On 3/10/07, Alex Fink <a4pq1injbok_0@...> wrote:
> On Mon, 5 Mar 2007 11:32:41 -0500, Jim Henry <jimhenry1973@...> wrote:
> >I obviously need to work on coming up with a balanced translation
> >challenge deck where all the sentences are of roughly similar
> >difficulty.
>
> I still like the idea of player-created translation challenges to, well,
> offer more variety than a fixed deck of cards can.
I like the idea too, but so far the one time I've tried it the
results were sub-optimal.
> I wonder whether there's
> any way to get non-linguistically-oriented players to generate sentences of
> reasonably balanced difficulty -- perhaps to ask for sentences of some fixed
> number of English words?
I think I asked them for sentences of around five or six words, but I got
several that were longer.
> Alternatively, it might be possible to write some computer program to
> generate sentences of uniform difficulty. Even if you don't want it to be
> necessary to call on a computer for every game, you could have it generate
> you a new deck of sentences once the players have grown too familiar with
> their current one.
Maybe so. I think it will require some playtesting to figure out
what kinds of words and sentence structure actually make a
sentence more difficult. Tentatively I might guesstimate one "point"
of difficulty for each of
- non-indicative mood
- an abstract word
- use of conjunctions
- use of preposition(s), esp. non-spatial ones
- use of relative or subordinate clause
Etc.
> >The playtesters had a few suggestions for improving the game, including:
> >
> >- starting out with some random phonemes and syllable cards already
> > face down, so players can start coining words on the very first turn
>
> Good. Or have a very fast first round where each player is constrained to
> put down a phoneme or syllable card?
That's basically how it started -- the first several players played
phonemes or syllables if they had them, and after we had enough
to start coining words, the next player coined a word on his turn.
> >- modular translation challenges with mix-and-match subject and
> > predicate cards
>
> Kind of a less extreme version of complete dynamic generation by computer.
>
> Would every predicate, and every subject, have the same complexity?
> Otherwise, what's to stop you from drawing a more-difficult subject with a
> more-difficult predicate?
Hm; I expect the modular translation challenges would tend
to vary randomly in difficulty, then. It would be a tight constraint
to come up with a set of subjects all equally complex and
a set of predicates all equally complex.
>
> >- action cards that let one player place requirements on what another
> > player can do on their next turn; e.g., must coin a word of a
> > specific part of speech, or must use their new word in a sentence
> > immediately
> >
> >- action cards to let players retrieve cards from the discard pile, as
> > in Chrononauts
>
> I like these two, and they look implementable without breaking anything else.
I've already made these cards though I haven't had a
chance to use them yet.
> >- having two-stage missions, with players drawing an "easy"
> > translation challenge card at the beginning of the game and then
> > drawing a "hard" translation challenge card after they have
> > translated their first sentence
>
> That would make for a really long game, wouldn't it?
Yes; maybe you would use one translation challenge deck
or the other when you have less time to play, and use the
easy deck followed by the hard deck when you have more
time.
> >- add more incentives to express sentences in the gamelang, besides
> > the basic game-winning goal of translating your challenge card
>
> Hmm, how?
I thought about letting players draw an extra card whenever
they manage to say a sentence in the gamelang.
> >- the translation challenge sentences should have significant lexical
> > overlap, so once players get familiar with the deck, they can't
> > guess what sentence someone else has by the first few words
> > they coin.
>
> There's a tradeoff here: the less lexical variety there is in a given deck,
> the easier it is for players familiar with it to guess what word someone's
> trying to define.
Right. Maybe the optimal number of times for a given lexeme
to occur in the deck is about two or three -- enough that
coining that word won't give away what challenge card you've
drawn, but not so many that it will occur in more games than
not and make it too easy for other players to guess your
charade.
> >I also noted that I probably need to have a slightly higher proportion
> >of syllable cards to phoneme cards, to get word-coining play started
> >as soon as possible in the early rounds, and need a higher proportion
> >of grammar change cards in the deck. After the game was over we
> >talked about how the number of phoneme and syllable cards in play
> >tended to grow without limit and that there should be some mechanism
> >for limiting the number of either in play at a given time. I'm not
> >sure what would be the best way to do that -- maybe add several cards
> >setting a maximum phoneme inventory limit (10, 15, 20, 30...)?
> >Or roll dice at the beginning of the game to set the initial limit,
> >and add action cards that allow a player to increase or decrease the
> >limit by an amount given on a new die roll...? What do y'all think?
>
> Would it help to treat the phonotactics cards more like the syntactic cards?
> So instead of having a card for each syllable structure, you'd have e.g.
> coda cards 'no codas', '(N)', '(C)', '(C)(C)', etc., and separate onset
> cards.
I thought about that earlier, but I decided the single set of syllable
cards might be easier to explain to linguistic novices.
>The problem then is that it's not clear what to do about replacing a
> card with a more restrictive one when there are already words with the less
> restrictive phonotactics: force a sound change?
I actually have a sound change card called "Eliminate cluster"
which does exactly that. It forces a sound change at the player's
option on all words formed with clusters of that kind -- insert
epenthetic vowel, drop the first consonant, or drop the second
consonant. It was played once during our first game, removing
the Fricative-Plosive-Vowel card and turning words like "tsi" into
"tisi".
> In fact, the same problem seems to manifest itself however you try to
> decrease the limit: what happens if there are 18 Cs in play and then someone
> plays the 'upper bound 10 Cs' card?
Yes, that's a problem. I suppose the player who plays the
upper bound card would have to devise a set of phoneme
merges that reduce the phoneme inventory to the new limit.
> Here's an approach that makes sound change more important: set a hard limit
> on the phonology via dice at the beginning of the game, and perhaps have a
> couple infrequent action cards that increase or decrease it. But then allow
> phoneme splits (and any other sound changes that introduce a new phone?) to
> disregard the limit.
That's probably good.
> >descriptions, or email me asking for a detailed list of the cards in
> >my deck if you're interested. I'm considering making an Esperanto
> >version for the next couple of E-o conventions I go to; depends on how
> >much free time I have before them.
>
> Yeah, I wouldn't mind seeing the list of cards. (Send it to 000024 at gmail
> if you want to do so offlist.)
I don't have the deck with me here -- I'll email you a list next
week sometime.
--
Jim Henry
http://www.pobox.com/~jimhenry/gzb/gzb.htm