Another piece I wrote a couple of years ago, unsuccessfully pitched (a shorter version) to a magazine.
I’ve been watching my parents’ eyebrows for 6 months now*. A weekly ritual of Skype calls, London to Sydney, and we haven’t quite perfected the art of a full portrait. Partial information exchanged over 10,000 miles. Other information, though, flows more easily. Our messages, planning for times together and catching up on separate lives, are replete with photos, google searches for the latest destinations, URLs exchanged. Nuggets of information more easily shared, portals into what we’re looking for in our respective places.
They’ll be visiting Sydney again soon. Until then we’ll engage in these exchanges. Their beach trip photo for mine. Between these contrasting photo exchanges – their ocean “blue” for my ocean blue – we trade in restaurant possibilities (some no doubt to expire before the visit), events, important visa and insurance details. The photos we trade in remind me of Google’s “visually similar images” function, which returns files that share key features – most notably colour and composition. You can drag a picture to your browser’s search, looking for images – not search terms – that Google thinks are related. Upload a profile picture to see where else it appears, a beach shot to see Google’s reckoning of similar vistas. The artist, Rebecca Lieberman, frames this in terms of the “poetics of search”. In a talk of hers in Amsterdam, she took the audience – me included – through a journey of this semantics of search, a “shared formal vocabulary” of disparate images that share key features, meanings lost and retained through this visual semantics. The colours, shapes, shadows, that Google classifies images by, identifying similarities across visuals. These journeys took us from your photo of a football pitch, to Google’s similar photos…photos of fields, a bowl of apples, block-green clipart…I’m not sure if my ocean doesn’t share more meaning with a handful of blueberries than it does with your grey ocean.
The information I share with my parents isn’t just itinerary planning or pictorial bragging. It’s about what we want to do, our interpretation of reviews, about sharing what we – in combination with the resources of the web – know about. When we engage with resources on the internet, in a very real sense we’re engaging with the voices of others, “well this reviewer says…”, “yeh but they sound mad…they got angry over the brand of tea??”, “these guys had a great time there…”. When we do that physically or virtually with our friends or families, we bring those conversations into the fray. How you read that article, those reviews, these pictures, is not the same as how I do, or how we do it together.
Because of this multi-vocality some researchers in information seeking – the key folks in how we use search engines – including myself, have recognised that when we engage in a search, it isn’t just about “finding the fact”, it’s about the social context, the ways language is used across the web and in our own lives. Drawing on the philosopher Mikhail Bakhtin, some have recognised this ‘many voicedness’ or polyphony on the web – that engaging with the web involves interaction with the thoughts of others, in our reading and comment on blogs, tweets, facebook posts, and exchanges of memes. That changes how we search, and what information we gravitate to, the language of search isn’t neutral, embedding the popularity of content (and of global languages) to its core.
I’ve been studying this phenomenon of the social and collaborative features of information seeking for some years. Sometimes these social features are implicit, algorithmic: your Netflix suggestions, supermarket deals, the articles in your newsfeed. In these cases, other people are mediating your access to information; it isn’t just an algorithm that drives personalisation, it’s the choices and preferences of “people like you”. Modern search engines, and any other systems that use recommenders, are bound up in this algorithmically mediated social interaction. It feeds off our own actions – we choose the movies, the articles to like, the brand to select – and joins us to others who make similar choices.
But other times, like with my parents, it’s more explicit: the teacher who leans over your shoulder to suggest a keyword; the friends who sit and plan a holiday together; the colleagues who split up a task to each find part of the needed information. In fact, researchers at Microsoft have surveyed professionals, and found that these kinds of collaborative search scenarios are pretty common across our home and work life. We don’t just search alone, we search together, assemblages of ourselves, others, and technology
What’s common to many of these tasks is that they break one of the conventions of modern search: they’re about finding lots of information, not one answer, and they tend to meander, and engage people in what Microsoft researcher Jamie Teevan calls ‘slow search’ (think ‘slow food’). Information seeking experts call these ‘exploratory’ searches, because they’re not just about precision – the answer that Siri and Google can give you immediately – they’re about investigation, and learning.
I’m just about old enough that I have needed to use library catalogue cards, although I haven’t the foggiest where that would have been, and only the faintest memory of how they even worked: a system of subject codes and bibliographic data that placed books on a similar theme close to each other. When search engines, or portals, first emerged they were like this, allowing you to browse to subject-based information, or query for particular terms across the metadata or content of a document. You had to know what you were looking for (initially at least), and it made sense to browse by subject. You might have used a thesaurus and book index to look for new vocabulary to write carefully constructed Boolean queries. Algorithms that look at how words are embedded in texts across the web can do this for me now. I don’t browse books physically and topically close on the shelf anymore, I browse books that people who’ve read my book have also read, whose reviews I can read over; cross-referenced through algorithms and annotated through social interactions.
The rise of Google and its pagerank algorithm marked this change. Suddenly searches weren’t just returning results based on the content of the documents. Our information was made more social by data. Data about the ways the documents were being referred to elsewhere, how they were interlinked and – even – how they contained similar meaning, but different words; data that is fundamentally social in nature.
The introduction of suggested search marked a further step change, relying on indexing relationships between key terms and being able to link my own queries to those of others; knowing that many are searching for the impact of Brexit, Google knows “Brexit consequences” is a search I might want to make when I start typing “Brexit”. These are further personalised; my searches for ‘Liverpool’ send me to the Sydney suburb; not so many scousers as my parent’s northern English results. The consequence of this has been much discussed, with Google mirroring back to us society’s prejudices, and seemingly able to predict societal shifts from elections to flu incidence.
I also watched the impact of these suggested searches in some of my research on how trios of children search for information together. Asked to research female role models they searched “How many women ha…”, the suggestions were all along the lines of “how many women has Charlie Sheen slept with”; sometimes, the things people around us search for aren’t so palatable, or so useful. These suggestions entered into their conversation, shaping their dialogue. While for another group this resulted in an entertaining (if, off task) discussion around the superpower they’d want, more insidious suggestions have led some to accuse Google of racism. These suggestions are based on what we search for, a mirror back onto society.
Search has entered our conversation. We bounce queries back and forth mid-discussion, we take its suggestions, Google’s spelling corrections accepted with grace. We appreciate the knowledge Google leads us to, and how we talk about that tells us something about how people treat information. When I conducted the research with children, we asked them to think about why the information they found was important. One group said repeatedly “because I didn’t know that before”, the novelty was important to them. Another was more interested in “key facts”, and authoritative websites, with another focussing on the quantity of knowledge, caring little about the quality of the source (“it had all the information”). It’s not just children either, when you’re searching for something with friends, sharing via facebook, Whatsapp, Skype, how often is the conversation focussed on the author, their venue and sources, compared to the content?
Perhaps, in fact, we should engage in more conversation with Google. Often, we tend to take the first result, the most linked-to (plus PageRank magic sauce) page, and rarely probe deeper. But that act, of delving into the second page of results, or of searching ‘laterally’ – about the source, the author, their expertise – rather than vertically within a page of content, could open up our conversation, and lead to deeper discussion on the who, how, and why, rather than the what.
Recent coverage of ‘fake news’ and ‘filter bubbles’ (including my own https://goo.gl/PkzpBS, https://goo.gl/GgVKZH) highlights exactly this need to talk about search, flagging that rather than take ‘news’ at face value we should explore, slow our searching down and search across sites, entering into conversation. We can do that both with the search engine itself, and in collaboration with others. Sometimes – like in my work with the children – that’s about looking for the justifications for claims, laying it out with your peers and trying to make the reasoning explicit. Other times, novices might work with experts to find or evaluate information; stories of patients turning up to Drs with internet printouts are not uncommon, and nor are stories of Drs searching for information during a consultation. In my own work it’s common for us to turn to Google Scholar for a quick check on the literature. But how we do that, our disciplinary and search expertise together, matters; in a recent meeting I showed people how you could search articles “citing” another piece, and limit by year, to only show recent papers, but which ones were relevant needed my student’s expertise to judge.
Like it or not, information seeking shapes, and is shaped by, our thinking. Concerns swirl around that Google is making us dumber, promises raised that it’s increasing our capacity, extending our mind. Pub quizzes are certainly easier. And, we might be less likely to share a calculator than let Google resolve a calculation now. But we still need to know what to search for, and how to work with what we find. As the Google researcher Dan Russell notes, most people would agree that knowing a factoid isn’t the same as having an education, so why do we treat Google as a threat to that education? We communicate with Google and it communicates with us, giving us information, and – through pagerank, suggested search, and google trends – giving us insight into how others are thinking about something. We need to learn how to speak its language, as it learns to speak ours. Social features pervade the way we seek and find information, from the various ways that algorithms mediate our access, to the social contexts and conversations in which we find ourselves reaching for Google and talking over its results. What we know matters, search engines are entwined with that, both algorithmically, in the ways we talk about everyday issues, and in those weekly Skype calls home…
* This characterisation may be slightly embellished