Joi Ito's Web

Joi Ito's conversation with the living web.

I've been trying to push against Clay's assertion that blogs exhibit a power law and that power laws cause inequality. You can't "fix" the system without breaking it. We've gone back and forth in different places and I THINK I've boiled it down to a few key points for me.

When Clay uses the word "inequality" he means "not the same" and indeed, in a fair system, the outcomes will usually be inequal. I won't argue with that. What my question was was whether the rules were fair and whether we could counteract the current bias towards those in positions of privilege and amplify those opinions that are currently underrepresented.

I think the notion of trying to modify or influence the system to push it towards a particular outcome sounds like regulation and hits a negative chord with the free market libertarian types on the Net. I am also against unnecessary regulation. However, I do think that we can and should try to influence the architecture to push towards an outcome that we believe in. I think this is the nature of politics.

Clay talks about the power law in his paper, Power Laws, Weblogs, and Inequality. As most of you are aware, power laws are a type of distribution exhibited by large networks that grow where people are allowed to link freely. Since new sites tend to link to sites that already exist or are famous, the links aggregate to the well known sites making the "rich richer". The power law shows that even with fair rules, the outcome will be very inequal.

Recently, Marko and I blogged about blogs and justice in the context of the power law. Clay recently blogged this:

We can and should talk about the type of inequality we want — right now, for example, most of the high-flow webloggers are men. We can ask why that is, whether we should do anything about it, and if so, what? We can’t ask how we can level out the difference between the high-flow end of the popularity curve and the rest of us, or at least we can’t ask that unless we are advocating the destruction of the blogosphere. The interesting and hard question is “Since there is to be inequality, how shall it be arranged?”

I think we are going to see an explosion in work designed to alter the construction and effects of this inevitable inequality (viz Sifry’s experiments on moving recent blogs up the Technorati list) and I am optimistic about this change, as I believe the concentration of real thought and energy on what is actually possible, as opposed to cycles wasted on utopian declations, will be tremendously productive.

So I'm glad Clay is willing to consider what we might do about the fact that the most influential blogs are by people in positions of privilege.

In Linked Albert-Laszlo talks a lot about power laws and makes a few interesting points. First of all, power laws on the web make two assumptions, that the network is growing and that people tend to link to sites that have the most links. Laszlo cites work by Paul Krapivsky and Sid Redner from Boston University, working with Francois Leyvraz from Mexico,

generalized preferential attachment to account for the possibility that linking to a node would not be simply proportional to the number of links the node has but would follow some more complicated function. They found that such efforts can destroy the power law characterizing the network.
He goes on to talk about Google coming in as a latecomer in the search game and how "fitness" or the likelihood that someone will link to you is not entirely determined by your existing position on the power law curve and that a site worthy of connecting to can quickly scale the power law curve if it exhibits exceptional fitness. All disruptive technologies and innovations break power law curves by exhibiting exceptional fitness.

If you think about the power law as themes or ideas instead of people and you think about fitness as the level in which an idea resonates with people, the power law could be viewed as an amplifier for ideas and memes that are sufficiently interesting. Because fitness so influences a nodes ability to climb the power law, I think the notion that I described in the Emergent Democracy paper, where the tail of the curve is where the creativity happens and the power law is how an idea whose time has come goes main stream still makes sense. I think the key to making the system "fair" is to make sure the tail is as inclusive as possible and to try to encourage technology and norms to value fitness over simply linking to those who are popular. As Ross shows in his three layers of creative, social and political, I think the power law is the final amplification part. In fact, the tail of the power law, the creative layer and the social layer where the initial deliberation occurs might be where we should be focusing our energies.

I have a feeling that the blog power law is like a real-time amplifier. I think it is key to note that nodes that lose the fitness that got them there in the first place retire very quickly and that fitness is amplified in scale-free networks. If we architect blogs to allow the amplifier to be sensitive to positive fitness and quickly retire irrelevant blogs, it will be a good amplifier. If the Technorati top 100 is the Marshall amp, maybe we should be talking about the guitar?


I finally figured out what I don't like about Clay's pitch.

It's not fair that Clay, who hasn't made the investment in blogging himself, should try to control what happens in blogs.

When he talks about justice, I see the hypocrisy. If you said the pavement should be ripped up and you were willing to wear a ten ton weight around your neck, that would carry some weight, because at least you would have something to lose.

In any case, hypocrisy aside, blogs are not like television networks. A blog with an audience of 15 makes total sense in lots of cases. You don't measure the value of a blog by how many people tune in. In fact, I believe the value of a blog goes DOWN as more people tune in. That's been my experience. You waste more time trying not to evoke readers' emotions. And that is a waste of time, and it turns you into an actor or a Washington pol, anything but a blogger.

So it's self-defeating, we don't need Clay's help, which reeks of carpetbagger-ism.

And Joi, in case you noticed, I'm not worrying as much about what people think these days. I'm saying what *I* think, and I encourage others to do the same. I'm sick of blogs that are all about being and staying friends with people. That, emphatically, is not what blogs are about.

BTW, your blog is *really* slow at posting. It never seems to come back after I press the Post button.

Nice, I like the amplifier metaphor. Little worried about what seems to be some premature jumping ahead going on though. A lot of building off of unproven assumptions appears to be going down.

Ross doesn't "show" anything with his three layers, he presents a hypothesis. Same thing with the notion that creativity lies in the "tail" of the curve. These are interesting concepts that should be studied, but I sure wouldn't trust them as foundations for anything yet.

Perhaps a bit more rigor should be placed towards testing some of these assumptions? Experiments perhaps. What happens when the exact same post gets published by a "tail blogger" and an A list blogger? Who gets credit with the idea? and is it fair?

Maybe seed 10 posts to tail bloggers and have them post. Wait a week or two. Observe. Then post on 10 different A list blogs. What happens? We all probably have our suspicions, but who wants to test them rather then architecting with early assumptions?

Couple of points, firstly what you're describing sounds similar to the topic-based ontologies that Topic Exchange and k-collector are currently trying to build out to the mainstream. Content is aggregated based on topic, not on *who* posted it. Dave Winer's category/Channel Z tool sounds like it could do the same.

My second point is that Sir Tim Berners-Lee has a theory he calls the "Fractal Society", which in some ways I think could be an antidote to all this handwringing about the power law (note I didn't say antidote to the power law itself). Fractal in Sir Tim's view means "structure on many levels". Here's part of what I wrote up the other day based on Sir Tim's interview with Christopher Lydon:

[start of excerpt]TBL wants people to try and achieve a "balance across the different scales". But achieving success on a global scale doesn't mean you need to become famous - it simply means "think global, act local". His suggestion is that we should divide our time over 10 channels, like so:

1 - You
10 - Your family
100 - Your social group
1000 - local community (eg your church)
etc up to the 10th level, which is a global scale.

Each of those levels represents the size of your audience, or the number of people you are dealing with (nb: TBL didn't specify what exactly the labels represented, but this is my understanding of what he said). He had an analogy of dropping marbles into 10 cans and the aim is to spread one's marbles around. TBL went on to discuss the fractal theory in terms of blogging:-

Blogging is an individualistic activity, in that you're expressing yourself via the Web in your writing and other multimedia. But blogging is also a fractal activity because, even though you're doing an individualistic thing, you're also "part of something bigger". When you blog, you're participating in a group activity. The question for us bloggers then becomes: which scale am I blogging at and therefore how much time should I be devoting to it in relation to my other activities? For example I am currently working at the 3rd level with my weblog - my audience can be counted in multiples of 10. So should my ultimate goal be to hop up a couple of levels and become an A-List blogger, where my blog reaches an audience of 10,000 or more? Or do I want to become a Citizen Blogger for my local community, so I move up to level 4 with an audience of around 1000? Or should I be happy writing for a small audience of people who share my interests? Maybe I can do a combination of these things - that is, different blogging activities aimed at different fractal levels. These are all questions that perhaps bloggers should be asking themselves. My initial impression is that Tim Berners-Lee's fractal theory helps us to balance blogging with other aspects of our social lives on the Web. Perhaps it's even an antidote to inequality in the blogosphere? [end of excerpt]

I don't know if I'm off-base here and sorry for excerpting such a large bit of my content ;-) But I think this "fractal society" concept has some potential in the blogging world...

Best regards, Richard

Thanks for your comments Dave. I generally try to write what I think and I think that trying to address Clay's points logically is something I'd like to try to figure out since I do think he has a strong argument.

I'm not sure exactly what blog are all about yet, but my friends do hang out here... Hanging out with my friends is not the primary purpose of the my blog, but keeping in touch with my friends is definitely part of why I blog.

Abe. Good point. I think experimentation and rigor is a good thing and needed to test some of these ideas. I agree, some of my assertions are more like theories or "hopes".

Interesting Richard. Thanks for the excerpt from the Chris Lydon TBL interview. I like that image a lot.

And sorry about the slowness Dave. I think I have a way to accelerate the rebuild. I'll try to implement it when I get back to Tokyo.

Joi, I think your point that "we can and should try to influence the architecture to push towards an outcome that we believe in" is exactly the issue. I've tried to summarize what I call the Three Mistakes in the Moral Mathematics of Blogging in reply to Clay's points. This is just a first step to clear the ground for a positive view.

Interestingly, the topic of justice is important and not only from a normative – what some would call a soft point of view. It is ultimately related to the issue of sustainable business models in this space. The question of justice relates to the business of social software, because social software is, well, social. And social systems need a degree of trust to be sustainable. A big part of that trust comes from a public conception of justice that all participants accept.

But I'm sure many will disagree with this last point. Time will tell.

For the emperically minded... Seth F. has an analysis of readership after a reference on Lessig Blog:

It would seem from his results that you'd need a number of such A-list refferals before you started to get into the power-law territory.

Following on Richard's comments, I think we need to increase our sense of subtlety and scale when we talk about complex systems in general, but with specific applications in mind.

A while back, I wrote about how web game hi-score lists might exhibit power law growth. My point was not that the ensuing rich get richer scenario is necessarily undesirable, but that it might not encourage the participation the web game publisher (in this case British Airways) wanted to achieve. In other words, it's a question of what we want to get out of these systems. Between the lines of that post, I suggested that a smaller range of participants might have made the hi-score list more effective in its intended purpose (namely, encouraging competition).

Joi, I think the idea of the amplifier is interesting, but my question is what does power law growth amplify, and why, and to what benefit or detriment is specific circumstances. Specific circumstances are important.

Inbound links are important for Yahoo! because they make money from advertising and, increasingly, customer conversion. Inbound likes are important for blogs because... well, for a number of reasons. For me (a video game researcher), I want to participate in an exchange of ideas specific to my field, that together might advance it. At the same time, I admittedly also want to stake out own influence within my desired sphere of influence. For others, their goals are different.

I don't like things like Technorati because they preclude the question of specific value in individual blogs and localized networks of blogs. What are blogs good for? If the blogosphere is just about posturing for a place in the blogosphere, then no thanks, I already went to high school.

I agree with Dave, you don't measure the value of a blog by how many people tune in. In fact, this is exactly the problem with the social network services like Friendster: they're still focusing on the size of a network instead of what you do with it, and how. LinkedIn gets a little leg up by focusing the kinds of interactions it enables, but the whole system is still stratified based on numbers of links, rather than the value of the work produced as a result of individual connections.

Blogs are a medium for exchange on a specific topic. If anything, they provide an exceedingly accessible way to level the playing field within a specific community, if your content is fit, as you say Joi, then it has a good chance to excel. We need to focus more on specificity and relative value in individual blogging communities -- on individual networks of blogs and their users, and what those networks do that is interesting -- and less on broadly formalist notions of the blogosphere itself.

I think the power law curve is an amplifier. Technorati is just a way of measuring and visualizing it. I think that people tend to learn about new blogs through other blogs and blogs that have a lot of inbound links or are popular have a louder voice. However, I've watched the number of inbound links fluctuate on my blog and they fluctuate heavily on a daily basis depending on whether I'm saying something interesting.

I guess my point is that a snapshot of the power law curve makes it look a lot like power law curves that are more static, but with blogs, I believe that you can quickly gain or lose your public voice depending on the relevance of what you're writing. Some ideas will propogate up the power law via other blogs and some blogs will propogate directly like Salam Pax.

Again, I don't think that the amplifier is necessarily the most important, but if you're trying to cause political action or widespread dissemination of some idea, at some point you're going to want to amplify it.

I think that diving into deep thoughts or discussing something creatively would be easier to do "under the radar" without the noice or the scrutiny. Once an idea's time has come, the amplifier can pick it up and send it out. That's what I'm trying to argue.

So although the value of your blog is not measured by how many people turn in, the president of the US and public policy will be decided by how many people are swayed and how many people vote. That's why I like Ross's characterization of that layer as the political layer. You can be just as happy playing your guitar without an amp, but you'll need one at the primaries.

Six Apart has just released a new version of Movable Type (2.661) that changes the behavior of to use redirects when linking to URLs given in comments. The goal of this is to defeat the PageRank boost given to spammers by posting in the comments on a weblog

This will have the interesting side-effect that 'normal people' (ie non-spammers ;-) who often comment on other people's weblogs will no longer get extra GoogleJuice (from the comment author name being linked to their blog) than those who don't comment. It probably won't affect Technocrati rankings (I assume they'll code around it), but might have some interesting effects anyhow.

Cool, a mention - somebody somewhere was reading :-)

Now, regarding:

"What happens when the exact same post gets published by a "tail blogger"
and an A-list blogger? Who gets credit with the idea? and is it fair?

This is easy - to a very good approximation, if it's published by a
tail blogger, it's ignored. If it's published by an A-list blogger,
it gets echoed-chambered around as much as any item. I mean, why is
this even an issue? Look at it this way:

"What happens when the exact same post gets published to 100 people
and to 10,000 people? Who gets credit with the idea? and is it fair?"

Whether you think it's fair or not is an exercise in moral
philosophy. What happens is an exercise in mathematics.

In general, much of the discussion seems to revolve around a few ideas:

1) Some People Write Just For Themselves

Indeed they do. But they don't care about power laws, so anyone in the
discussion is automatically excluded from here.

2) Fame Is Fickle

Well, yes, at the margins, someone can hit it big, or fade from view,
it happens. And celebrity requires some care and maintenance. This is
pretty mundane stuff. It doesn't automatically somehow mean that
fame/celebrity is therefore "fair".

2) You Can Be Big Fish In Small Pond

Sometimes people want the respect of a few peers, a professional circle,
more than a global audience. For example, they'd rather impress the
New York Times Review Of Books, than Oprah's Book Club. This tends to
be argued as if it were a rebuttal to the limits of having a global
audience. It isn't. It's both a different goal, and within its own
terms, it has the exact same problem of only a few slots (we can't all
have a million readers, and we can't all win the Nobel Prize For
Literature either).

3) Don't Worry, Be Happy

Cultivate your garden. Merrily pour time, energy, effort into your
work, and if it reaches just one reader, consider it a job well done.
It is very easy for a "have" to say this to a "have-not".

I could say more, but this is long enough (and I worry about being
a blog-ant among blog-elephants).

Respect for fittness is the key to maintaining a healthy society in a networked age. Respect for fitness is akin to Social Justice.

Also, while power laws do come into play, the blogosphere is very tolerant of human scale projects. I have a readership of about 450, and that's pretty comfortable for me. Small things have value too, especially when there are a great many of them, and when some may contain the seeds of greatness.

It's a utopian ideal, but it works pretty well. You do what you do because you love doing it. Because you love doing it you do it well. When you do things well, it's natural to share them with other people.

That's what this is about. Popularity only matters if you're trying to sell things.

I'm sure you're tired of hearing me belabor this point, but this discussion of fitness and "amplification" reminds me of Adorno and Horkheimer:

Interested parties explain the culture industry in technological terms. It is alleged that because millions participate in it, certain reproduction processes are necessary that inevitably require identical needs in innumerable places to be satisfied with identical goods. ... any trace of spontaneity from the public in official broadcasting is controlled and absorbed by talent scouts, studio competitions and official programs of every kind selected by professionals. Talented performers belong to the industry long before it displays them; otherwise they would not be so eager to fit in.

Doesn't that describe the selection of "ideas and memes that are sufficiently interesting" by market-like processes? Doesn't the expectation that gives rise to the question of "inequality" (like the problem of envy in Rawls) arise from the expectation that, as bloggers, our aim to build a relationship between producers and consumers, no different in kind from that between a corporate content-producer and its consumers? "Popularity is only important if you're trying to sell something."

Furthermore, what is interesting is what is interesting to most people. But most Americans believe that Iraq was directly involved in 9-11. Since this is not the case, it seems that most Americans are interested in the wrong things — the amplification of this meme seems pretty pernicious me. Is it utopian to expect better of our fellow citizens?

Robert McChesney:

Given the dominant patterns of global capitalism, it is far more likely that the Internet and the new technologies will adapt themselves to the existing political culture rather than create a new one. Thus, it seems a great stretch to think the Internet will politicize people; it may just as well keep them depoliticized. The New York Times cites Wired magazine approvingly for helping turn "mild-mannered computer nerds into a super-desirable consumer market," not into political activists (Keegan, 1995, pp. 38-39). In particular, having mass, interactive bulletin boards is a truly magnificent advance, but what if nobody knows what they are talking about? This problem could be partially addressed if scholars and academics shared their work with and tailored their analyses to the general public, thereby engaging in a public dialogue. Unfortunately, such behavior runs directly counter to the priorities, attitudes, and trajectory of academic life.

Well, this discussion, at least, provides a counterexample to that last bit of pessimism. Still, the general tendency of this debate is to accept the (terribly flawed and awfully literal-minded) metaphor of a "marketplace of ideas" with a complete lack of critical reflection. Tell you what: you read CyberMarx and I'll bone up on my network economics ...

9 TrackBacks

Listed below are links to blogs that reference this entry: Inequality and the role of "fitness" in power laws.

TrackBack URL for this entry:

In this article I draft guidelines for a Fractal Blogosphere and suggest that Read More

Joi Ito has an interesting post where he raises the concept of blogs acting of amplifiers. It's a good metaphor,... Read More

Joi Ito has an interesting post where he raises the concept of blogs acting of amplifiers. It's a good metaphor,... Read More

Joi Ito has an interesting post where he raises the concept of blogs acting of amplifiers. It's a good metaphor,... Read More

Powerlaw Puzzle from Ascription is an anathema to any enthusiasm
May 27, 2004 4:19 AM

Joi Ito takes a stab at getting his head around what he thinks about the power-law puzzle. Dave Winer comments that he can't think about the issue because he's decided he doesn't like the messenger's credentials; the messenger in this case being Clay. ... Read More

Policy Entrepreneur from Ascription is an anathema to any enthusiasm
May 28, 2004 10:32 PM

I've been thinking about roles in the economy of ideas. Three roles: those that make ideas, those that consume them, but most importantly the idea middlemen. I'm reasonably sure it was one of Paul Krugman early books that introduced me to the idea of ... Read More

Policy Entrepreneur from Ascription is an anathema to any enthusiasm
May 28, 2004 10:46 PM

I've been thinking about roles in the economy of ideas. Three roles: those that make ideas, those that consume them, but most importantly the idea middlemen. I'm reasonably sure it was one of Paul Krugman early books that introduced me to the idea of ... Read More

There's been some follow-up discussion on folksonomies (by Peter, Clay, Victor, and Jon) and what I believe is a related thread about Wikipedia vs. traditional Encyclopedias (by Joi, Matt and Cory) Both discussions seem to boil down to a few... Read More

Tags from Preoccupations
September 10, 2004 5:13 AM

Taxonomies have been much in the news recently. Laughingmeme, for example: complex hierarchies of information suffer from many of the same problems that hierarchies in general suffer from, in particular they're brittle. They adapt poorly as your domain... Read More