usage vs. meaning

In a tag combining discussion about Egypt vs, Kemet it was claimed that usage should outweigh meaning (http://www.librarything.com/topic/86466#1838989). I think, put like that, it's absurd.

For example, I could start using "Goldylocks" for both George Washington and Atilla the Hun. That would constitute "usage". So would we then have to combine Washington and Atilla?

Okay, being deliberately flippant here. But I'm sure you get the picture. To me it all starts with "meaning".

Anybody?

And how do you define meaning?
To expand: I don't think usage should necessarily overrule meaning when usage is patently bizarre, as in the examples you give. But when differing usage indicates a differing meaning, then we're not talking about two separate things there (usage vs. meaning), so much as differences in meaning.

I'm neither a linguist nor a philosopher, so I wouldn't hazard trying to define either of those words (especially since English isn't my native language). But I do think that for practical purposes most of us have a reasonable grasp of both.

Of course usage can - and must - modify "meaning", and in doing so it inevitably redefines the "meaning" of any word. But most words / names do have basic meanings beyond "everything they could possibly mean as well", and it's those meanings that are usually more understandable than the modified "secondary" meanings (especially internationally).

But don't get me wrong. I'm not trying to convince anybody of anything here. But I do think we need to have this discussion about the general principles of tag combining.

But we don't do meaning already. You wanted to combine /zon/ with /sun/, even though the Dutch word for sun doesn't mean anything like Mandarin word for grandson. How about "love" which can mean love in English or faith in Danish? /love/ is combined with /kärlek/ (Swedish meaning love) and not /tro/ (Swedish meaning faith). Or /read/ which is combined with /gelesen/ even though the English "read" is polysemous; among other meanings, it is a command to read something. That is using "read" for "to read" or "to compel others to read" is perfectly valid English meaning. Which is not to mention that "read" is West Frisian for red.*

The only way tag combinations work is that we ignore the inherent multilingual meaning of the tag and work on what people are using it for. How pure that usage has to be, and whether and how much we take into account potential usages, is an open question, but I don't see meaning as viable for non-trivial combinations of any simple tag.

* (I suspect that in a few years, I won't be nearly as handicapped here, but the English Wiktionary isn't nearly as massively multilingual as it could be; I'm forced to pull these from mainly Western European languages. If I were pulling from something with a thousand languages, I suspect I could get a dozen meanings for /love/ and /read/.)

Just because usage can be random doesn't mean in real situations it is.

This is the anti-tag argument that because someone can tag something anything, over time every book will fill up every possible tag "slot" there is. That is, Macbeth will eventually get tagged "bluebird" and every other possible thing, and the result will be meaninglessness. This was raised as a serious argument in a scholarly journal. Needless to say, the problem is that tags aren't like subjects—where every subject is equally true the moment it's applied. Rather, you judge tags by frequency. Yes, given all time, someone will tag Macbeth with "bluebird," but there will be a million more tags of "Scotland," "tragedy," "drama," "murder," etc.

That statistics applies here. So, yes, in THEORY, I could use Goldylocks to mean Atilla the Hun. In reality, I'm unlikely to do so, and if I do, I'll be a very small percent of the taggers out there. Even in free environment, stuff clusters. That stuff is meaning.

It's also perspective. As Clay Shirky writes, "Queer Politics" and "The Homosexual Agenda" may mean the same thing, but they encode very significant differences in perspective. Ditto "Greece" and "Hellas," "South Africa" and "Azania," "Hebrew Bible" and "Old Testament," etc.

By combining two such tags, you make it harder for people to find the books they really want—the queer politics books and the "homosexual agenda" books get all mixed up. Thus arises the rule that tags should be very similar in both meaning AND usage.

If frequency of tag usage is to settle usage debates, then time should be considered an important factor as well. Arguments involving the usage of a particular set of tags are usually supported by list of "top books tagged X". But with new books and tags constantly being added to LT, those lists might look much different in six months or so. "Stable" tag clouds for works (like scotland, drama, murder, knives for Macbeth) and stable work lists for tags take time to crystallise. And, of course, it requires enough members applying similar tags along similar principles, which is an issue clouding the kemet vs Egypt discussion -- the former is used by 11 members. At which point are usage statistics assumed to have achieved stability, a level of reliability?

Arguments for favouring usage are open to another weakness: the decision to combine or not combine a pair of tags may influence the way in which users apply these tags: I get some of my tags from works' tag clouds, or take a clue from existing usage. I don't think I'm the only member to look to existing tags, tag clouds and tag combinations to bring my cataloguing at least somewhat in line with other users'.

Consider the tag American Presidents There is no President of America so logically this, in the same manner as European Monarchs, means the president of any country in the American continents. However most usage of it is for President of the United states. (These are combined).

Another case user A has tag Box 3 and user B box #3. These both have the same meaning in that both users have at least 3 boxes of books and they are listing the third. But what use is the combination? In practice the tags meaning is restricted to a single user's library.

The thing about time, though, is that tag combination is reversible. So, if at some point "Kemet" comes to be a fully standard and unmarked English-language term for Egypt, and the items-tagged "Kemet" differ in no real way from "Egypt," we can change it.

For the box3 stuff, there's no right or wrong answer. They're both kind of absurd. I'd recommend combining them, because then related tag-clouds will only have one useless tag in them... :)

#7: I think 11 users are enough to build large-scale patterns like Kemet has making it distinct from Egypt. Digging around Google Books makes me feel it very unlikely that the nature of the tag would change; Kemet users usually come from an entirely different perspective for Egypt users. In any case, we can change it in the future, and I feel no guilt about encouraging the use of Egypt over Kemet unless you're tagging something that uses Kemet (i.e. the type of book currently tagged Kemet.)

> 5 / 6: That's part of why I think the current system of tag voting is a huge improvement, and why I think it's useful to have these discussions. To me "Kemet" was simply another way to say "Egypt" (which is why I made that proposal in the first place), but you guys convinced me there's a different connotation that's potentially useful and should therefore be respected. In the old days I might have simply combined them myself*

Of course none of us have sufficient knowledge of all languages ever used. Let alone of the cultural difference that lurk behind and even within them. But, for example, I would never propose combining "Engels" and "English" just because usage suggests it (and meaning doesn't exclude it) - there is also a minority usage for Friedrich Engels, which doesn't deserve to get drowned out in the massive "English" tag.

I do think, on similar principles, maybe "love" should be separated from "kärlek" (and all the other translations of the English version) if it can also mean "faith" in Danish.

Thus arises the rule that tags should be very similar in both meaning AND usage.

And that should - in my humble view - remain the basic rule. Neither of these should always overrule the other.

* (Probably not, because I wasn't much involved in tag combining before - exactly because I didn't feel I could simply rely on my own judgment to that extent).

I think the Danish word is already so drowned by the English useage that it is impossible to do any more harm by combining.

(Funnily enough, "love" seems only to be combined with misspellings. The real tag "kärlek" is still uncombined: http://www.librarything.com/tag/k%C3%A4rlek)

I do think, on similar principles, maybe "love" should be separated from "kärlek" (and all the other translations of the English version) if it can also mean "faith" in Danish.

If this were the rule then there really should be no combining of foreign words at all, or most English words for that matter, since the vast majority of English words have multiple meanings.

I hope "hope" is a separate word at least, or 1 Corinthians 13:13 would seem pretty weird.

> 13: Most words in other languages have multiple meanings too. It's ultimately like that act of balancing rotating plates on sticks - with a near infinite number of plates involved.

> 14: Well, I guess it could also refer to Bob Hope, or be a typo for Hopi, or...

>15 Nicole_VanK: or... it could refer to the town of Hope near Vancouver B.C. I've used the tag twice so far.