Here’s A Solution for Transparent Hearsay

Media | Language/Structure
If only it were easy to find out what a given person thinks of a certain entity, and by extension, what a given person thinks of anything, or what people think of a certain entity. This, I believe is the goal of “information transparency” that is anticipated by the blogger visionaries.

Let’s concede that it’s very difficult to unearth this sort of information by reading through someone’s blog. You’d like to find out where they stand, and whether their position has changed over time. And, as difficult as that is, it’s logistically impossible to measure across many blog posts and many blogs. We need to develop a social technology specifically designed for this purpose.

The proposal is ironically named of course; the whole point is to eliminate hearsay. My first introduction to hearsay was at summer camp, when you’d mention a popular musician that you may like, you’d here “Oh he’s anti-Semitic! Says bad things about Jews and Israel at his concerts,” and I’d have no idea if this were true– seems that the easy way out was to stop listening to such a musican. Ten years later, I was still hearing these sorts of queries on Internet mailing list– somebody went to a hip-hop concert in New Jersey and heard it for themselves. So they immediately wanted to check with other other people. Though the act of checking involves repeating what could be hearsay to other people. So it would be desirable to have an database/index which could be easily searched.

The various websites on “urban legends” such as Snopes have some information to answer some questions about contemporary myths, many involving hearsay, but it’s hardly a generic database. I have suggested that social bookmarking systems such as bloglines could expand their blurb system to allow quantifiable rankings. But, again, these are limited in scope by rating only blogs and other online publications which support RSS. So let’s look at a generalized proposal.


The Hearsay Network consists of Questionnaire servers which publish questions, along with a framework of suggested answers. Now, anyone can submit a question, but, by some organic process some questions will rise to the top, either as a result of its ingenious wording, or the general reputation of its author.

Anybody the world can subscribe to these questions, and they can make their answers available to certain servers, to be called Answer/Summation servers. These servers should not only show the current answers for a person or entity, but they should provide historical information as well. So you can find out when a person changed their position on a given entity.


These are the types of questions that I envision would be asked:

  • Documents: what do you think of a given document resource? These would be identified by an Internet URL. I have developed ViewPoints as a rating system for document information.
  • Issues: These are based on documents, and may use URL’s. In addition to ViewPoints, the best questions should accept quantitative answers– not whether you support a given issue, but how much you are willing to pay for it.
  • People: What do you think of a Zephyr Teachout? Jeff Jarvis? Tiger Woods? Mother Teresa? me? or anybody? This would use an addressing system that is provided by the i-name framework, and implemented through organizations such the Identity Commons. This may also benefit by breaking down into various qualifications: as a worker, as a friend, as someone who can rate other people (today these attributes often rated by social networking sites)
  • Endorsements: Who are you endorsing for a President? for DNC Chair? for class clown? for best restaurant in town? The choices given would be i-names, as above.


For each answer, the user should be able to select an option or quantity and an explanation, which is a free-form text of limited size (25 words or less, but a URL may also be specified to point to a longer answer).

All of the popular blog/civ tools should support a way to do this. I envision that you could use them to load some fresh questions, using criteria of subject, popularity, or recommended by a question authority.

Now, to address what is now called the “access divide” or “leisure divide” (The “digital divide” is a misnomer as a majority of the population now has access to computers.) But people don’t necessarily have the time to participate on blogs or Internet forums. And they’re right– these take too much time investment to join and keep up with. The general population likes single-use tools: email, news on the web, IM, and maybe perhaps, their little polling pal.

This would have the added benefit of making polling fun. There’s nothing people love more to do than to answer poll questions– except in the case where they get a hundred tedious questions from Zogby Online or Harris Interactive. Opinion polling should be light and painless.

Also, your answer stays until you change it. This would encourage individuals to be familiar with their own answers– and not to have to answer the same question repeatedly.

Users may choose to make their answers anonymous for a given question. But for questions about people, the Answer/Summation server should not openly publish– anonymous explanations. There is no ethical benefit whatsoever to this. It should only be of interest for the subject of the question to retrieve anonymous advice directed at themselves.


For a politician, or a candidate, not to answer a question of serious public interest (as judged by the number of people answering it) would be unconscionable.

Should journalists be similarly expected to participate and answer questions? The consensus from some of the folk bloggers (Jay Rosen, Dave Winer) is that journalists are just like normal people and ought to make their opinions known. I argue otherwise. I conclude that it is are more interesting and useful to rate, using the collective hearsay network, the documents that a given journalist has published.

Can people asked to be removed from being ranked? That may be desirable, notwithstanding the enormous social value of getting unsolicited feedback. As for minors, I would expect most parents, and educators, and child psychologists (I am none of those) to give a flat-out NO.



We wouldn’t need polling firms. It’s quite ironic that this group hasn’t met the wrath of the online civic activists as much as, say, traditional media has. Newspapers at least have a letters page and an ombudsman. A polling firm, by contrast, can find out that you were against the Iraq war, and that your neighbor was for it initially, but has soured after 2 years– and the polling firm has no concept at all that you might have some common ground to talk on.

Yes, true, the polling community would insist that these polls aren’t scientific. They don’t have to be. They should be deliberative and conversational (this is not fully the same as James Fishkin’s Deliberative Polling, which demands that people meet to discuss what they are being asked before giving their final verdict). Scientific polling is, by it’s nature, anti-democratic. It doesn’t invite everybody to answer the questions. It doesn’t allow people to follow up constructively. It’s expensive enough to process simple yes/no questions and Likert scales. Instead of asking the vague question “How strongly do you feel about a gas tax?” our type of poll can ask quantitative questions like “How much would you be willing to pay for a gas tax?”

Product Surveys

The current approach to consumer product reviews is contained within privately controlled (though quite open) entities like Amazon and Epinions. I don’t participate in these, since it’s one more account I’d have to maintain, and I haven’t bothered to spend the time figuring out how these work. I don’t know how easy it is to change one’s review of a product– suppose it breaks. Or one’s review of a book. But I do suppose that it would be easier to do this if you can control your own ratings.

Institutional Reviews

As I wrote in Other Spheres of Media, there’s a number of sites on the Internet devoted to critiquing one particular institution — Dan Rather, Starbucks, Walmart, the MBTA. I called this the “contrasphere,” but that’s a misnomer, since its participants don’t really share information with eachother like the blogosphere does. Nonetheless, this group’s needs should be considered for the Hearsay Network. Each site has developed their own taxonomies specific to their institutional focus.


Hearsay was the first name that came to mind, though, regrettably, is owned by a domain squatter. So that remains to be chosen. We need more ways to get this into the language– we need an equivalent for “read the blogs” How about “read the pulse“? This would lend itself to an alliterative endorsement in and of itself by an editor or politician. “I don’t read the polls. I read the pulse.”

Maybe this will represent a giant revolution in social communications on the Internet. I heard Esther Dyson say at the last Berkman Center conference that the use of the Internet during the political campaign was all talk and no listening. Such a system would actually discourage the free-form talk one finds on the Internet. And maybe by reducing the talk, and encouraging people to qualify it, we’ll enable more people to listen better.

UPDATE January 23rd: I’ve finally updated some of the spelling typos and dangling fragments here, and made one semantic change: I had written that “polling is anti-Democratic,” but I did not mean to say “big D” but “small d” democratic. I regret the error. Well, I fixed these all not a moment too soon– some 149 visits to this page in the last 48 hours, in the last week. At the Berkman webcred conference, everybody seemed to ask for “new tools,” so I scribbled out a sketch of this on paper, handed to a few people in front of me at the big table, and Dan Gillmor liked it enough that he posted it to his blog.