This entry has moved to a new page.

http://www.mcgath.com/filkdef.html

-- Gary

(Some of this is adapted from a posting I made to rec.music.filk on April 4, 2001.)

When discussion of the question "what is filk?" comes around again, I tend to be far less interested in the specific definitions that people come up with than with the ways in which the discussion illustrates the nature of human categories and the ways in which they are constructed and negotiated. Needless to say, I am an academic.

Specifically, I have a background in cognitive linguistics -- that is, the study of how language use can be used to study how the mind functions (and conversely, how the nature of the human mind affects how we use language). One of the fields that contributed to the development of cognitive linguistics is the study of categories: how we define them, how we label them, how we hedge them, and always and ever how we talk about them.

The study that particularly comes to mind each time the discussion "what is filk?" starts to get heated is W.B. Gallie's article "Essentially Contested Concepts" in the Proceedings of the Philosophical Society (vol. 51, 1956). (I warned you that I'm an academic. You'll be lucky if you get out of this without reading footnotes.)

The basic idea is that there are some concepts (or, in usual application and practice, some categories) that, because of the nature and structure of the category, are _inherently_ destined to be contested by those who use the concept/category. The attributes that lead to a category being contested include:

- existence in a relatively complex framework, either in terms of how the category is defined, or in terms of how the category naturally occurs;

- intense social evaluation regarding membership (i.e., either membership or non-membership in the category tends to be strongly valued -- and may be valued differently by different people, even in whether they value it positively or negatively);

- the category emerges either from a cluster of highly-related, but non congruent "central cases", or from a central case that is based on extensive, highly-specific details;

- thus, in order to extrapolate the category from this central case or this central cluster of cases, people must prioritize various aspects of the complex frame;

- because the complexity or non-congruence of the central case(s) means that all extrapolations must focus on only a subset of attributes, and because people prioritize and evaluate those attributes differently, these extrapolations will result not only in differing, but even possibly in nearly non-overlapping category structures for different people;

- and because people place a value (positive or negative) on category membership, these differences in larger category structure will be a subject of public debate and of a public struggle for the right to define the category.

One of the (literally) textbook examples is the category/concept "art". Another popular one to use in teaching the concept is "mother". (Where the first is generally seen as being the "non-congruent central cluster" type, and the second is generally seen as the "highly over-specified central model" type.) To elaborate only slightly (before moving on to filk), the category/concept of "art" has a central cluster involving attributes like "human-created", "representation of the world", "skilled hand-created artifact", "induces an emotional response", "esthetic rather than functional", and so forth. The point is not that any one person necessarily defines "art" by applying _all_ of these, but that when you study large scale patterns of judgement about whether things "are art", these are themes that show up being shared by significant subsets. And when one of these attributes is significant to a person's definition of "art", their judgement of how good an example of "art" something is will be proportional to how well it fits the attribute. A category-model of this type means that it's possible that there is no single object that _everyone_ will agree is an excellent example of art.

In contrast, a category like "mother" begins with a central model that exists within an extremely complex frame, with an extensive set of attributes that _can_ co-exist, but where many cases also occur where some attributes are lacking and yet there are useful reasons to evaluate the example's category status. For example, the central model of the category "mother" will include attributes like:

- female provider of genetic material
- locus of gestation
- female provider of nurturing
- partner of father
- etc.

If all of these attributes occur together, it's unlikely that anyone would dispute that the person in question is a "mother", but it often happens that a person will be considered a perfectly good example of the category without having all the attributes. (Note that the context in which the category is evaluated can affect which attributes are important in defining the category. For example, in the particular circumstance when risk for a genetic disease is being evaluated, it wouldn't make sense for a step mother or adoptive mother to insist "But I'm her _real_ mother", whereas if the question is, "Who should I send a card to on Mother's Day?" entirely different criteria are prioritized.)

"Filk" seems to me to be an excellent example of an essentially contested concept. It arises in a complex frame; it is strongly evaluative among those who use the category; there is a group of concepts/attributes that are associated with central model/models; and there is a constant and consistent desire to extrapolate the core category in order to evaluate "marginal" items for their category membership. (Is this song a filk? Is this performer a filker? Who should we give filk awards to? Etc.)

The central structure of the category partakes somewhat of both types discussed above. From a historic viewpoint, the "overspecified" model (i.e., the "mother" type) tends to make sense. From a contemporary viewpoint, the "cluster" model (i.e., the "art" type) seems more accurate. In either case, a lot of the relevant attributes are going to be similar. So the trick here is to discard the notion that we're talking about "necessary and sufficient conditions" -- that if we say "characteristic X is part of the central model of "filk", we're somehow excluding or devaluing all songs or performances that don't include characteristic X.

So without weaseling around any more, what are some of those characteristics? We can identify one of them by looking at how "filk" is used as a transitive verb. If I were to say, "I filked X" (where X is either a song or a song-writer), I think there's general agreement that I could only mean, "I wrote a parody of X (or of one of X's songs, in the case of a writer)". So part of the central cluster of attributes is "parody". What else can we uncover from default assumptions when the word "filk" is used? Suppose I were to say to you, "I had a lot of fun filking last weekend." I'll assert that your first assumption will be that I attended a convention last weekend. Only with other cues would you _first_ assume that the filking happened in some other context. So my second central attribute will be "done at a sf convention".

Another way to discover attributes that are part of the central cluster is to survey people's unexamined judgements as to "better" and "worse" examples of the category. In the laboratory, there are all sorts of ways of testing this -- most involving extremely rapid judgement (to get the "unexamined" part), and often involving actually measuring milli-second differences in how long it takes people to decide category membership. ("Good" members of a category are identified quickly, more marginal members take longer.) I'm not in a position to do these sorts of experiments, so the following is unscientific observation.

Due directly to the historic development of filk by sf fans and writers in the context of sf conventions, it is unsurprising to find a subject-matter core characteristic that might be generalized as "dealing with the same topics, settings, motifs, and concerns as those found in the creative material that inspired the creation of sf conventions". That may be overly broad, but I'm feeling cowardly at the moment. I'll go more out on a limb and assert that, certainly from a historic point of view, acoustic performance is likely to be a central characteristic.

Another way of evaluating central characteristics is whether the lack of a particular characteristic or group of characteristics creates an identifiable (and named) sub-set of the category. So, for example, in the category "mother" the existence of the category label "birth-mother" points to the distinction between the "locus of gestation" and "provider of nurturing" models. A woman who has carried and lovingly raised her own genetic offspring is technically a "birth-mother", but to call her one automatically creates the implication that she was the gestator (and possibly genetic contributor) but _not_ the provider of nurturing. Similarly, the existence of the category label "found filk" emphasizes the existence of a central characteristic "created within the 'community'" (where "community is defined by other central characteristics), as opposed to an appropriation by the community. And as long as I'm touching on the slippery topic of "community", I'll offer as a central characteristic that "filk" is performed by and for that community (where "performed", "by", and "for" are all relevant).

So the theory is that an event that involves _all_ of these "central characteristics" is one where nobody would dispute that it is a member of the category "filk" -- even if they might not place it at the center of their understanding of the category. And, if we come up with the proper list of central characteristics, then an event that comes up negative for _all_ the characteristics is one where there will be universal agreement that it isn't a member of the category "filk". But between those two extremes, different people will evaluate membership differently, depending on how they prioritize and emphasize the various central characteristics. For one person, the parody sub-model may be the key one extended to map out new territory for the category. For another, the crucial _sine_qua_non_ may be the aspect of "created within the community". For another, subject matter may be at the top of the list. In all cases, people are not likely to be extending only a single central model, but a cluster of them, perhaps with different weighting (e.g., "It's _more_ important what the subject matter is, but it's really hard to consider it filk if nobody has ever actually sung it at a filksing.")

It has been objected that, if accepted, this way of "defining" the category of filk is useless, because it is no definition -- it leaves us in doubt, confusion, and disagreement. But clearly (based on the discussions I've seen) we _are_ in doubt, confusion, and disagreement as to exactly what the category "filk" includes and excludes. Is it more valuable to reject the existing state of affairs in order to create an official "necessary and sufficient conditions" definition for filk that has the virtue of precision, or is it more valuable to describe the "definition" in a way that reflects the variability of actual usage, and yet captures the patterns of agreement that do exist?

The point of understanding "filk" as an essentially contested category is that it will _never_ be possible to draw up "necessary and sufficient conditions" for identifying something as filk -- and that this isn't a problem. Rather, that any definition of "filk" must understand it in terms of a cluster model with extrapolations, and with the understanding that people will prioritize different parts of the cluster and different extrapolations differently.

The link for the Nick Smith article is no longer usable since I split up the filk FAQ a few months back. It should now be the same as the link to my page on filk definitions, since it's the main entry thereon. That's http://home.earthlink.net/~kayshapero/filkdef.htm

Good idea, and if it's ok with you I'm going to link to this page from the filk definition page.

Hi,

I'm working with ShowMeCon, and we're trying to get a filking track started. Any ideas on how to let the filking fans know about it?

Thanks!

Mary

I have always tried to offer my friends and family a fairly simple definition. This way, it can be summarized without too much difficulty and -perhaps- understood that much better. This is how I define filk:

Filk - A form of music not identified by its harmonic, rhythmic, lyrical, racial, or "traditional" musical qualities. Rather, it is music defined by it's content: music that is by, for, and/or about science fiction, fantasy, horror, science, and/or subjects of interest to fans of speculative fiction.

Wordy, yeah ... a bit. But also rather direct.

What do you think?

Base URL: http://www.filking.net
All postings, columns, articles and artwork belong to their respective authors. Everything else ©copyright 2007 Debbie Ridpath Ohi.
Reproduction and/or distribution of the whole or any part in any form
is forbidden unless prior permission has been granted.