Clay Shirky
( Archive | Home )

Liz Lawley
( Archive | Home )

Ross Mayfield
( Archive | Home )

Sébastien Paquet
( Archive | Home )

David Weinberger
( Archive | Home )

danah boyd
( Archive | Home )

Guest Authors
Recent Comments

pet rescue saga cheats level 42 on My book. Let me show you it.

Affenspiele on My book. Let me show you it.

Affenspiele on My book. Let me Amazon show you it.

Donte on My book. Let me show you it.

telecharger subway surfers on My book. Let me show you it.

Ask Fm Anonymous Finder on My book. Let me show you it.

Site Search
Monthly Archives
RSS 1.0
RSS 2.0
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline


« Who's afraid of Wikipedia? | Main | Popularity Slider: Diving into the long tail »

March 1, 2005

Matt Locke on folksonomies

Email This Entry

Posted by Clay Shirky

Wonderful Matt Locke piece on folksonomies, which introduces not one but two substantial ideas to the debate:

Perhaps this illustrates the limit of folksonomies - they are only useful in a context in which nothing is at stake. [Emphasis his] Folksonomies are, in essence, just vernacular vocabularies; the ad-hoc languages of intimate networks. They have existed as long as language itself, but have been limited to the intimate networks that created them. At the point in which something is at stake, either within that network or due to its engagement with other networks (legal, financial, political, etc) vernacular communication will harden into formal taxonomy, and in this process some of its slipperiness and playfulness will be lost.

He relates this to the idea of play from finite and infinite games. (I’m more optimistic about the shift here than he is, for reasons I’ll discuss below, but I think he’s spot on about the gap between palyful and serious categorization.)

The other idea, from Bowker and Star’s marvelous Sorting Things Out, is about the inherent tension in classification generally:

Bowker and Star identify three values that are in competition within classfication structures: comparability, visibility and control. Folksonomies have elevated visibility, but at the expense of comparability (being able to translate classifications across taxonomies or contexts) and control (the ability of the classification to limit interpretation, rather than interpret ‘emergent’ behaviour). Whilst nothing is at stake, and there is little lost by not being able to transfer taxonomies from one context to the other, or users are not disadvantaged by the need to independently assess and contextualise meaning, folksonomies will provide a useful service.

Just a fantastic post.

The only place I vary from Matt (it’s not even a disagreement, really, just a prediction about the future) is in the eventual value of folksonomy. He likens folksonomies to vernacular vocabularies, but this doesn’t describe their first-order importance, at least not where systems like are concerned.

Here’s what’s radical about what protends: My vocabulary on folksonomy is personal, not vernacular — no one knows or needs to know which class I’m talking about when I tag something ‘class’, or that I use LOC to mean Library of Congress. This isn’t the same as, say, the dictionary of thieves slang from the mid-18th c. because no one else needs to know my bookmark system, and I don’t need to know anyone else’s, or, to quote Adam Smith: “It is not from the benevolence of the butcher, the brewer, or the baker, that we can expect our dinner, but from their regard to their own interest.”

This is really, truly different, because it uses the intiution of markets — aggregate self-interest creates shared value. Locke points to the loss of control as one of the downsides of folksonomic classification (at least in its del-style form), but there are significant upsides as well. The LOC has no top-level category for queer issues, but does, because its users want it to.

By forcing a less onerous choice between personal and shared vocabularies, shows us a way to get categorization that is low-cost enough to be able to operate at internet scale, while ensuring that the emergent consensus view does not have to be pushed onto any given participant.

Which is why it mystifies me that both Matt and danah are so concerned with exclusion — who’s excluded here, who isn’t also excluded from using the internet generally? Put another way, is anyone excluded from using who has better representation in other classification schemes?

The answer is “If you don’t like the way something is tagged, tag it yourself. No one can tell you not to.” Prior ot, controlled vocabularies were almost inevitably vocabularies that pushed the politics of the creators onto the users; that is upended here.

Comments (9) + TrackBacks (0) | Category: social software


1. David Sabel on March 1, 2005 11:48 AM writes...

I'd add that the more different "personal" ways people choose to tag the same thing actually increases meaning.

For example, two users of our site,, have classified the artist Robbie Fulks - one chose to label him as "Roots" another as "Americana."

Seeing both labels together adds nuance to understanding who people think Robbie Fulks is. And, when you can drill into each user to see what else they have classified using their tag (Americana also has Gram Parsons, Emmylou Harris, Lucinda Williams while Roots has Ry Cooder, Sister Rosetta Tharpe etc.) you get even more user supplied context and meaning.

Permalink to Comment

2. Bud Gibson on March 1, 2005 3:54 PM writes...

Interesting post. For me, it comes back to the idea that folksonomy at any aggregate scale is about data collection. How a social system aggregates that data will determine whether or not there is exclusion.

Permalink to Comment

3. Julian Bond on March 1, 2005 4:31 PM writes...

We need a taxonomy of folksonomies. No really! ;-) The problem is that is not the same as flickr is not the same as technorati tags. We've got group tagging of individual entities, individual tagging of multiple entities and individual tagging of individual entities. And of course shades of gray between these.

Some situations like tagging wiki pages lend themselves to wiki-like convergence and agreement among taggers on terminology. has enough feedback, volume and common language to encourage convergence purely on statistical terms.

But now what about tagging in an environment where something is at stake, like money. If we just took eBay, tossed out the categories and replaced them with a folksonomy tagged by auction owners, we'd have tag spam in 3 seconds flat. This may be why in the marketplace of web page attention and search engine placement, meta keyword tags ended up being useless when a significant number of people gamed the system. There was no positive feedback and no incentive to converge on "better" tags.

I'm really hoping that we can find ways of using mesh like tags instead of tree like category hierarchies to navigate things like Craigslist, eBay, DMOZ. But while it's tempting to get the members to do all the work, maybe there's no way of encouraging convergence instead of divergence and so preventing it ending in chaos.

Permalink to Comment

4. weaverluke on March 1, 2005 6:08 PM writes...

Julian feels we need a taxonomy of folksonomies. I would suggest we need to find approaches to data management that organically integrate structured and fluid modalities. Human community can be seen to evolve through a phase-patterned play of structure and fluidity, but the social net does not as yet embody the possibility for such dialectic processes:

Permalink to Comment

5. Ian Smith on March 1, 2005 9:08 PM writes...

Wouldn't it make more sense to grow folksonomic interfaces than build them with taxonomies? Think of various tag-centric sites as fitness landscapes, sooner or later some of the objects will spill out of one system to another and be tagged on different orders - item, group, metagroup. There seems to be the opportunity to jig up some sort of meta-tag viewer by bouncing a URI off of any given number of tag-centric site APIs and graphing out the results to keep them in context with one another without placing any one over the others. Finally, it also seems that as long as there is a means to bridge between the so called stake-less and stake-full systems when necessary there doesn't seem to be any reason to try and prematurely impose a taxonomy onto the stake-less system, particularly since one will emerge on its own if we resist the temptation to play human.

Permalink to Comment

6. Julian Bond on March 2, 2005 2:45 AM writes...

Weaverluke, I have no idea what you're talking about ;-)

Permalink to Comment

7. Saurier Duval on March 2, 2005 9:42 AM writes...

Great post, just a remark: the playful (personal, vernacular, folksonomic) system seems to me to be a superset of any (or the combination of all) 'serious' categorization schemas. Usecase if anyone wants a LOC conform classification there, just create a canonical user (or use an intra/inter-group tag as flag signifying the mode of formality) who's tagging behaviour follows those rules and naming conventions. You still get all the rest for free.

Permalink to Comment

8. Mark Wubben on March 2, 2005 10:53 AM writes...

Ian, that's what I've been thinking about too... now if somebody had some time left for me I could write about it.

Permalink to Comment

9. John Mayer on March 10, 2005 3:34 AM writes...

A vernacular is exactly what students are learning when they study a new subject. I work in legal education one of the biggest barriers is learning the legal jargon. The education link here is that students build their own internal understandings (constructivism et al) and folksomonies seem to be outward representations of these internal structures. Does this go both ways? Can an instructor observe a student's folksomony to see if they "understand" the material being taught? Or, is it too too personal.

There does seem to be a continuum from internal structure to folksomony to vernacular to formal taxonomy and the point of education is to bridge the formal to the internal ... or to create a replica of the formal in the internal (i.e. indoctrination in both positive perjorative sense). As any politician knows, if you can frame the argument with your own button-words, you can win it on your own ground. Look at how politics handles the environment, left v. right, abortion, racism, war, ad nauseum. Aren't "talking points" just an advocacy-intended folksomonie that wants to grow up and become a vernacular that is eventuall passed as law into a hard taxonomy? Isn't law the hardest of taxonomies? Whoaaa. my brain just exploed.

Permalink to Comment


TrackBack URL:

Listed below are links to weblogs that reference Matt Locke on folksonomies:


Email this entry to:

Your email address:

Message (optional):

Spolsky on Blog Comments: Scale matters
"The internet's output is data, but its product is freedom"
Andrew Keen: Rescuing 'Luddite' from the Luddites
knowledge access as a public good
viewing American class divisions through Facebook and MySpace
Gorman, redux: The Siren Song of the Internet
Mis-understanding Fred Wilson's 'Age and Entrepreneurship' argument
The Future Belongs to Those Who Take The Present For Granted: A return to Fred Wilson's "age question"