Bubble Trouble

by Henry on September 5, 2011

The American Prospect has published a review essay I wrote on Eli Pariser’s The Filter Bubble (I’m not quite sure when this went up on the WWW; I’ve been travelling). It’s an interesting book, which takes some of the empirical developments that Tyler Cowen enthused about in The Age of the Infovore and comes to diametrically opposite conclusions about their normative implications.

What Cowen sees as enhancing individual autonomy, Pariser sees as restricting personal development. Instead of constructing personal micro-economies that allow us to make sense of complexity, we are turning media into a mirror that reflects our own prejudices back at us. Even worse, services like Google and Facebook distort the mirror so that it exaggerates our grosser characteristics. Without our knowing, they reshape our information worlds according to their interpretation of our interests … We are beginning to live in what Pariser calls “filter bubbles,” personalized micro-universes of information that overemphasize what we want to hear and filter out what we don’t. … Cowen’s ideal world—where the private vice of self-centered information leads to the public virtue of a lively interactive culture—is unlikely to be self-sustaining. It’s also difficult to see how regulation could pop information bubbles. … As Harvard political theorist Nancy Rosenblum has argued, partisanship creates its own checks and balances. As long as partisans are contending for a majority of public support, they have to temper their own beliefs in ways that will allow them to appeal to the public and to respond to potentially persuasive arguments from their opponents. Democratic competition is not a complete solution. It does not protect individuals from a narrowing of their horizons. … Even so, democracies are far more robust against information bubbles than Pariser believes.

{ 33 comments }

1

ezra abrams 09.05.11 at 10:22 pm

am i being to snarky in saying that anyone who uses the word normative, outside of highly technical specialized writing, is a dunce ?
I mean really, its almost as bad as the term college admissions officers use “rising senior”
gag me with a spoon

well, I read a little more and go this:
we are turning media into a mirror that reflects our own prejudices back at us
uh, hello, what do you think media has been doing since the pharoahs sent out the daily papyrus ? almost all media exists to support the prejudices we have already; no one spends good money on something that regularly offends them….

2

Steve Laniel 09.05.11 at 10:33 pm

The central thesis, and its political implication, seem like quite loud echoes of Sunstein’s “Republic.com” (available for a penny used on Amazon now). Yet I don’t see it mentioned in the review at all. If you’ve not checked it out, you might like to.

3

tomslee 09.05.11 at 10:41 pm

Cowen’s ideal world—where the private vice of self-centered information leads to the public virtue of a lively interactive culture—is unlikely to be self-sustaining

I wonder about this. It’s easy for technologists to see content selection as a problem for a matching algorithm, where we want to see more of what we previously liked. But that’s really just a guess isn’t it – never mind the implementation? Tech companies take the same attitude to advertisements (“relevance” being the key) but this paper by Joseph Turow and others, that Alex of this parish linked to shows that we dislike them.

What evidence is there that we enjoy life in our bubbles?

4

Andrew F. 09.05.11 at 10:53 pm

Oh for the pre-internets days when reported information was more accurate and individuals were more likely to be confronted with contrary facts or different viewpoints.

I too am gravely worried by the obvious tendency of Google to personalize my searches. Soon Wikipedia will be inaccesible via search – who wants a neutral viewpoint or facts? Snopes and politifact are clearly on the verge of extinction.

I yearn for the time when we would adventurously seek out different demographics, not self-select our friends and acquaintances, and subscribe to multiple newspapers for the edification of clashing editorial viewpoints.

And, obviously, the difference between Google ranking certain political sites higher based on what I always click, and Google flooding me with links to blogs I have no intention of visiting, is profound. Surely if you show an avid reader of Mother Jones enough links to the Weekly Standard, he’ll click them – even though he doesn’t want to click them.

If anything Henry, I think you were too mild in your criticism.

5

C.P. Norris 09.06.11 at 12:39 am

The “filter bubble” argument always seems to be in comparison to a world where all Americans get their news in perfect balance from Peter Jennings. And maybe that was a better world. But that form of news was a relatively recent invention. Before that, people got their news from partisan newspapers. Or they just learned everything they knew from their parents and the parish priest.

My RSS reader is pretty biased and partisan, but it’s still less filtered than what my parents or grandparents read at my age.

6

tomslee 09.06.11 at 12:55 am

#4 and #5.

I don’t really care whether the Internet filter bubble is a bit bigger or a bit smaller than the Peter Jennings filter bubble. You don’t have to yearn for ye golden age of broadminded yokels to be concerned about the argument. And the TV vs Internet phrasing is a bit 2004 anyway: Google is TV’s friend, as Eric Schmidt told the UK television industry in Edinburgh last week.

7

piglet 09.06.11 at 1:29 am

A really interesting question, and one to which I haven’t yet seen an answer, is why the hugely increased availability of information hasn’t made us collectively better informed although we personally feel much better informed than we ever were (I assume that is true of most who post here).

8

CHRISHHH 09.06.11 at 3:04 am

These “filter bubbles” are merely the reflection of what social products we have at this particular moment. I mean, you could *easily* imagine a Wikipedia-like Facebook product that operates totally differently than what we got now.

For instance, if I could quit my job tomorrow I’d make a timeline-based personal history social website. You could write your own history, which would satisfy the narcissists out there — but it’d also allow anyone to contribute to entries for events you’ve described. Events would be open to all, and searchable and editable by time and location.

You could imagine how this type of social product would allow for contribution and dialog, rather than fortress-building. I think such sites will be more common, as they encourage debate (something folks seem to like) but will also tear down the “bubbles.”

9

Salient 09.06.11 at 3:33 am

What evidence is there that we enjoy life in our bubbles?

The Internet has made us much more social, and almost all forms of voluntary socialization are bubble-forming, bubble-reinforcing, and life-satisfaction-improving.

why the hugely increased availability of information hasn’t made us collectively better informed

Perhaps it’s better to ask why hugely increased availability of community interaction hasn’t made us ‘better informed’ — I’d say, it has, if you allow ‘what acquaintance X thinks about topic Y’ to be an acknowledged valid form of information. I know quite a lot more about the thoughts and perspectives of quite a few people with whom I’d be pleased to share conversation over lunch, and that’s mostly what I’m looking for from the Internet. (I doubt I’m in the minority there.)

10

Chris Bertram 09.06.11 at 6:51 am

I think it is worth combining this with something else you wrote Henry, which, iirc, showed that an universal expansion of personal choices was compatible with a narrowing of the global range of options.

11

Phil 09.06.11 at 7:37 am

Let’s say, for the sake of argument, that the ‘filter bubble’ model is valid & generalisable. Then take a couple of examples from the piece – MoveOn as an example of polarisation leading to activism & enriching the broader debate, & the ‘bubble’ represented by political parties in the era of mass membership. Your conclusion is that “democracies … [have] survived bubbles for hundreds of years” – but the bubbles you’ve described are precisely those which feed back into the public sphere & form the terms of democratic debate. The question is whether that’s still happening, or can still happen. (On the other hand, Tom might ask if there’s any reason to ask this question other than habitual doom-saying, and he might have a point.)

Tangentially, I think you misrepresented the ‘party social life’ phenomenon, making it sound as if it was just one of those things parties do. In the case of Italy (the example I know most about), the Communist Party after Liberation worked long and hard to set up a network of local bases with their own social life – not because it was something other parties did (which it wasn’t) but because it was something the Church did. The story of Labour and Co-operative social clubs in Britain isn’t all that different – they took over where Sunday schools left off (the relationship with the churches was less directly oppositional in this case).

12

Henry 09.06.11 at 1:06 pm

Hi Phil – the point that I was making was a weaker one than the one you think I was making. It was simply to say that democracy can survive having people in different ‘bubbles’ as long as politicians have some incentive to reach beyond those bubbles as well as within them in order to garner votes. And I’d actually see the history of Italy a little differently than you do – I think that this is less about the creation of partisan forms of social life post-WWII than the revival of local bases that had come into being in the nineteenth century, in a duel between ‘reds’ and ‘whites’ where the latter were responding to the former as much as vice-versa. This is the Arnaldo Bagnasco story anyway, and it seems to ring true to me.

13

Henry 09.06.11 at 1:28 pm

Chris – was it this piece that you were thinking of?

14

Tangurena 09.06.11 at 2:47 pm

Oddly enough, I recently finished two related books: The Big Sort, and Bowling Alone. One of the premises of The Big Sort is that people were moving into like-minded communities, which resulted in that kind of filter bubble occurring in meatspace. I suspect that the issues raised in Bowling Alone (that people were withdrawing from community and communal events) is driven by the fact that folks don’t feel like we’re all Americans (did we ever? or was it only a mythological nostalgia that says we were?), and that a more us vs. them attitude has been going on for a long time. While the internet could be used to break down cultural and tribal boundaries, what it appears to me to be going on is the opposite: reinforcing attitudes and biases to the exclusion of other information.

This self-reinforcement may have unhappy consequences for politics. Pariser, who is inspired by the philosophy of John Dewey, argues that ideological opponents need to agree on facts and understand each other’s point of view if democracy is to work well.

I think unhappy consequences is an understatement. I also think that a model for the spiraling decline in discussions between ideological opponents is that of the abortion debate. Neither side is able to actually talk to the other side anymore, and neither side really wants to bother. Folks who try to bridge the gap get marginalized by both sides.

And as for agree on facts, I claim that the model to understand is the delaying tactics of the tobacco industry. Scientific evidence that tobacco caused disease started appearing at the end of the 19th Century, yet they spent a fortune on advertising and false science trying to convince the public otherwise. And they succeeded in delaying the inevitable for almost a century. The books Trust Us, We’re Experts and Toxic Sludge Is Good For You show how those tactics of muddying issues and denying evidence have been appropriated by other industries.

15

Chris Bertram 09.06.11 at 6:17 pm

No, not at all. I’ll try to find it now Henry.

16

Chris Bertram 09.06.11 at 6:42 pm

No, I’m not finding it. Anyway, I think it was you Henry who pointed out that an increase in the range of available consumer choices for individuals is compatible with an increased global homogenization of, say, styles of music or other cultural goods. The internet makes each of us aware of more of what was previously only locally available, so we get to choose from the whole world, but within each local segment variation goes down. Something like that.

17

tomslee 09.06.11 at 7:26 pm

Chris: my ego is telling me that you must mean this one.

18

Chris Bertram 09.06.11 at 7:45 pm

Yes Tom, that’s the one!

19

piglet 09.06.11 at 9:23 pm

Tangurena 14: I found that the Big Sort was quite shallow and very weak on evidence. I think the only actual quantitative evidence for the “Big Sort” was electoral clustering.

Salient 9: I don’t see any evidence that the internet has made us “much more social”.
Also, I agree that many of us individually are better informed thanks to the internet but collectively that doesn’t seem to be the case, which is precisely my point.

20

Salient 09.07.11 at 1:01 am

I don’t see any evidence that the internet has made us “much more social”.

It might be that you’re limiting ‘social’ to ‘face-to-face.’ Getting away from that, the internet’s a popular communications tool that has made many forms of previously impossible interaction possible, and quite a lot of people take advantage of that regularly. (I’m too lazy to pre-emptively go find evidence of this assertion, ‘least not until someone explicitly disputes it. It seems pretty self-evident.) I’m pretty sure the Internet has only made a negligible number of people less social, so long as we’re careful to acknowledge any form of potentially two-way or multi-way communication as equally social, e.g. [1 hour of reading blogs and commenting when moved to do so] = [1 hour of listening to a friend over coffee and interjecting when moved to do so]. IMO interacting with someone over the ‘net really shouldn’t count for less than interacting with someone face-to-face.

I agree that many of us individually are better informed thanks to the internet but collectively that doesn’t seem to be the case

Okay, let me acknowledge upfront that folks are not terribly well-informed about the consequences that follow from achieving specific political objectives, relative to what either of us would consider satisfactory. But a lot more people are better informed about, for example, the lyrics to their favorite songs, the goings-on of sports franchises they support, current weather conditions in the cities where their distant relatives live, etc.

Ok, the topics I just listed aren’t really germane to this discussion, set them aside. There’s one topic that is germane, which I tried to mention earlier — a lot more of us know a lot more about the opinions and thoughts of people we like, specifically including the things that people we like and identify with believe are true. (Truthy things, if you like Colbert’s neologism.) There’s a natural shortcut in the brain, accepting that category as a stand-in for things that are true. Beliefs stand in for facts when their source is well-liked. My guess is that you’d assert Obama is a socialist doesn’t count as information, and a person doesn’t count as better-informed if they receive and accept an Obama is a socialist message from their favorite blog. We agree, a lot of those opinions and thoughts are based on false or questionable facts, and/or derived via faulty or sloppy reasoning, but I’d say the recipients of that information are highly unlikely to care about that (they’re likely to feel a natural disdain for the source that conflicts with the source they like).

So the problem’s not really how do we achieve a more informed or less ignorant population, the problem’s how do we successfully change what people want, so that they want to obtain the kind of information we feel is relevant and vitally important? How do we get folks to desire the types of information we want them to desire, enough to at least google around and navigate through easily-available sources, critically but receptively? The distinction is important: improving access to information won’t matter nearly so much as constructing a context in which that information is desired and corresponding misinformation is heavily disdained.

People find bubbles they like, and the more we enable and expand forms of social interaction, the more likely it is that people will successfully find and embed themselves into bubble-communities they like. So our goal should be to make sure folks have positive early life experiences that motivate them to seek out bubbles we approve of, or at least, bubbles we don’t disapprove of.

(Peoples’ bubble-preferences can change in adulthood, but I’m skeptical that peoples’ bubble-preferences can be changed in adulthood with any consistency. People are fairly resilient to that kind of pressure.)

21

tomslee 09.07.11 at 1:20 pm

It might be that you’re limiting ‘social’ to ‘face-to-face.’ Getting away from that, the internet’s a popular communications tool that has made many forms of previously impossible interaction possible, and quite a lot of people take advantage of that regularly.

I think this skews the answer. It goes like this:
1. Let’s say social = face-to-face in the offline world
2. Internet lets us do all kinds of new things (twitter, disagree with people in comment threads, etc etc)
3. We must redefine social to include all those internetty things, and then we see how social we are.

But this neglects non-face-to-face forms of communication in the offline world. Does being in a crowd with 20,000 people at a football game count as 20,000 contacts? How about walking down a busy main street with a thousand others? What about posting a “has anyone seen this cat?” appeal on a lampost? I think if you are going to redefine social for the Internet, you have to redefine it for offline forms of social interaction too. And I have no idea how you do the comparison after that, at least in a one-dimensional framework of “more” and “less”.

22

Alex 09.07.11 at 1:38 pm

The “filter bubble” argument always seems to be in comparison to a world where all Americans get their news in perfect balance from Peter Jennings.

Yeah. Its implicit assumption is a perfectly balanced liberal republic with an electorate in which preferences are normally distributed along an axis from, eh, Republicans to Democrats, what are you a commie or something? And where the media delivers all white citizens a perfectly balanced evening news report on the horrific and utterly stupid war their government just kicked off in Vietnam.

23

piglet 09.07.11 at 3:20 pm

Salient, when I say the internet hasn’t made us as a society better informed, what I mean is something like this: http://www.wired.com/culture/culturereviews/magazine/15-07/st_infoporn

I should recognize that I’m using a narrow definition of what I consider “well-informed”. Maybe you are right and people are in fact better informed about the weather than they used to be, but not about politics. Otoh newspapers had weather maps before the internet and whoever was sufficiently interested had access to that information. Also, the weather has itself become political. Do Americans know that the current Texas drought is the worst on record? I doubt it. The deeper problem may be, as you suggest, why most information consumers don’t seem to care about verified information. Verifying information by using simple internet search tools is a habit that I suspect the vast majority of internet users hasn’t acquired. Maybe there’s nothing surprising about that. Maybe it has always been thus and it would be naive to expect anything else.

24

tomslee 09.07.11 at 3:31 pm

@Alex – I see the opposite case being made too. The old world was Mainstream Media Establishment Opinion Delivered By White Guys In Suits, and the new world – in contrast – is the Non-Establishment Voice of Real People on blogs and so on.

But this suffers from the same issue as I pointed out at @21 – if you are going to include niche media on the Internet, you have to include niche media in the offline world too when you are doing your comparison.* Does a modern-day Chomsky** or I F Stone have more or less impact than the Vietnam day equivalents? Has Wikileaks had more or less impact than the Pentagon Papers?

Personally, I think that the technology reshapes the channels through which we operate, but I don’t see it altering the overall balance between establishment and anti-establishment views very much.

* Self-promotion alert: I tried to sum up the general fallcy here.
** Yes, I know Noam is still out there doing his admirable thing, but you get what I mean.

25

Salient 09.07.11 at 4:28 pm

I think if you are going to redefine social for the Internet, you have to redefine it for offline forms of social interaction too.

Nope, because my definition is conduit-independent. Social behavior = intentionally exchanging nonphatic sentences. I’m using phatic to mean something like “minimal polite formalities.”

There are probably cases where that equivocation is counter-intuitive, but I think it gives a pretty reasonable definition of social, and it definitely addresses all the cases you suggested. (Being in a stadium with 20,000 fans is surely social, unless you don’t speak at all with any of the people around you, which I think would be antisocial behavior. Attending an orchestra concert is probably a better example for what you were listing, since in that case the default assumption is that attendees won’t chat with anybody.)

Maybe you are right and people are in fact better informed about the weather than they used to be, but not about politics.

People seem to uniformly highly value having access the opinions of folks they like and highly value having the opportunity to voice their support or concern (consider that we’re both exchanging opinions here, for example). The internet has provided all manner of new opportunity to [1] discover new folks to like, who you can fairly effectively pre-screen to ensure they share general characteristics with you, [2] access their opinions, and [3] respond to their opinions with your own. We probably agree that this much is true, so the question would be whether it’s relevant to the low-verified-information problem (and honestly it sounds like we probably agree pretty thoroughly in spirit, and only disagree over some nuance in how to best characterize and address the concern that we share).

Maybe it has always been thus and it would be naive to expect anything else.

Nah, consider — one second grade teacher with a forty-year career and a measly 10% success rate at instilling those values would still teach hundreds of kids to value what we’d like them to value — and we can teach teachers to do that, assess effectiveness, refine our teacher-teaching, fight school boards to put more and better lessons about those values in mandatory lesson plans, etc. There’s room enough for hope and optimism.

26

Kaveh 09.07.11 at 4:58 pm

My the problem with the filter bubble criticism is that it treats people too narrowly as consumers of culture, not as potential producers. It assumes that the only benefit of removing gatekeepers is from the range of choices that makes available to (and then, actually taken by) potential consumers, as opposed to directly valuing their participation in production. If I’m writing a blog post that will be read by a few dozen people, that’s time I can’t spend reading novels.

What about a world in which everybody consumes a bit less culture, and (in some senses) more narrowly than they used to, but many times more people are producers of culture in some capacity, able to reach an audience of some sort, even if it is a very small audience? Isn’t that what we’re actually looking at? That doesn’t directly refute the claim that we are enclosing ourselves in bubbles, but it prevents an apples-to-apples comparison of people in smaller bubbles w/ a mic to people in bigger bubbles w/o a mic.

Salient @20 Online interactions are equal to offline ones for purposes of being exposed to new authors/information, but while they certainly have an emotional and affective dimension, they are definitely not the same as offline interactions in that respect (I won’t say “unequal”, although one probably could say this).

27

tomslee 09.08.11 at 12:10 am

Salient – how has your day been? And how are the kids (should you have any)?

Social behavior = intentionally exchanging nonphatic sentences. I’m using phatic to mean something like “minimal polite formalities.”

To me, that reads like you are deeming information exchange as significant, while social niceties are relegated to the sidelines of triviality. I would have thought that “social niceties” – much of which we do offline — are by definition a part of “social behaviour”. In fact, without phatic sentences (now I know what it means – thanks for introducing me to the word) can any exchange be called social? I can exchange information with a bot.

I’ll buy you a drink if you are in my neck of the woods sometime.

28

Salient 09.08.11 at 12:32 am

tomslee – Yeah, a few minutes after writing that I thought, “wait a minute, why did I exclude phatic interactions? That was silly. Worse than silly. Phatic interactions are a core component of socializing.” I think I had originally meant to do somewhat the opposite, carefully giving some kind of sympathetic credit to people who might exchange “hi” with coworkers on the way in yet feel lonely and under-socialized, but [1] I’m not sure there was any reason to be preoccupied with that and [2] my refusing to acknowledge what little social interaction such a person does have in their live is obviously not a good way to be sympathetic.

Kaveh — definitely, it’s natural for people to value certain forms of social interaction over others, so “hi / hi / [walk past each other]” probably counts for something but not very much, unless someone happens to really like that exchange. Obviously there’s no point in being prescriptive about how a person should order the value of various forms of communication, I just suspect that if we tried to collect empirical data, a plurality of people would rank internet exchanges much much much more highly than most folks tend to presume off the cuff.

29

BJN 09.08.11 at 2:49 am

Tom,

Did you see the recent Pandagon post “The Internet Is Good for Offline Living”

http://pandagon.net/index.php/site/comments/the_internet_is_good_for_offline_living

Thoughts? We seem to be having parallel discussions about the internet and political life and the internet and other life. At least on the latter point I think Amanda is pretty convincing. The modern era has taken away a lot of hassles that we seem to forget we ever dealt with before.

30

rick 09.08.11 at 11:57 am

I made it through this thread, including most of the links!

1
ezra abrams 09.05.11 at 10:22 pm

am i being to snarky in saying that anyone who uses the word normative, outside of highly technical specialized writing, is a dunce ?

I thought about this comment. It is precisely the use of terms like “normative”, “deontological”, “eschatological” and now, “phatic”, that keeps the trolls away. It’s very hard to comment on posts and threads if every sentence has a word you don’t understand. In effect, “vocabulary as defense against trolls”.

Enjoyed the discussion, but haven’t formed a real opinion yet.

31

tomslee 09.09.11 at 3:19 am

I’m in moderation but don’t know why.

32

Bill Benzon 09.09.11 at 10:20 am

To the extent that this ‘filter’ hypothesis implies that people are the helpless victims of their informatic environment, it’s nonsense. People who want to remain informatically insulated have always been able to do so. People who want informatic diversty have always been able to find it, no more than ever.

Further, to the extent that this ‘filter’ hypothesis implies that we are helpless informatic victims, it is not only wrong, but dangerous. For it continues the pervasive post-modern notion that we are always already the hapless victims of hegemonic systems beyond our control and even beyond our ken.

33

tomslee 09.09.11 at 12:15 pm

Bill – from what I can see (haven’t read the book) there’s nothing here about “helpless”. There’s no prison, just a continual nudging towards the centre of our own road. The relevant question is not “can a person who wants informatic diversity find it?” but, “how has the awareness of the population as a whole changed?” and those are very different.

I’d write more, but having two comments filtered out means I’ll keep it short and link free.

Comments on this entry are closed.