Text Patterns - by Alan Jacobs

Monday, April 12, 2010

"a colossally elaborate manipulation"

Tim Burke is one of the most consistently thoughtful bloggers I know — and I’m choosing that word carefully: Burke doesn't blog very often because he actually takes time to think before posting. The problem with the web, of course — and especially with the world of blogs — is that not many people follow Burke’s example. In a recent post, he gives a good reason why this is so:

There’s really very little to be said for trying to carry on a conversation (online or otherwise) with people who have nothing but an instrumental view of conversation as a means to their own anti-pluralistic or illiberal ends, who concern-troll every debate in the hopes of getting someone to take the bait. There are a set of writers who work hard every day trying to create a framework where the only right answers can be some kind of dogma, who will never for one passing second acknowledge the legitimacy of evidence which contradicts their own pet doctrines, who are never even momentarily in any danger of being persuaded by any countervailing viewpoint. For these writers, all online discussion is a colossally elaborate manipulation. I spent too much time in developing this blog arguing for an indiscriminate openness to conversation. Pursuing conversation with the comprehensively dishonest is a fool’s errand, and I’ve sometimes been just such a fool.

This seems exactly right to me. In an environment dominated by “ideological amplification”, very few people have any interest at all in thoughtful conversation, and they are hard to find in the midst of all the shouting.

I have thought about this a lot, with no results. I have argued for years that the post-plus-comments model is fundamentally broken — it works fine for a blog with a readership the size of this one, but it simply doesn't scale — but I can't for the life of me come up with any alternative to it, except the famous Slashdot karma model, which has the opposite problem: it works only at a very large scale.

When I wrote for The American Scene, for a time we had a wonderful community of conversation and debate, but then the comboxes were overwhelmed by trolls and other unhelpful voices, and became unreadable. Sure, I could look for the remaining thoughtful commenters, but only at the cost of having to wade through a great deal of garbage to get to them. Increasingly that came to seem too much trouble; and commenting itself came to seem too much trouble for a number of the people I most valued. If a typical post had three hundred comments instead of thirty, a Slashdot-like system could have filtered out the crap and left me with a reliable body of interesting comments to read; but with just thirty comments, one thoughtless vote can have a disproportionately great effect.

I have returned to this topic many times over the past few years, because I can't find any answer to these problems. The person who figures out a new architecture for online communication that encourages real conversation and filters out the trolls will have performed a great service for humanity. Though of course the trolls are always with us.

11 comments:

  • except the famous Slashdot karma model, which has the opposite problem: it works only at a very large scale.

    Perhaps what is needed is a way to switch between the basic comment model and the karma model as the number of comments increases. For example, a basic combox would appear until a predetermined number of comments were received (say, 25-30) and then it would automatically change to a karma ranking system.

    I'm not sure how feasible it would be to implement but it seems like the type of feature that could catch on fairly quickly.

  • I'm not quite as convinced that the model is broken. I write for a Gawker Media site, and spend a fair amount of time in those comment sections, and for all the noise, I've seen a lot of very fruitful conversations take place. In part, I think this might have to do with the brand -- it's sort of involved-cool-media enough to scare away a lot of the folks who comment on newspaper websites (and seriously, are newspaper site comments not frighteningly bad?), maybe.

    The Gawker sites aren't as moderated as they used to be, I don't think -- at least, back in 2008, they let go the woman who was approving commenters, and now it's more of a crowdsourced effort. But the site editors and some commenters are pretty active in setting the tone, and certainly, there are prevailing points of view, but also a lot of healthy back-and-forth.

    I don't know -- I think Gawker rightly takes a lot of flak for some of what it does, but as far as commenting goes, they make it work as well as I've seen on any blog with such high readership. The A.V. Club does a good job, too.

  • Oh -- meant to add: While it's totally important to reduce noise, I think there's a bit of an inherent, and now anachronistic bias in the notion that the blog producer needs to find a way to showcase good comments and filter out the bad. If one of the things new media have done is blur the lines between producer and consumer, then it's also up to the reader to learn to sift through content and ignore the gunk.

  • There’s really very little to be said for trying to carry on a conversation (online or otherwise) with people who have nothing but an instrumental view of conversation as a means to their own anti-pluralistic or illiberal ends, who concern-troll every debate in the hopes of getting someone to take the bait.

    This seems very close to me to saying, "You can't argue with most people," who succumb to the kinds of problems described here. I'm reading Bryan Caplan's The Myth of the Rational Voter right now, and it's a wonderful book on many levels—most notably for what he calls the voters' preference for irrationality, which is often a form of ignorance. As Caplan points out, the cost to the voter for irrationality is zero, since one vote effectively never alters an election (he develops the basic ideas much more, including fairly sophisticated descriptions of what rationality and irrationality look like).

    In any event, you can see where this is going: most bloggers have no cost to the vitriol; they have no cost if they're wrong; and they have no immediate cost to their ideological blinders. In this respect they're not like journalists, who can be fired if they're wrong or make gross misrepresentations (unless they work for Fox News), or academics, who will be professionally embarrassed if they have tenure or fired if they don't.

    This framework translates well to the Internet. It's also an issue that's been on my mind since I wrote "Trolls, comments, and Slashdot: Thoughts on the response to Avatar," which deals with the trolling tendency and, perhaps worse in some ways, the thoughtless tendency. This can be fought in part through deliberate culture, as Paul Graham does at Hacker News, but even then the fight is a) difficult and b) makes a site more like a garden that needs constant tending. This latter issue, by the way, is why I moderate both the blogs I contribute to, especially Grant Writing Confidential, which attracts a fair number of commenters who don't actually understand grant writing.

  • Very good insights, though no solutions. I've often wondered about scalability and why getting bigger is such an implicit drive when so many times small is optimal. With blogs having large readerships and comment sections, I neither wade through the comments nor offer my own. Similarly, if the gems among Amazon product reviews don't sort to the top, I'm disinterested in mining through tens or hundreds of submissions.

  • Josh wrote: "If one of the things new media have done is blur the lines between producer and consumer, then it's also up to the reader to learn to sift through content and ignore the gunk." I think this is true, but it's also true that many, many smart people who could make excellent contributions to online discourse are simply unwilling to do the sifting: too much trouble for too little reward. I'm not claiming to be a person of such quality, but I'm one of those who has opted out. I simply do not read blog comments anymore — ever — except for the ones on this site.

    By the way, following jseliger's link to his "Trolls" post will lead you to several interesting conversations on this subject.

  • Dr. Jacobs:

    I have been following this line of your thinking since you wrote a determined good-bye to blogging in Books and Culture back in 2006, and in his comment above, Brutus articulates part of my own reaction when he mentions that small is optimal. In 2006, when I read your good-bye to blogging, I laughed to my wife and claimed, "This is the complaint of one who is used to being published." At the time I had been blogging for not too long, and I loved that it gave me a venue for publishing my writing to a small audience and interacting with them. Online, as so often in person, small does appear to be a key for thoughtful interaction--thus the importance for me to have sought a college with small class sizes. When it grows big, it is difficult to moderate and filter for quality. In a blogging situation, it appeared that your "classes" kept swelling from small seminars to auditorium lectures, except the audience was more able and tempted to share their jaded cynicism than they would in a real auditorium.

    Anyway, on my little blog of small audience I have rarely encountered such problems, though even there Burke's observations about concern-trolling apply. I posted some light comments about Bill Bryson's book on Shakespeare and suddenly had an anti-Stratfordian reader terribly concerned about my teaching.

    Thanks for sharing your always developing thinking on this technology.

  • @Alan: Well, I think some kind of human hand -- to set the tone for the site, to be present in the comment section, to moderate and keep trolls out -- is necessary. Which is a good thing, right? I mean, if a major complaint about the Internet is its ability to dehumanize (a term I'm using really broadly here, for shorthand's sake), it's a relief to me to think that thoughtful discourse can't necessarily be automated.

    (Thanks for the note about jseliger's link -- I do venture into comment sections; I don't often click links in them. But I'll check it out.)

  • "Well, I think some kind of human hand -- to set the tone for the site, to be present in the comment section, to moderate and keep trolls out -- is necessary. Which is a good thing, right?" Um . . . I think so?? I do know what you mean, but that kind of role is far easier for an extravert to fill than an introvert such as myself. I'd rather just crawl back in my hole. . . .

    And Geoff, please note that the only "Goodbye" in that essay you cite is in the title, which I didn't choose. The essay itself is a critique, not a farewell.

    Also, all anti-Stratfordians need to read James Shaipro's new book Contested Will — it leaves their case in tatters, without even making that a primary goal.

  • I doubt that "small blog conversation" model is scalable, any more than the "small dinner conversation" is scalable. Indeed, I suspect that to believe that it is - and I don't know that slashdot really resembles a serious conversation, but I haven't really spent enough time there to say for sure - something akin to the belief in the possibility of "multitasking." To give serious, sustained attention to what your interlocutors are arguing is necessarily a slow, deliberative process and certainly not something that one can do with 15 people chiming in over 10 hours during a work-day, especially when at least a few of them will necessarily be up to mischief of one sort or another.

  • This post just came out and is relevant!

Post a Comment

[Basic HTML tags can be used in this comment field.]