Still On Reviews

Posted: October 21, 2009 in Braak
Tags: ,

It’s tricky for me, once I start thinking about a thing, to stop thinking about a thing.  Probably because I never reach any actual conclusions, and this hinders my ability to establish closure on subjects.  But I was talking about reviewing on Monday, and there are a couple of things I still want to have some words about.  Specifically, I want to go back to utility reviews, and how they’ve been affected by Amazon.

So, there’s a couple things here.  The first and most important thing is this:  art (er.  Entertainment, anyway.  Art, literature, movies. &c.) is how human beings learn morality and identity.  It’s been this way forever, whether we like to admit it or not.  Identity–both personal and social–is what literature is FOR.  Now, interestingly, in Olde Time Dayes, most of the people that were familiar with a piece of art were also neighbors.

There were traveling theater troops, obviously, starting in about the 15th or 14th century (and probably linked to a troubadour tradition that was a few hundred years earlier), but that’s actually pretty late in terms of the invention of art.  Nonetheless, while these troupes and troubadours got around, they didn’t get around that much.  For one thing, they had to walk everywhere.

Likewise, even after the invention of the printing-press, it was still a while before books became a medium of mass-entertainment; they were expensive, and most people were illiterate.  It was theater up until about the 17th or 18th century, and even then, it was still theater for a while.

So, okay, bear with me here.  Once books start becoming popular, there are two, and basically only two ways for people to know what books are good:  word of mouth and newspaper reviews.  Both of these review models have a very short range–they are city or social-circle specific.  Yes, the books get spread around–but the machinery for understanding and appreciating them doesn’t, and it’s going to be vastly different from locality to locality.

Obviously, that’s not how it works anymore.  The internet happened, and now instead of individual localities gradually collecting and interpreting the literature that defines them, we’ve started working the other way around:  picking the literature that we like and allowing ourselves to be defined by it, then connecting to people (sometimes thousands of miles away) and forming an ad-hoc identity group through it.

This is what I think is interesting about Amazon.com.  Amazon has perfected a system of “utility reviews”:  not reviews that provide in-depth criticism, not reviews that are objects of entertainment in and of themselves, but short, shallow descriptors that are geared entirely towards telling people whether or not they’ll like the book.  People rate the book, then they rate the ratings, and theoretically the most helpful ratings will get sorted to the top, and then the people that might like the book will know.

Of course, all people don’t read all books.  Amazon has got their fancy recommendation system–which I quite like, actually, it’s how I found Jeff VanderMeer and K. J. Bishop–that works exactly the way the human brain works:  it takes a familiar data set and makes inferences based on it, updating its predictions according to positive or negative responses.

What does this mean?  All right, well, we’ve been assuming that every human being’s taste is fairly unique; but if the point of art and literature is to build social identity, obviously that can’t be true–there have to be some similarities, at least.  I think that an in-depth study would find there’s actually very little variation in taste.  Rather, there are something like sixteen “exemplars”–sixteen maximally-different book-buying patterns, with the variant patterns clustered around them in increasingly rarified circles.

The Amazon utility reviews that come in on books are mostly going to come from the cluster-patterns that prioritize those books.  If you mostly buy cookbooks, for instance (this is a hyper-simplified example, obviously), you’re less likely to buy Excavations at Ur, and similarly less likely to write anything about it.  Mostly, except for the handful of rigorous review writers, people write about things that either exceed or fall short of their expectations:  they write that a book is mediocre or bad if they expected it to be good, that it’s good if they expected it to be mediocre.  

If you’re buying outside the focus of your buying-pattern, you’re less likely to have any expectations about the nature and quality of the book, and similarly less likely to have the language necessary to describe it.  Even if you DO write a review, it won’t be a robust review–and the other people for whom this book is part of their pattern-focus will just rate your review down, removing it from consideration.

Regular review writing that happens in magazines and newspapers and on blogs and such has to be an object of entertainment in and of itself–but the nature of entertainment is not universal.  It is based on the pre-existing, anticipated elements in the psyche of the audience.  Amazon reviews are not entertaining, they’re purely utility–but they operate at a thoroughly unprecedented scale.  This scale is large enough to become, practically speaking, a kind of engine of identity:  a self-selecting and self-reinforcing system.  If you accept Amazon’s recommendations and reviews, you will drift closer and closer to one of the maximally-different exemplars, and subsequently writing reviews that are a product of that exemplar pattern.

All this time, I thought the functions in Amazon were there to sort books; really, the systems are there to sort US.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s