004) It also echoes the Pew finding that most respondants seeking religious information online went first to sites they already knew rather than to search portals. (Fallows 2005) Mental models The small body of literature on “mental models” which has arisen in library and information science can be understood within Gee’s division of different kinds of language and literacy use into specialties, Discourses, or domains. Specifically, the literature on users’ mental models of search engines can shed some light on their domain knowledge of the Internet, which as I suggested earlier ought to be considered separately from their knowledge of other content. (I NEED TO ELABORATE ON/CLARIFY THAT ABOVE) Gee posits that the process of comprehending what we are reading involves engaging our existing mental simulations or models of how objects described in a text work in the world. Gee also spends a great deal of time talking about how players “read” the possibilities for action presented to them in video games. (Gee, 2003, 2004) This perspective on reading strongly suggests that we understand users’ mental models of search engines, blogs, and other Internet infrastructure in our analysis of how they are reading links, banner ads, search results, and other elements of the web. In the information studies literature on mental models, Efthimiadis and Hendry contributed a number of studies. Among some 232 university students in their study --179 of whom were majoring in information studies -- few drew sketches which indicated an understanding of how search engines index pages or match results to queries. Spiders, indexes, and even ranking were among the elements of a search engine which appeared in less than 30% of participants’ drawings. (Efthimiadis and Hendry, 2005) Zhang (2007) similarly instructed users to draw out their mental models of “your perceptions about the Web” on paper. Categorizing these, she found four types of model: technical (composed largely of computers and other technical elements, with some people); functional (based on behaviors one performs on the web); process (with the search engine as the center of the web); connection (emhpasizing connections between pages, people, information, and technical elements). Zhang did not find that having any one of these models seemed to give users an advantage in conducting searches. She did find some gender differences, however, with more men drawing technical models of the Internet and more women drawing process models. Blandford et al (2007) discussed users’ mental models of the Internet in their case study of users’ perceptions of libraries. They concluded that in fact, “most participants did not clearly distinguish between different kinds of digital resource, viewing the electronic library catalogue, abstracting services, digital libraries and Internet search engines as variants on a theme." Even among their participants, who were masters’ students in their own department of library and information science, they found little understanding of the process which a search engine uses to return results in response to a search query. Participants mostly understood differences between search engines at the superficial level of the interface; they could not articulate how differences might affect what they should enter as search queries. Misunderstanding of search engines led users to trial-and-error search methods, which Blandford et al suggested might be remedied by “rich feedback” integrated into search engines by their developers. (Blandford et al 2007) In sum, Blandford et al concluded that users misunderstand how search works, and this inhibits their ability to fruitfully use search engines. However, some of the participant quotations they include seem to also indicate that users’ misunderstandings of the domains they were seeking within may have been as much the culprit as their misunderstanding of search engines. Their finding that users’ lack of awareness of the range of documents which might be available to an engine they were using – which they found tended to discourage users -- could also be seen as an interaction between the engine and the domain of knowledge it covers. Statement of question Building on the research in information studies and in new literacy studies, it behooves us to ask: What are the conceptions of the Internet and other specific knowledge domains which bring gumbaby commenters to search engines and comment boxes, and lead them to comment? In context, how are they reading, and to what end are they writing? In order to investigate this, we must also know about the Internet environment they encounter: what aspects of the construction of Internet elements – like blogs, search engine algorithms, advertising, and portals – are also implicated in the production of gumbaby comment threads? (Fallows, 2005; Graham, 2003; Henry, 2007; Leu, 2007) Fallows, D. (2005). Search Engine Users. Washington, DC: Pew Internet and American Life Project. Graham, L., Metaxas, P. (2003). “OF COURSE IT’S TRUE; I SAW IT ON THE INTERNET!” Critical Thinking in the Internet Era. Communications of the ACM, 46(5), 71-75. Henry, L. (2007). Exploring new literacies pedagogy and online reading comprehension among middle school students and teachers: Issues of social equity or socal exclusion? Unpublished Dissertation, University of Connecticut. Leu, D., Reinking, D., Carter, A., Castek, J., Coiro, J., Henry, L., Malloy, J., Robbins, K., Rogers, A., Zawilinski, L. (2007). Defining Online Reading Comprehension: Using Think Aloud Verbal Protocols To Refine A Preliminary Model of Internet Reading Comprehension Processes. Paper presented at the Conference Name|. Retrieved Access Date|. from URL|.