Wednesday, August 31, 2011

Speculations on the future of reading

In 2008, Nicholas Carr published an essay in the Atlantic Monthly titled "Is Google making us stupid?" The article got a lot of play and was later turned into a book titled "The Shallows."  At its heart, Carr's thesis is a simple one. He argues that the extensive internet reading – meaning the the copious amount of reading that we do on the internet as well as our need to be always "plugged in" into, for instance, email and Facebook – is changing the way we think. He is explicitly worried about the future of reading. He thinks that the art of reading deeply – think about being immersed in a novel for a few hours – is dying out; that, instead, reading has become a "sampling" activity: a little bit here, then a quick glance through email, another little bit there. Since reading did not come naturally to the human brain and in fact helped shaped the brain as we know it today, this new form of reading  - all stops and starts - will change it as well. If that happens, will the decline of quiet contemplative deep reading result in the decline of deep thinking? (Obviously Carr poses this question rhetorically; his answer is an emphatic yes.)

Should we be worried about this?  My sense is not really.  In my opinion, the main problem with Carr's analysis is that he relies on cognitive and neurological studies and hopes to prove that skimming is bad for us because it changes the neural structure of the brain.  This seems to me both wrong-headed and unnecessary. Certainly, studying what happens inside the head when we read (or write) is interesting for its own sake. But by understanding reading as a mental or neurological process, we are inevitably stripping the activity of reading of its context and its entanglement with our day-to-day life. It makes it seem as though reading and writing are activities that occur by themselves, independently, which is clearly not the case. In this post, I want to try and map out how reading will look like in the age of the internet.  Obviously, these are mostly informed speculations -- but I hope to show that by looking at reading in all its different contexts, we will come to realize that the future of reading is not as dire as some people imagine it to be.

Here is a first stab at delineating the contexts of reading. Reading and writing occur in a variety of places. We read at home, we read at work, we read when we travel, we read on the way to work (in buses and trains and airplanes).  Reading can be classified in terms of what we read for or in terms of what it means to us. We read for pleasure, we read for edification, we read for work. We read so that we can produce content of our own (scholarly papers, articles, reports, presentations, blog-posts, etc.) It can be thought of in terms of what we read. We read genre fiction, we read literary fiction, we read non-fiction, we read updates on Facebook, Twitter and email. We read newspapers, magazine articles and blog-posts. We read scholarly papers, white papers and consultant reports. Or we could think of it in terms of the media (or the instruments) we use: the web browser, the e-book reader, the codex, the pdf file-reader, the mobile browser, or the RSS feed-reader. Finally we could think about it in terms of time: reading in our leisure time, reading during work hours, reading on weekends and so on. Note that these are meant to be over-lapping categories (and are certainly not exhaustive). We may read for edification during our leisure time or we may read fiction for pleasure in our leisure time. We may read fiction or non-fiction for work. We read email for work as well as for our personal lives. We use the web for work as well as for our day-to-day shopping.

These contexts of reading are social structures. They structure and shape how reading and writing get done. Moreover, even a cursory examination of these contexts of reading will show that most of them predate the internet. Fears of the demise of deep reading, like Carr's, often take for granted that the only form of reading that matters is the reading of literary texts. Other forms of reading and writing in other contexts (technical reading, reading for work, scholarly writing etc.), are often not even mentioned. Which is why once we start to look at contemporary reading and writing practices as embedded in structured contexts, and as more than just the reading of literary fiction, the idea that in the age of the internet we don't read as deeply as we used to stops making sense. Instead we can start thinking about how each of these structuring frameworks have changed in the recent past and how that has transformed the practice of reading and writing.

Let's take the first and most important category: reading for pleasure, clearly the most common form of reading. When most lay-people say that they "like to read," this is what they mean: that they like to come home from work and settle in with a good novel, or that they read a novel while commuting to and fro from work every day, or that they usually go to the park on weekends and read for an hour or two. It need not be novels, of course: it could be biographies, self-help books, or books on history and politics. It could also be literary fiction although this is probably rare – genre novels like romances, science fiction and thrillers are far more likely to be read for leisure (and pleasure!) than Tolstoy or Proust. How will practices of leisure reading change with the Internet?

The amount of leisure reading that we do is already substantially less than it was before, thanks to television. Will the Web reduce it further? On the contrary, the Web might make us read more, rather than less. However it is likely that is likely that the reading of magazine articles, blog-posts and other forms of web-content will occupy a larger share of our reading time than they did before. We are also more likely to skim this web-content than deep-read it (more on this below). However the core of our leisure reading, novels and some form of non-fiction, will be the way it is now – from start to finish, in one continuous sweep – or in other words, "deep reading.” Other peripheral changes, however, are likely to occur. Perhaps we may increasingly start to use e-readers like the Kindle, rather than the codex, for our leisure reading. Perhaps the length of printed books (especially non-fiction) might be substantially reduced. Typically, the size of the book (200 pages, 300 pages, etc.) is determined by a publisher's economies of scale. If e-readers replace printed books (a very big “if”), and the cost of publishing plummets, then the length of a book could become less standardized. To summarize, some of our novel-reading will be replaced by the skimming of shorter-form web content, but most of our leisure reading will still involve long-form content, and we will consume this content by “deep reading.”

However, the share of newspapers, magazines, and blog-posts in our leisure reading is likely to increase. And day-to-day practices around the reading of newspapers and magazines will, and have, almost certainly changed with the rise of the World Wide Web. It is here – in the context of reading newspapers and magazines – that it seems to me that changes in long-established practices are occurring and where deep reading may well be on the decline. But it's not clear if this is a cause for worry. Let's examine this in more detail.

When newspapers were available only in print, the standard practice was to subscribe to one or two, usually locally available, newspapers and read them as fully as possible every day. A standard newspaper tries to cover all possible topics in the limited space available: national politics, international news, sports news, art and culture, and so on. The printed daily newspaper was also sold as one single commodity with all the articles bundled together. The assumption is that the reader will read only a newspaper or two in a day and the idea is therefore to provide him with a little bit of everything. The combination of the two factors, the reader's ability to only read one or two newspapers (or magazines) in a day (or month) as well as the publication's tendency to cover as many fields as possible, meant that the reading practices around reading printed periodicals emphasized “breadth.”

With the Web, it is no longer necessary that the day's newspaper or the monthly issue of the magazine be one packaged entity; instead individual articles or sections can be sold separately. As newspapers and magazines move to the Web, readers too do not have to limit themselves to only one newspaper or magazine. This suggests that an interested person could opt to read, say, only the economics section of many different magazines rather than reading any one magazine in full. Which, in turn, suggests that newspapers and magazines could choose to focus on certain topics in depth, while leaving others out completely, since their readers probably prefer to read about them in other places. That is, publications might become more specialized, catering to a certain niche of topics and readers. Both these trends together -- that readers can sample more publications, as well as the possibility that publications themselves might focus on only certain topics -- might result in a transition from a reading practice that emphasized “breadth” to one that emphasizes “depth”.

This brings us back to the problem of skimming. One of the central cultural anxieties today is that skimming a text is an inferior way of engaging with it.  If indeed there is a shift towards reading more newspaper, magazine articles and blog-posts online in our leisure time, and these shifts result in more depth-oriented reading that emphasizes skimming large amounts of text, should we be worried about it? I am not convinced. Here's why.

First, it's worth keeping in mind that skimming newspaper and magazine articles was something we always did, even before we started reading these on the Web. In fact, newspaper articles are structured precisely so that they can be skimmed. Second, even if one skims most articles that one reads, the sheer amount of reading one does means that it is far from shallow. But most important, this anxiety about skimming and deep reading ignores the central issue by focusing attention on the skimming itself. Instead the key issue is: what is this skimming for? How does it fit into the context of our other activities?  This, it seems to me, is what needs to be investigated empirically. How do people decide what to skim and what to read in detail? And how do they skim? And how do the artifacts of reading like the blog-post, the web article, or a Wikipedia essay structure these reading practices? Brain-scans and mental models will not help here; what we need is concrete observations of what people do and what it means to them.

Deciding what to skim and what to read deeply is always a difficult question. I would argue that this was never a problem in leisure reading before because people never had that kind of content at their finger-tips. On the other hand, the problem of what to skim and what to deep-read has always been faced by those who read for work (graduate students, academics, researchers, journalists, some other categories of white-collar workers). Let me call this category of people "discourse producers." What may be new is that the decision-making problem of what to skim and what to deep-read, usually faced only discourse producers (who are reading for work), is now also faced by lay-people (those who do not read and write for a living) -- and it is faced while reading for leisure. Note that discourse producers are also used to skimming long-form (usually non-fiction) books; lay-people will be less interested in skimming books as such, especially works of fiction. However skimming will play a big role in the short-form content that they read (blog-posts, articles, emails, etc.) To read well then, it not only becomes necessary to skim but it becomes essential for everyone (not just discourse producers) to skim well. One way to understand Ann Blair's work on note-taking in early modern Europe is to think of it this way: in early modern Europe, discourse producers started to face the problem of what to skim and what to read deeply, which is precisely the problem of “information overload.” In our age, this problem has stopped being solely the problem of discourse producers, hence we are experiencing “information overload” all over again.

This background therefore allows me to restate the worries about the demise of deep reading as follows: given that there is a possibility that the Web will involve a slight shift in our leisure reading from long-form books to more short-form content (blog-posts, magazine articles, etc.) and that this engagement with short-form content will be more depth-oriented rather than breadth-oriented, should we be concerned about the demise of the so-called practice of deep reading?

Again, it seems to me that the increasing focus on the neurological aspects of reading leads researchers and cultural critics to fetishize the act of reading itself rather than to focus on the question of what this reading is for. For instance, discourse producers have always faced the question of what to skim and what to read deeply. If the rise of the Web has increased the magnitude of this problem for them, then has their output (journalistic and scholarly articles, reports, essays, etc.) significantly diminished in quality? Even though this question is too broad to be answered in any meaningful way, it could be broken up into smaller, more empirically tractable parts (say, by looking at only journalistic output or the scholarly output of historians). For lay-people, the research questions are even harder to frame. If there is a shift in leisure reading from a deep-reading of long-form books to a depth-focused reading of short-form web-pieces, what are its implications exactly? Concerns that this may make us less “thoughtful” are too broad and frankly, too elitist, to mean much. A better research question could be: does this shift from long-form books to more short-form web content focusing on politics and current affairs make us more politically conscious? Cass Sunstein and others have speculated that the internet with its tendency to exacerbate homophily (i.e. the tendency of people to talk to people who are similar to them in some respects) may increase the political polarization of the electorate. However, empirical work on this topic is still inconclusive. We need to develop further research questions on similar lines rather than simply thinking about deep reading/skimming dichotomy in isolation.

The brief segue into political polarization allows me to bring up another closely-related issue: the problem of gate-keeping on the Web. Let's assume that the key effect of the Web has been that the traditional problem of what to deep-read and what to skim, faced only by discourse producers, is now also faced by lay-people. Clearly, discourse producers have institutionalized ways of deciding what to skim and what to deep-read. This depends on their training, their disciplinary identity as well as the task at hand. What kinds of practices will lay-people evolve to solve the same problem and how will it affect them? On one hand, this is an empirical question. But on the other hand, it could also be posed as a normative question. Will the practices of lay-people ever measure up to those that discourse producers have evolved over centuries (which Blair brings out)

Or we could think of this in yet another way.  Before the internet, there were a set of mechanisms which ensured that what people read was of a certain standard. These were mostly on the side of content-producers.  For example, newspapers have editors and reporters, each of whom are trained to judge the quality of a publication. Publishing houses have teams of editors who decide what gets published. With the internet, these mechanisms can be (apparently) bypassed because on the internet, anyone can get published (more on this below).  Instead the burden of discerning the wheat from the chaff, the "good" content from the "bad," has instead shifted to consumers of this content, in this case, the lay-public.  How do we make sure that the lay-public is able to separate the "good" content from the bad, the wheat from the chaff? 

It's worth clarifying a few points here.  First off, it just isn't true that anyone can get published on the internet. Well, this is true in a trivial sense that anyone can create a Blogspot account and start publishing their writing to the Web. But this is not the same as saying anyone can get read on the Web.  The best way to be read on the Web is to make sure that another web-publisher links to you.  And because linking is so important, the blogosphere has evolved its own informal norms around the practice. Bloggers always reciprocate with links: if X links to Y, then Y links back to X. They are very careful in attributing where they found any piece of writing. And even when they excerpt from another blog-post or news article or a Brookings Institution study, they link to it so that their readers can go ahead and sample the article itself. An article that gets linked to by many people gets read more and will also rank higher in an internet search. Being read on the internet then is a matter of convincing the right people that an author is worth reading.  And in that sense therefore, the blogosphere or the internet is no different from the rest of society or the publishing industry.  In particular, the norms of the blogosphere are startlingly similar to those of the scholarly community: scrupulous citation, an emphasis on linking to the source of a claim so that readers can judge for themselves if it is true, and so on.  What I'm trying to say here is that there is a quality-control process operating on the internet as well.  But this quality-control happens through an informal economy of links rather than through a formal, organizational process. This is why it makes people uneasy.  And yet, this is precisely the reason for its relative “openness”: newcomers have a better chance of being noticed than in the formal world of publishing.

That still doesn't tell us what we can do to help lay-people make "good" judgements about what to skim and what to deep-read - or in other words, how to establish the quality of the content they are reading.  It strikes me that we are barking up the wrong tree here.  These practices will evolve, just as they did for discourse producers.  Perhaps we could start teaching school-children how to skim as well so that they have a repertoire of techniques which they can employ when they read on the internet.  But our central concern should be with making sure that the online world remains, to some extent, "open."  The internet is not some magical medium that is innately free.  It's only what it is because of the practices around its use.  We need to make sure that the open linking practices of the blogosphere persist and are not eroded (by, say, increasing corporatization and the tendency to link only to one's own content). A second way may be to ensure that linking practices retain the element of serendipity. This means making sure that the barrier to entry remains low enough, so that content from newer people can still be read by many. It means making sure that the mechanisms of sharing (say, Facebook and Twitter) do not lead to stifling homogeneity. In other words, we need to make sure that the depth-oriented reading that the internet facilitates is not too depth-oriented, that some amount of breadth does creep in.  How this is to be done, I am not sure, but someone will think of something.