Definitions

[Charlie Reams]

If one were to try to design a word game in which definitions did matter, what would you come up with? The obvious approach of just demanding a definition for every word played introduces a lot of problems with subjectivity, with specificity (yes, a towai is a tree, but that doesn't seem like much of a definition), and with the basic complexity of definition writing (what is love?).

Any better ideas?

[Gavin Chipper]

It depends on how the game works, but if it's done on a computer, it could come up with a list of possible definitions (four or something) and you have to select the right one. Also, it could be that you lose all your points for the whole game if you get one wrong (or at least enough to make it a deterrent) so it's not just a case of having a 1 in 4 chance of guessing right but much less than that because you'd have several per game. You then might want an indeterminate end to the game to stop people doing the old last round gamble when they know they're going to lose if they don't make a random guess. Or you could just accept the last round gamble as part of the game.

[Matt Morrison]

Gev on the nail with the "pick from four" method - if every word has its definition listed too then it would be a simple "supply the correct definition, pick three random from the whole of the rest of the database, randomise order" for the computer.

[sean d]

A couple of issues... Many many words have multiple definitions, presumably any definition in the dictionary would be acceptable. In the case of SET or PLAY or HEAD there could be a dozen or more. If you're given a choice of four definitions are you expected to know all possible definitions? eg I see TANGENT which I know is a maths term but the correct definition offered among the four is about TANGENT in terms of irrelevant conversation. (I'm sure there are better examples of words with two or more obscure unrelated meanings)
Also, are players going to have to define every word, or where do you draw the line in terms of obscurity?
Then there's the question of how tight you have the definition... For OAK is 'plant' the definition, or 'tree' or 'deciduous tree' or 'a tree of the genus Quercus'? And how tricky do you make the alternatives?

[Adam Gillard]

Free Rice does this if anyone wanted to look at a working model (Google it; it's a charity word game).

[Charlie Reams]

Gev on the nail with the "pick from four" method - if every word has its definition listed too then it would be a simple "supply the correct definition, pick three random from the whole of the rest of the database, randomise order" for the computer.

I think you'd have to be a bit smarter than this. Often the general form of a word tells you what part of speech it is, which would eliminate a lot of the definitions. Also the multiple definitions problem that Sean mentioned. Most difficult of all, the program would have to ensure that none of the other definitions actually matched the word. But there's probably the kernel of something workable in here.

[Matt Morrison]

Gev on the nail with the "pick from four" method - if every word has its definition listed too then it would be a simple "supply the correct definition, pick three random from the whole of the rest of the database, randomise order" for the computer.

I think you'd have to be a bit smarter than this. Often the general form of a word tells you what part of speech it is, which would eliminate a lot of the definitions. Also the multiple definitions problem that Sean mentioned. Most difficult of all, the program would have to ensure that none of the other definitions actually matched the word. But there's probably the kernel of something workable in here.

Oh that's true yes. So perhaps database would need "type" as well as "definition" to choose randomly from other words of the same type. Not matching accidentally would be more annoying to work around.