The article Definition of Videogames1, which I commented earlier, reminded me of a discussion about defining games at Digra 2007 conference. When I mentioned Wittgentein2 critique on definitions based on necessary and sufficient conditions, I got a reply that dismissed Wittgenstein critique, because he changed his mind. The presented argument, he changed his mind, is an example of ad hominem argument which is logically invalid argument.
Wittgenstein critique on the categories of natural languages seem to be accepted in social psychology (see, e.g., Kunda3); the categories of natural language has no clear borders, and whether something belongs to a category is based on similarity rather than necessary and sufficient conditions. There are of course categories that are based on strict definitions, e.g. odd numbers. But there are also categories, such as art and game, where using necessary and sufficient conditions (or cluster definitions) in defining what belongs and what does not belong to the category has turned out to be problematic.
Maybe we should take Wittgenstein critique on formal definitions seriously and start to think what implications of that in game studies, instead of proposing different definitions that can be used to categorize an arfifact just by looking at its formal features.
References
- Tavinor, Grant. (2008). Definition of Videogames. Contemporary Aesthetics 6. Available http://www.contempaesthetics.org/newvolume/pages/article.php?articleID=492.
- Wittgenstein, L. (1958). Philosophical investigations. New York: MacMillan Publishing Co., Inc., 3rd edition. Translated by G. E. M. Anscombe.
- Kunda, Z. (1999). Social cognition: Making sense of people. Cambridge: The MIT Press.
Petri Lankoski 
Greetings.
A sufficiently intricate “traditional” definition can get arbitrary close to any one concept. This is trivially true, because anything can be defined as “meets the criteria A, B,… and is not (contained in) a, b,…”, where a, b, etc. can be single cases or broader sets. If given definitions don’t work, it can be useful to take their intersection or union (or combinations thereof) and see if those are better.
Personally, I do find such definitions useful, even when they are imperfect; they often provoke new ideas. That may just be me with half a year’s studies of philosophy, though.
That said, I doubt it is possible to precisely define anything complicated without changing the meaning, unless the definition is arbitrarily restricted to a given domain. Defining something and then investigating the implications does sound useful.
-Tommi Brander
I agreed that definitions can be useful, e.g., normative definitions for design are fruitful to make visible aspects of phenomenon. Also making definitions can help to understand phenomenon a new way, even if the definition is flawed.
Definitions of cultural phenomenon, nevertheless, encounter problems that relate to extension or changes (within time) on phenomenon that definition tries to explain. This has been case with the art, as there have been appearing pieces of art that defies the definition. Due to this, some philosophers have proposed institutional (e.g., Dickie) or narrative (Carroll) theory to explain why something is art instead of using traditional definitions.
Traditional definitions do have the problem that people can and do use them as bludgeons, hence starting pointless semantic debates and stopping all useful conversation. Notable example: “That’s not roleplaying/art/whatever”. Actually building a border around a concept that changes is likely to be easier with other methodologies.
That said, I find that definition is useful if and only if it helps in thinking/talking about the thing definined in a new or more precise way.
Some people, such as the ones who give other people money if they make art, may find other use for definitions.
-Tommi Brander
“Definition time. I’ll go for normative ones, even if they can be argued against.”
Notably, we need definitions in order to discuss things.
The question is more about the scope of a definition. Useful definitions usually have only limited scope, (with that I mean claims that are formulated like “this theory is relevant to phenomenon defined as follows…” or “my essay considers subset of games that share following features…”). Alternatively, one can claim that typically phenomenon consist of … (e.g., prototypical features of games are rules or algorithm based systems, goals, and opposition).