Do we need such broad tags on questions about specific works?

Question

It has come up in chat that the usage and existence of such broad tags as russian-literature or short-stories (or mexico) on questions asking about specific works are debatable.

On one hand, they help categorize a question, for the purpose of finding it - and we have agreed that we want to use author tags for questions about specific works, for the same reason.

On the other hand, there's a whole wagon of tags that can be applied to a particularly taken work of literature.

So, do we need broad category tags on questions about specific works?

Answer 1

As a counter to Robert's post, I am of the counter argument crowd.

Many fans of literature (and dare I say "experts") will be interested in a particular language-of-origin. To me, a first-time or casual visitor will be interested in russian-literature and use that as a starting point for searching questions and/or answers. It may also aid in Google searchability; I am thinking of a question like "Are there common themes in Russian Literature?". An expert, such as a Russian Literature professor, may also do the same, as they will be familiar with many different authors in that classification. We want users to be experts in certain tags (e.g. for gold badges) and it's perfectly reasonable to expect someone to be a expert in language-of-origin tags.

In addition it may help users who are not native English speakers. While posts on this site do have to be in English, someone who has a different mother tongue may only want to find questions or give answers about their nation's literature.

My final thought on the tagging system, is I always ask myself, "What harm does this tag do?" (I originally came to this thought from discussion about character tags of SFF.SE). To expand a bit on what "harm" means: is this tag being misused in a way that detracts from the quality of question? In most cases when the answer is "yes, definitely" the tag can simply be removed. However, in this I feel that tags can do more benefit than harm.

Answer 2

Yes, absolutely! Our primary tags should be body-of-literature tags. Tags about individual authors may be useful for fine-grained filtering but they're useless for area-of-expertise filtering.

The primary purpose of tags is to convey areas of expertise, to allow experts to quickly filter questions that are within their area of expertise. Thus Stack Overflow uses programming languages as its primary set of tags, Super User uses operating systems and applications, History uses geographical areas and historical periods, etc.

The experts in literature are professors of literature. So how are university departments organized? Let's look at a few examples:

• Harvard splits literature among the following departments: Classics, Comparative literature, English, Celtic languages and literatures, Germanic languages and literatures, Romance languages and literatures, etc.
• Oxford has a Faculty of Classics, a Faculty of English Language and Literature, a Faculty of Medieval and Modern Languages which itself is subdivided in sub-faculties by languages and groups of languages.
• Paris-Sorbonne has “French and comparative literature” and a number of faculties for language and literature of linguistic groups (English, Arabic and Hebrew, Germanic and Nordic, Iberic and Latin American, Italian, Slavic).

Clearly “my language” is an important category, and “not my language” gets subdivided into language families. A given professor is usually an expert in the literature of a specific language. So languages are a primary classification for experts. Time periods are another.

The bulk of the participants on this site are readers. Many books aren't even translated, and most people are best read in the language that they are most fluent in. So for readers, too, language and period tags are important.

My conclusion is that most questions should have two body-of-literature tags: language and period. (And genre as well where applicable.) This is somewhat similar to History's use of geographical locations (united-states, europe, soviet-union, india, …) and time periods (world-war-two, 20th-century, ancient-history, middle-ages, ancient-rome, 19th-century, …) as two major categories of tags, with a third major category (military, war, economy, political-history, economy, …) being topics whose history is being studied (similar to themes and genres for literature).

Answer 3

The pragmatic answer is that these overly-broad tags will tend to be overused (and used inconsistently), so we should probably do without them entirely.

The problem with overly broad tags is that, in finding them, the author doesn't feel compelled to drill down into what the question is actually about, so you end up with a site where most of the questions are tagged in only the most superficial sense.

book question american novel ← _{a minor exageration, but we've actually seen this}

The Counter-Argument Considered

Over the long haul, we might (just might) accumulate the tiniest sliver of question about the actual subject of "Russian literature" (as a collective whole). But what is way more likely to happen is folks will label every question that happens to have a Russian author as "Russian literature" and be done with it. That's not a useful way to attract a very targeted search into your site.

_{^* Russian literature is just an example.}

But I'm only interested in Russian literature

I can appreciate that folks have various areas of interest; it's just not a great way to divvy up a site. I might only be interested in the classics (for example); someone else in satire; someone else will want to see only non-fiction. We had a site where proponents insisted that questions agreeable to vegetarians be labeled as such. It didn't work.

We already allow for questions grouped by title and author. Beyond that, you simply have too many axes to divvy up content before you run out of tags (and make tagging itself too onerous)…

…so I suggest avoiding this class of "coincidental tags" and sticking to tags where the [subject of {x}] is directly addressed in the question.