The Future of Book Recommendations

Jonathan McCalmont @ 23-04-2008

Welcome to Blasphemous Geometries, a cross-media criticism column where Jonathan McCalmont pokes the foetid corpse of genre to see what oozes out.

Blasphemous Geometries by Jonathan McCalmont

In his introductory column, Jonathan examines new ways in which retailers might decide what to place in our paths next time we’re shopping around for some sf-nal entertainment.

###

My name is Jonathan McCalmont, and this is a column about science fiction.

Well, I say that it’s a column about science fiction, but in truth it’s just a momentary distraction. A few small insights or entertaining thoughts that may – hopefully – distract you while you are at work, prompt you to write a response or consider something from a different angle.

In the old days, people used to say of newspapers that they were so ephemeral that they were nothing more than tomorrow’s fish’n’chip papers; if you tried to take fish’n’chips home in this column you’d end up with nothing more than unsightly burns.

Online columns are borrowed electrons … they are tomorrow’s illegally downloaded Lady Sovereign mp3s, or this evening’s transsexual amputee porn. They are the bacteria that consume and bloat the corpses of science and art. They are the ultimate perversion of human communication; a naked homeless person screeching his uninvited, obscene and unhinged opinions at innocent passers by.

Ladies and gentlemen… I am filth. Welcome!

As sexually and emotionally gratifying as that little exercise in self-deprecation might have been, I did have a serious point in mind when I started it – namely our relationship. If I’m to write about science fiction, then you need to know why you should pay any attention to what I say. I need to know what type of stuff you want to hear about.

Should I just cut’n’paste some press releases about the latest Australian fat fantasy novel or Vampire-shagging epic to emerge from the diseased colon of the publishing industry? Or should I simply spend my time chastising and insulting all of you that don’t own Ian McDonald’s Brasyl and Peter Watts’ Blindsight? Traditionally, a critic finds his voice and his audience finds him, but here at Futurismic we like to do things differently… we like to move things forward instead of falling into old patterns.

That is why this column is all about the scientific future of book recommendations.

We’ve all encountered this phenomenon in one way or another. In the Wired article in which he originally formulated the Long Tail hypothesis, Chris Anderson pointed out that websites make recommendations based on stuff you’ve already consumed.

Anderson’s example features obscure ska-punk bands, but the experience most of us have of this kind of auto-criticism is after we buy a book about the royal family for our mothers, when our recommendations list is suddenly swamped with biographies of people without chins. Not to mention the time I bought Thucydides’ History of the Peloponnesian War and Amazon decided I was some kind of Nazi-sympathiser with a bottomless interest in tanks. Screw you, man … I am not my last six purchases!

The problem of course is that the links between the different products aren’t based on anything other than guesswork. It’s hard to get excited about a recommendation that states that a fondness for light bondage and the works of G. K. Chesterton mean that you should read M. John Harrison’s Viriconium books. Clearly what is needed is a more concrete basis for recommendations, maybe one similar to the one they use down in Human Resources. How about psychometrics?

Nowadays if you apply for a job or sign up on a dating site, the first thing you’ll be asked to do is fill in a questionnaire so that they can determine your personality type. If personality types can be used to determine such unimportant facts as which jobs you should have and who you should date, why can’t they be used to determine important matters – like whether you should devote the next fifteen years of your life to fighting your way through the eight hundred-odd volumes of Robert Jordan’s Wheel of Time?

While the idea of a personality type probably goes back to ancient Greece with Hippocrates’ Four Humours, the idea in its modern form is most closely associated with the Myers-Briggs Type Indicator (MBTI), a questionnaire originally used to objectively categorise people according to the personality theories of Carl Gustav Jung.

The MBTI gives us the distinctions between introverts and extroverts, sensation and intuition, thinking and feeling and judging and perception – as well as the sixteen types of personality most of us will be familiar with in one form or another (although there are very nearly as many competing theories of personality as there are Star Wars tie-in novels).

Apparently an introvert’s flow is directed inwards towards ideas because introverts draw their energy from being alone. Conversely, extroverts flow outwards as they draw energy from interacting with people and things in the outside world. Might these attitudes not map onto tastes in books?

For example, one could argue that introverts might be more likely to be interested in escapist fiction as it would provide them with tools with which to withdraw inside themselves. Similarly, extroverts might fail to understand the need to escape and might want books that actively engage with the real world. Alternately, someone who has a noted tendency towards making decisions based on feeling might find it difficult to read books without protagonists they can associate with, while thinkers might fail to engage emotionally with what they read, preferring to consider books as fictionalised collections of ideas and arguments.

Obviously this approach is problematic. So problematic, in fact, that I’ve struggled to find any trace of anyone pursuing it as an avenue of research – despite the obvious marketing possibilities.

For one, it’s not easy to encode a book. Most books do a number of different things in a number of different ways – and that’s without mentioning the fact that there’s a degree of drift within anyone’s personality. This is beautifully demonstrated by the Forer effect, which showed that if you give people a vague assessment of their personality they’ll most likely think it’s accurate, as there are times when we’re more introverted than extroverted and more intuitive than sensing.

As a result, most attempts at automated book recommendation tend to revolve around what you’ve previously read or what vague categories you feel most attracted to at a given time. But of course, these types of systems will – at best – only ever tell you about books that resemble other books you like. They’ll never reliably tell you what you should be reading.

Just think – right now you could be neurologically attuned to the works of Laurell K. Hamilton. All your life you’ve been unhappy and never known why … you’ve experimented with sex, drugs and even religion but the one thing that might bring you true happiness is a story about a wereleopard who is all angsty because he’s got such a gargantuan georgie-porgie that he couldn’t get any proper sex with girls until he hooked up with Anita “fill my cavernous lady-bucket” Blake!

Given that I can’t go by your personality, what other means could I – as a spunky young futuristic SF critic – use to tell you what kind of books you should be reading? How about that old zombie favourite … the brain.

One of the problems with advocating one book over another is the idea that all I am doing is offering an opinion, and that all opinions are effectively equal. So why would you listen to me rather than the New York Times, or that creepy guy at the bus stop who always smells of milk?

One answer to that question is that I might be a more reliable source of advice; I might be less likely to be bamboozled by marketing (sometimes the publishers send us reviewers packets of sweets … I’m still waiting for the uncut cocaine and highly-skilled courtesans) or hype (I hear that Richard Morgan’s new fantasy novel The Steel Remains is going to be dark and gritty!) or generally better able to perceive an aesthetically pleasing book than the guy who decides what books get laid out on the tables at Waterstones.

The growing field of neuroaesthetics aims, as Oliver Elbs puts it, to supercede the maps of art that have been generated by art and music critics and replace them with maps based upon what said art and music does to the brain. Sadly, as Martin Skov points out in his excellent primer to neuroaesthetics, little research has been done into how the brain reacts to literature.

One suspects that this is partly because the senses are currently better understood than the more abstract processing faculties that are used when reading, and therefore it’s easier to produce testable hypotheses about how we appreciate art than it is to produce hypotheses about how we appreciate literature.

However, while the science may not yet be there, this has not prevented some critics from jumping into the future and deploying neuroscience as a theoretical lens through which to observe literature. Consider, for example, this piece by A. S. Byatt about the poetry of John Donne. It is full of talk of the brain, and how the poetry conforms to the aesthetic principles built into it by our evolutionary history.

So, according to this model, I could recommend books to you based on how a particular book stimulates your brain. Maybe at some point in the future it might be possible to go to a local bookshop, stick your head in a hole in the wall and receive a print-out of all the books that would suit your precise neurological parameters.

Unfortunately, this future still seems like a long way away.

The neurologist, doctor, poet and swordsman (okay … not that last one) Raymond Tallis has recently reacted to neuroscientific attempts to colonise the humanities by pointing out that neuroaesthetics is nowhere near being able to explain human consciousness fully and – even if it could – that a purely physical description would be overly reductive.

This means that not only are we nowhere close to being in a position whereby I can say to you “read Rudy Rucker’s Postsingular, it will stimulate parts of your brain other books cannot reach”, but even if we could reduce talk of books to talk of the brain, there would still be stuff left unsaid in the same way as a full explanation of seeing the colour yellow could not capture what it feels like or means to see that particular colour.

So it would appear that – for the moment – you are stuck with the likes of me and my hopelessly old fashioned attempts at convincing you to read one book rather than another.

I’m not much happier about it than you are … I could be watching Mysterious Cities of Gold on DVD. *sigh*

###

Jonathan McCalmontJonathan McCalmont is a recovering academic with a background in philosophy and political science. He lives in London, UK where he teaches and writes about books and films for a number of different venues. Like Howard Beale in Network, he is as mad as hell and he’s not going to take this any more.

[ The fractal in the Blasphemous Geometries header image is a public domain image lifted from Zyzstar. ]

Be Sociable, Share!

3 Responses to “The Future of Book Recommendations”

  1. Fredosphere says:

    Dude, the solution is encapsulated in just two words.

    Two words. Are you listening?

    Collaborative filtering.

  2. Ian Sales says:

    I bought a book on Amazon once and Amazon told me that “people who bought this book also bought underpants”. I remember it being quite a good book too…

  3. Paul Raven says:

    Was it a scary book, Ian? That might explain it …