JFGI or “I will use Google before asking dumb questions”: An investigation into subjectivity and language usage in search queries within the context of traditional questions.

I recently gave a paper at Exeter University's College of Humanities Postgraduate Conference entitled "Shifting Boundaries" 

The paper itself is comprised of a part general summary of my PhD project so far, a summary of the main argument of my first chapter and some comments on how we treat search engines and the ways in which Google has become deeply integrated into our lives:

JFGI or “I will use Google before asking dumb questions”: An investigation into subjectivity and language usage in search queries within the context of traditional questions

Today I’m going to briefly outline the wider scope of my PhD and then focus on what I’ll call the transition from questions to queries.

So I study the nature of web search engines and the effects they have on traditional ideas concerning knowledge. Google is increasingly becoming, the font of all our questions, if not all our answers too. And Google works so well! Or at least we’ve grown very satisfied that the answers Google gives us are the kind of answers we were looking for in the first place. And that the way these answers are provided, a hierarchized top down, linear list representing the previous preferences of other users is just the way these answers should be presented. But to criticize the effect Google has on our lives seems like the complaints of a spoilt child.

Google knows what I mean so much of the time, right? It gets me webpages it knows I’ve been before, finds me new ones which it knows I’ll like, recommends me restaurants or the cheapest place to buy new cds, shows me the quickest way to walk across town when I’m late for a meeting, or gets me a page scan of a book for when I’m too lazy to walk to the library to sort out my essay citations. 

Google sorts me out, so much of the time.

Increasingly, the phrase “just Google it” means, use your computer to find the answer. And in an age of computer ubiquity, when most of us carry around a small computer in our pockets, we increasingly assume “just Google it” as our mantra for finding answers. The “just” in “just Google it” serving to remind us of its simplicity, speed, ease and above all transparency. You don’t “just” do things which have complex or ambiguous outcomes.

With this in mind lets take a look at Google’s mission statement.
Google’s mission statement is, and always has been: "To organize the world’s information and make it universally accessible and useful.” Not the web, but all the world’s information. I find it very difficult to imagine what all the world’s information would look like, let alone how we could make all of it accessible and useful.

And don’t be deceived that this is simply an ambitious way of describing web search. Google stepped away from simply directing people to things on the Web long ago. Google Earth, Maps and Streetview are attempts to systematically capture “all the world’s” visual and geographical information to make it accessible and useful. Google Glass enables our daily lives to become machine-readable and allows the data of our everyday experiences to be placed in a functional context, rendering all stimuli meaningful and useful. Finally, Google Books, perhaps the more important of these developments, which last year announced it had scanned and made searchable 30 million unique books and aims to have scanned all published books ever (Google estimates this at around 130 million) by 2020. These scans both help Google’s algorithms better understand our natural language inputs and also means that Google’s results aren’t just webpages but all of history’s published thought.

Jorge Luis Borges’s short story “The Library of Babel” comes to mind. In which a universe consisting of a “total library” contains every possible combination of that universe’s 25 letters in every possible order. All possible knowledge has been gathered but the story tells of suicidal librarians who have gone insane trying to separate out the valuable volumes from the mass of incoherent possibility.

Here the idea of relevance becomes key.

In 2004, Larry Page, one of Google’s co-founders, said that "The ultimate search engine would basically understand everything in the world, and it would always give you the right thing.”

Which begs the question, even if Google manages to create a machine which “understands everything in the world” what does giving me the right thing mean?

Does that mean structuring all of human knowledge around the opinions and belief of the masses or academic elites?
Should Google tailor its results to people in my country, or specific communities or personalise to me as an individual?
Could knowledge ever hope to be relevant or useful if it was structured in the same way for everyone?
And if we think knowledge structures should be localised, what does this say about our increasingly globalised society?

These are just some of the questions which stem from beginning to think about Google, but in order to be academically rigorous we need to focus our attention on specifics. Therefore for the rest of this paper I want to talk about a transition from traditional questions to Google queries.

To do this I want to return to something I mentioned earlier. The phrase “just Google it”. As I’ve already said Google has become a byword for finding things out, the “just” reinforces that the task is easy, will be unambiguous and that the answer found will carry with it a relatively unquestionable validity.  We make jokes about trusting Wikipedia too readily but we’re never as dismissive about Google results, even if, as is often the case, the top Google result is a Wikipedia page which may or may not be plagued by factual inaccuracies or ideological biases of its authors.

One of the reasons for this discrepancy is the way in which Google self-describe as driven by algorithms as if there is little to no human intervention. As if technology magicians simply enable Google to come to life letting it feed on data and sustaining itself while producing machine made objectivity. We forget that algorithms are fundamentally ideological. An algorithm is simply a step-by-set list of commands which can be deployed in different situations. In this way an algorithm is much like a recipe. When baking a cake, if we decide we need 50g of flour for every egg we use we are building assumptions about what the outcome needs to be like. A simple recipe which allocates 50g of flour for every egg draws on cultural conceptions of what cakes should taste like. If we wanted we could run the recipe through a cooking computer which completely automates the cooking process but we still can’t get away from the fact that we designed what the outcome should be like in advance. If the machine goes wrong and we make tweaks or corrections, we are doing so in order to align the outcome with what we think a cake should be like. Google and its users often shift blame onto ‘algorithms’ as if they have no human guiding them and we should be sceptical of this.

The phrase “just Google it” has entered into popular discourse through common usage, news coverage and become a simple answer for complex issues. The “just” in “just Google it” reinforces certain positions mentioned above such as speed, easy, simplicity and most importantly that all questions can be posed in somewhat obvious terms. We can see from this Google trends graph (which shows relative popularity of search terms) that the term has grown in popularity due to its use in articles and news reports. 

This attitude of “just” doing something, covers all types of Googling from simple to complex. It would not be uncommon to answer “just Google it” if asked “what are the opening times of John Lewis” in the same manner you might to the question “is the Central African Republic more or less politically stable than a year ago”. It flattens out variation and simplifies the process of understanding.

This “just” attitude is part of the transition from questions to queries (this paper’s final section).

Another attitude builds out of this disposition and helps reinforce queries over questions, one we can describe as JFGI, or, Just Fucking Google It.

I don’t want to labour a point nor be crass for the sake of it, but I think the phrase, Just Fucking Google It, and its popularity both as a meme in the Internet-slang sense as well as in the more generalised Dawkins terminology is important for understanding a certain widespread attitude towards Google search that isn’t encapsulated in “Just Google it”.

The phrase expresses a kind of exasperation, but more than that, it expresses an underlying anger and outrage that can be unearthed at a moments notice. How dare you attempt to waste the time of a human when a computer can “just” do the job?

Although often used on messageboards, comments sections and blogs the meme “Just Fucking Google It” had a heightened popularity around 2005, when Google was around 8 years old. This isn’t a very long time to reach a point whereby not only did many of us feel we could reject some questions out of hand but that anger would be a legitimate response. As if asking us to do the work of a machine was belittling or that we were embarrassed that we could know someone so stupid that they their first port of call would be a human.

2005 signals a shift, whereby many traditional questions were becoming queries. We knew or strongly presumed that somewhere these questions had been asked and answered a thousand times over and should not be considered as questions anymore. The outrage of “just Fucking Google It” signalling their death as traditional questions. Now that these types of questions were now fundamentally queries it was insulting, belittling or downright dumb to ask them to people.

In order to treat what this strange claim might mean we need to take a step back and think about some ideas we take for granted, specifically we need to assess how we define questions.

Questions are strange in that we really don’t have a clear definition of what constitutes a question, yet we can all casually identify them in common usage. Multiple definitions exist but each highlight different features at the expense of others. This is particularly evident within academia, it is not a matter of comparing the definitions of linguistics, philosophy and sociology but navigating the entrenchments and disagreements within each discipline that define it in multiple ways. For time reasons I won’t outline all these different definitions but just mention how for example, in the field of linguistics definitions tend to group around one of two camps, the semantic and sociolinguistic, which are essentially mutually exclusive, either defining a sentence as a question because of its grammatically composed like a one (starts with a WH, ends with a question mark) or is a sentence which elicits a response that grammatically looks like an answer.

However, these kind of definitions lead to approaches which ignore context and often just fall into a spot the difference or parlour game of sorting questions from non-questions.

Instead of having such a rigid distinction I want to outline the description Elias Canetti gives questions, not because I think it is the most accurate but because it discusses power in a way usually ignored by other definitions.

Elias Canetti devotes a section in his 1960 work Crowds and Power to the topic of “Question and Answer”. The work as a whole deals with the relationship individuals have with the mass psychology of societies and large communities. In the section concerning questions we can gather a new perspective on the questioning process (and how it might be changing) by focusing on the interactive behaviour of individuals. Canetti argues that

“all questioning is a forcible intrusion. When used as an instrument of power it is like a knife cutting into the flesh of the victim. The questioner knows what there is to find, but he wants actually to touch it and bring it to light”

Therefore, adopting this reasoning, rather than describing questions formally or semantically we might instead describe questions as a more qualitative measure of the shifting dynamics of power between two or more individuals. Thinking for a moment not about the sentences themselves but about the contexts in which questions are utilised. Describing all questions as forcible intrusions proposes that, no matter the context or familiarity of participants, questions are always centres of power which in certain situations become conspicuously “instruments of power”. In this sense, power always underlies the questioning process and is only sometimes instrumentalised. At the heart of every question is a not only a process of empathy in order to work out what can and should be asked but also an ethics of interaction more generally.

This ethics of interaction relates strongly to the answers given by individuals or more specifically, the process of forcing individuals to respond.

Canetti says:

“even in normal circumstances, an answer restricts the movement of the person who gives it; he has to abide by it; it forces him to take up a fixed position and to remain there, whereas his questioner can shoot at him from anywhere, changing his position as it suits him. […] The ability to change his ground gives him a freedom which the other is denied. He probes his defences with questions, and, when he succeeds in piercing them, that is, in forcing him to answer, he has him pinned down and unable to move. ‘Who are you?’ ‘I am so and so’. From this time he must be himself, or his lies will entangle him. He is no longer able to escape by transforming himself.”

Questions are a way, says Canetti, to negotiate subjectivity in an active way but only by interrogating an other. Being questioned fixes us, articulates our subjectivity for us within the language of the dominant discourse and makes us passive targets.

By defining ‘questions’ in this contextual mode we can start using a definition which covers most examples (unlike semantic or speech act theory) and helps us to think about what traditional questions are.

So if we hold onto this definition for now, how are traditional questions changing into queries?

The critic Antoinette Rouvroy in a discussion of Big Data provides us with a critique of a much more general issue which she names: “Algorithmic Governmentality”, her term for when algorithms start governing our lives. An issue which, I think, provides a productive juxtaposition to Canetti.

Big Data, for anyone who doesn’t know, is the practice of using datasets larger than any human could conceptualise and larger than most computers could cope with in order to find correlations and patterns. It drives a number of contemporary tech successes, from Netflix suggestions and personalised nectar point vouchers to projects to projects predicting the international spread of flu viruses, sending aid to the most probable areas. In the words of Viktor Mayer-Chonberger and Kenneth Cukier “Big data is about what, not why. We don’t always need to know the cause of a phenomenon; rather we can let data speak for itself.”

Rouvory is very hesitant about this attitude of “data speaking for itself”, like the cake recipe example earlier, datasets and algorithms always have ideological positions. What kind of data has been collected, using what method, by who and for who and the way it is processed all represent certain attitudes. Rouvory’s fear of algorithmic govenmentality is that it is all too easy to imagine “raw data” and objective algorithms while ignoring that machines have been constructed with certain attitudes inbuilt.

So what does this have to do with Google. Well, when we search Google our query is treated with this Big Data mentality. Meaning is ignored in favour of correlation. Rouvory sees this as a problem because Google therefore can’t be held accountable for the philosophical, social and ideological structure which its search results impose. Rouvory draws on Kant’s category’s of “Judgment” and “Critique” as our usual ways in which to address the world. Whereby “judgments” structure, classify and fix the world, and “critique” addresses these categories, questioning their assumptions and stretches them out of shape, ready to be remade. Google’s way of structuring the world, however cannot be described as a Kantian “judgment” as its fluid use of user data works by avoiding fixing categories. Google’s structures instead inhabit and take over the fluid space of “critique”. As a moving and changing target it is much harder to critique “query and result” than it is “question and answer”.

Google is constantly gathering information every time we us it, allowing it to alter its results for us based on location, preferences and previous histories. Every time we search our results become more personalized, more relevant. These selective results fixes Google’s version of our identities through personalization but also fixes our identities in a very real sense through not showing us results which may not be relevant to us, because they are outside our comfort zone or portray a very different political or cultural perspective than those we have experienced in the past. Simply put, we have not clicked on that kind of thing before.

So now if we go back to thinking about Canetti’s definition of questions. Where he proposed that the questioner retains a movable identity “can shoot from anywhere” and is in a place of power while the person being questioned is with every answer interpolated “pinned down” and forced to adopt a fixed position from which he is no longer able to escape by transforming himself. Even if they refuse to answer a question they are shaped by Canetti’s formulation of “who are you?” “I am so-and-so”.

Google queries therefore completely flip this formulation. When I search, my queries fix my identity and allows Google’s “judgments” its fixing of the world, to remain fluid. Every query I ask provides Google with more information about me and in turn limits the field of view for what will be returned to me. Every query is an action of “This is me, I am so-and-so”.

Moving from questions to queries is a shift of power. It means that in trying to find something (in opening ourselves up to the world’s uncertain possibility) our identity becomes more fixed and less fluid. It is therefore the querier who declares “I am so and so” and this declaration doesn’t shape how the world sees us but instead fundamentally changes the way we see the world.

And for Google to work effectively this needs to happen, there isn’t a universal, impersonalized Google.com, and there couldn’t be. For a set of results to be relevant they need to be different for searchers around the world, in different contexts, with different pasts, desires and aspirations. Google interpolates us not only through our previous searches, but the language and phrasing we use, who it thinks we used to be and who it thinks we want to be. It forces an identity of its own construction onto us in order to provide us with what will be relevant to us.

Google wants us to be stablized and through Big Data correlations grown very good as predicting who we are and what we want. As if our very subjectivities were only an amalgamation of our queries: our memories, desires, actions and silences.

Canneti’s formulation of questions came out of the ‘60’s and a fear of totalitarian control in which our answers made us who we are. If Google has reversed this formulation, if now the things we ask, the things we query make us who we are, what does that mean in a culture that “just Google’s things”. Or in fact a culture in which individuals, even if they can help one another, respond “just fucking Google it!”

Richard Graham

Phasellus facilisis convallis metus, ut imperdiet augue auctor nec. Duis at velit id augue lobortis porta. Sed varius, enim accumsan aliquam tincidunt, tortor urna vulputate quam, eget finibus urna est in augue.

1 comment :