Why a bat may be a bird: the semantics of the ‘tag’

Print Friendly, PDF & Email
[I’m still in the middle of writing this one … but feel free to read the draft below. A final revised version will find its way into my ICHIM 2007 conference paper.]

There’s been some lively discussion in the steve.museum mailing list over the past week on the subject of ‘correct’ and ‘incorrect’ tagging of (museum) artefacts. It began with the following posting from Belinda Chu:

We are looking into a scoping exercise with regard to implementing social tagging on our online collections and Picture Library. The issue of users incorrectly tagging an object was brought up by our Collections Information Manager (eg: tagging an item as a landscape when in fact it isn’t). I would be interested in hearing from any institutions out there that are currently utilising social tagging that may have come across this issue and how it was addressed. For instance do you moderate the tags before they go live on your site.
Also, is anyone out there currently allowing user generated tagging to be applied in a bilingual environment on indigenous artefacts (For example in our case we have Maori and Pacific Island artefacts in which members of those various communities may want to tag in their own languages…?)
I look forward to hearing back from you.

Rob Stein from the Indianapolis Museum of Art responded that “The approach we’re taking with the steve project is one where we are building tools that will allow individual institutions to ‘review’ terms submitted by users against their contributed works of art”, which elicited from Chad Petrovay the response: “Isn’t the purpose of the entire steve project to democratize object tagging? If institutions start to make decisions about what is ‘correct’ vs ‘incorrect’ doesn’t that negate the entire purpose?”

Catherine Styles, of the National Archives of Australia, followed up with a link to a ‘useful rave’ by David Weinberger about the relationship between tax- and folks-onomies, ‘Can tags be wrong?‘ His article is thought-provoking enough for it to be worth looking at in some detail; and, although I shall quibble with some of his points along the way, I shall find myself–with some reservations–more or less in agreement with his conclusions. Weinberger writes:

Personal tags are not about truth. If I tag a photo of the Bay Bridge in SF as “golden gate” because I think that’s what it is, I’m wrong. But, when I go to look for that photo, my tag is still useful to me so long as I’m still wrong about which bridge is which. Of course, I might have tagged it “golden gate” because I’m doing a report on bridges near the Golden Gate Bridge, in which case my tag was true, even though a stranger who is not privy to my mental innards would assume I’m mistaken. But many personal tags aren’t primarily about truth at all: If I tag the photo “homesick”, “examine closely later” or “write poem about”, the value of the tag isn’t in its representation of something true about the object.

He’s stating, of course, what we all intuitively know and what we all so take for granted that we don’t much reflect on it: personal tags, like the piles of papers on our desks or perhaps the organisaton of our bookmarks in Firefox, are about making things easy to find, not about what they ‘mean‘. The matter of ‘correct’ and ‘incorrect’ (or ‘accurate’ and ‘inaccurate’) tagging only really become an issue when we move from the personal to the public and social:

Now we’re in the realm of folksonomies, i.e., the topology of tags generated by lots of strangers. At its simplest, a folksonomy reports on which tags are most popular for particular objects … Folksonomies get their value by reflecting the viewpoint of the plurality, not what an authority thinks is or ought to be the viewpoint.

So again it’s not about meanings but about ‘viewpoint’: where it’s located in those great big majority-rule public piles of papers, as it were; a folk classification, it is consequently different in kind—rather than simply an alternative way of organising a domain—from formal (scientific) classification. Arguably, it’s still, as with the personal tag, less a statement about what kind of thing an entity is than about where the majority of taggers think it is, what metaphoric ‘pile’ it may be found in and retrieved from, functional rather than semantic. When in a visit, a while back, to the online Philadelphia Museum of Art, for example, I tagged Paul Cézanne’s untagged The Large Bathers as ‘Impressionist’, I did so not because it was part of the meaning of the painting but because I just happened to know that this was at least one of the ‘piles’ in which it belonged and that my tagging it as such would help others find it.

Yet this is where I suspect we’re stepping into what is still uncharted theoretical territory. For the personal tagger, is the choice of tag routinely determined, after all, by (in the broadest sense) semantic knowledge? 1 I was possibly thinking taxonomically as much as folksonomically when adding the ‘Impressionist’ tag to Cézanne; and yet, although I don’t believe that either one excludes the other, my concern had been simply to put the painting where I believed it should be located for purposes of access rather than what kind of painting it is. The functional folksonomic tagging nonetheless was undoubtedly determined by taxonomic thinking against a background of what I’d assume to be common knowledge. And then going beyond personal tagging, is there some point at which the “viewpoint of the plurality” crosses over from the functional to the semantic? I don’t know; but I’d in any case want to say that in the former case, in which the tag is used effectively as just a mnemonic, it makes less sense to ask whether it is ‘right’ or ‘wrong’ than whether it is ‘useful’. I want to return to this issue later; but to, in the meantime, bear it in mind when reflecting on Weinberger’s question, “Can a folksonomy be flat-out factually wrong?”, for which we therefore have to assume a semantic interpretation. He continues:

Suppose the main tag for “The Protocols of the Elders of Zion,” an anti-Jew forgery, was “history”? While we cannot know what’s going on in the heads of all the people who tagged it that way, it seems as straightforward a misclassification as tagging bats as “bird.”

This is the point at which I feel bound to ask what may at first glance seem a misguided, if not uncomfortably odd, question: if the tagging represents the “viewpoint of the plurality”—presumably something like Schutz’s notion of “what everybody knows”—then isn’t it as arrogant as it is wrong to call it a “misclassification”? If pretty much everybody tags bats as ‘birds’, then they’re birds; it merely turns out to be the case that what popular consensus deems in this instance to be the meaning of ‘bird’ is probably somewhat different from what the word means for a zoologist.

But this is me now wearing my anthropologist’s hat. If, for the Karam of New Guinea, a bat is a yakt (‘bird’) and a cassowary is a kobtiy (roughly, ‘flightless biped’), I will not say that the Karam are guilty of ‘misclassification’. Rather, it happens to be the case that they classify things somewhat differently from the way that we do. And this, it seems to me, is both an interesting and an important point, particularly in light of Belinda Chu’s original posting: “we have Maori and Pacific Island artefacts in which members of those various communities may want to tag in their own languages”. It occurs to me that tagging in one’s own (non-European) language may inevitably and ineluctably entail tagging that also captures the indigenous classification systems for which that language is a vehicle. Were the museum devoted to the Karam rather than to the Maori and Pacific Islanders, for example, I would expect a bat to be tagged yakt and not kobtiy, a cassowary to be tagged kobtiy and not yakt. (They might also be tagged in other ways by non-Karam speakers; but that’s an issue I’ll come back to.) And I would also hope that the tagging might reflect the Karam taxonomy: 2

no hair
head not bony
can fly
small & medium size
small legs
no brain
head bony
large size
strong heavy legs

The very act of tagging then becomes an integral early step in the process of discovering and organising the knowledge systems that in an important way constitute (in the words of ICOM’s statutes) “the tangible and intangible evidence of people and their environment”.

But for Weinberger it’s not simply a matter of “misclassification” alone. He goes on to write that:

If we took a poll about either the “Protocols” or bats and the results showed that the majority of those polled believed “Protocols” is true and bats are birds, we wouldn’t say the poll was wrong. We’d say the people were wrong. Likewise, the folksonomy would be a true reflection of the popularity of false beliefs.

I’ll come back to thorny issue of The Protocols of Zion later. I want first to provocatively think a little more about the claim that “the folksonomy would be a true reflection of the popularity of false beliefs” (my emphases). Folksonomies, he goes on to add, “reinforce belief systems, since we believe (rightly) that what most people believe is a (generally) reliable guide to truth. Folksonomies make visible, and thus magnify, the effect of belief systems.”

I want at this point to return to the questions I raised earlier: is folksonomic tagging, no matter how innocently functional it may be in the mind of the tagger, guided by taxonomic or, more broadly and generally, semantic thinking? is there some point at which the “viewpoint of the plurality” crosses over from the functional to the semantic? can coherent classification schemes emerge from undirected personal tagging by individuals? Weinberger apparently assumes so; yet I suggest we need to look at the issue in a little more detail. My starting point will be another question: is the “viewpoint of the plurality”, as represented by the highest ranking tags, best understood to be [a] the collective and consensual viewpoint of the majority of taggers? or [b] simply the sum total of all personal tags? Or, asking the question in a slightly different way, how reasonable is it to argue the bona fide existence of a consensual ‘viewpoint’ in sense [a] as (presumably) an emergent property of [b]?

There can be no straightforward answer, of course. There will always be likely to come a point at which the functional tagging of active taggers becomes semantic labelling for others, a point at which concurrently the functional for some is a keyword index to a belief system for others. It is nevertheless an important question when considering Weinberger’s claims with regard to ‘correctness’ of tags in general, to the ‘truth’ of tags, and in particular to possibly ‘false beliefs’.

One of the key features of a shared ‘belief system’ is that it is more than simply the cumulative and quantitative sum of the personal beliefs of individuals. A belief system is ‘shared’ not because every member of an epistemic community contingently believes the same things but because each member believes that every other member holds (or, at least should hold) the same core set of beliefs. We might think of belief systems as ‘dynamic conversations’ in which, in the words of Berger and Luckmann (1967, p.173), “all who speak this same language are reality-maintaining others”. Does it, in that light, make sense to think of tags as, so to speak, ‘asynchronous conversations’ between taggers? Probably not: tagging remains a personal rather than a collective, collaborative, and negotiated activity—we typically do not consult with others or solicit the opinions of others before tagging an item.

Let’s go back to Cézanne’s The Large Bathers, now tagged “20th century, cezanne, french, impressionist, large bathers, oil on canvas, water” [blah blah …]

if we allow multiple folksonomies (along with traditional taxonomies), the folksonomy becomes not a statement of how the world is ordered but a reflection of the different ways a crowd orders material—some ways wrong, some right, and some just useful.

[I’ve tagged this posting “folksonomies, knowledge-interface, museums-and-heritage, tagging, web-2.0”]

For interest, a video interview with David Weinberger:


Bulmer, R. (1967). ‘Why the Cassowary is not a Bird’, in M. Douglas (ed.) (1973), Rules and Meanings. Harmondsworth: Penguin. [Amazon]

Law, J. & Lodge, P. (1984). Science for Social Scientists. London: Macmillan Press. ISBN: 0333351010. [Amazon]

Schutz, A. & Luckmann, T. (1974). The Structures of the Life-World. London: Heinemann Educational Books Ltd. ISBN: 0810106221 [Amazon]

Weinberger, D. (2007). ‘Can Tags be Wrong?’, Journal of the Hyperlinked Organization. 4th May 2007. Accessed 19th June 2007 at:

  1. I wonder whether this may, in principle at least, be an empirical question: if you want to know why a person tagged with the word s/he did, ask him/her. But I’m not yet ready to think this through.
  2. From the summary in Law & Lodge, 1984, p,69

Leave a Reply