towards a constraint deck

[Alex Fink]

The intent of this thread is to kick off discussion on what might go
in a Glossotechia constraint deck. I sat down and brainstormed some
things, below, but I haven't done the research and dredged up
constraints that appeared in prior games or anything like that yet.

Superordinate to this, I guess it's useful to discuss how to allocate
the scores for the constraints. To first order, I'd imagine that the
right thing to do for balance is to have each constraint score roughly
one point (or X points) per word that you'd expect needs to be coined
to be able to fulfil it, from a condition where the game language is
somewhat developed but no-one has been working on the particular
constraint in question: so (with this scale factor) one point for the
simplest constraints, up to the order of four or five for complicated
things.

There's also the question of who should be allowed to score a
constraint: just the first player to it, or everyone once (as long as
it's in a different enough fashion to previous approaches), or a
combination (everyone with a bonus for first)? I suppose this
interacts with whether there are private constraints; if there are,
then perhaps everyone else should get a crack at them after their
original holder fulfils them, to mitigate the potential unfair effects
of being dealt constraints of low value. Either that, or once you
fulfil a private constraint you draw to replace it.


First, some formal constraints, which I like because they establish
strategic incentives for various uses of the language change cards and
whatnot. Several of these seem naturally gradable, so they could be
scored as some factor times of the number of words involved (up to a
cap) or the like.

Make a sentence with rhymes (one point per rhyming pair of words?).
With alliteration (one point per word after the first with the same
initial?).

Make a sentence in some particular meter (iambic, dactylic, whatever:
a few copies of this with different meters could be good). This would
of course need a stress system to have been established to work.

Make an acrostic (with respect to the phonemic transcription). (It
could be allowed to be syllabic, for phonotactics that don't allow
initial V.)
Make a palindrome -- this probably shouldn't necessarily be the whole
sentence, but just some substring of words within it.
Make a sentence which is (or a substring of whose words is) an anagram
of some earlier played sentence (or constituent thereof). (That could
be a little fiddly?)

Variations on number of phonemes used: make a pangram, or a lipogram;
presumably to make the latter contentful for short sentences it should
be "use at most N different phonemes". Or these could be split out
into vowel and consonant versions of each (e.g. for the latter, use
only two Vs, resp. only 4 Cs?)

(One could throw in alphabetical order constraints here if there was
an alphabetical writing system created, but there's not.)

Make a tongue twister (bonus points if it catches someone who can
otherwise fluently use the phonemes involved?)

Make a haiku -- I'm thinking just the 5/7/5 constraint. Are there any
other well-known poetic forms or devices short enough for this?
Limericks are too long, I suppose. Kennings? Parallelism of the Old
Testament sort? ...

Make a (language-internal) pun. Make a cross-linguistic pun.

The above are largely phonological; I tried to think of some
morphologically based constraints, as another way to encourage
creation of nontrivial morphology in some games, but didn't get nearly
as far. There's always "use as many morphemes / morphological
processes in a word as you can" (people love the
_Avrupa-lı-laş-tır-ama-dık-lar-ımız-dan_ 'one of those whom we could
not Europeanize' thing).

Semantically-based constraints: the easiest of these are just of the
form "make a sentence about X", and I think these could be represented
a not-small number of times in the deck without seeming too
monotonous.
One good set of values of X would be aspects of the speakers' culture
(to be played with or without the culture rules): "make a sentence
either stating or otherwise reflecting something about the speakers'
religion / government / ..." The top-level headings in
http://www.frathwiki.com/Dr._Zahir%27s_Ethnographical_Questionnaire
would make a good dozen such cards (with a few tweaks, e.g. fold
"suicide" under "death").

A similar set of constraints I'd like, though I don't know how to set
them up, are ones like "make [a sentence of some type which is made
easier to express if there is morphological expression of some
category]". Daniel named evidentiality, aspect, topicality, agency,
modality, definiteness in the other thread.
If this doesn't cover them, throw in some of the trickier phenomena
from the extant challenge sentence cards as constraints, like
counterfactuality.

Append a sentence to a collectively-written story. This would work
best if players were encouraged to do it more than once, so the
scoring should be specially adjusted here.
In a same fashion, append a reply to a collectively-written dialogue.

Are there any interesting refinements of the idea "make a sentence
responding to a previous one" not of the cumulating sort like the
above?

Demonstrate a Lakoffian metaphor (LOVE IS WAR, ANGER IS
PRESSURISATION, LATER THAN IS BELOW, that sort of thing) in the game
language.
(Create and) demonstrate an idiom.
Posit a proverb.
Demonstrate a euphemism.
Demonstrate the use of a non-neutral register (vulgar or refined or
whatever). -- For this one I suppose the most common things for
languages to do is just variant lexical items, which is a little
boring, though other strategies are perhaps more interesting? (This
problem would go as well for "demonstrate a non-standard or dialectal
variant".)

Perhaps some cards on the "uses of language other than impassively
conveying facts" theme: make a joke or witticism (a non-pun-based one,
if puns are separate?), make a threat, an insult, etc.

Translate some song lyrics, a literary quote, a quote from a famous figure, etc.
Translate something amusing one of the players said -- I'm thinking
for instance of the amusing wrong guesses that always come up in the
charades phase, but other things would work too.


Alex

[Jim Henry]

On Sun, Feb 24, 2013 at 11:49 AM, Alex Fink <000...@gmail.com> wrote:
> Superordinate to this, I guess it's useful to discuss how to allocate
> the scores for the constraints. To first order, I'd imagine that the
> right thing to do for balance is to have each constraint score roughly
> one point (or X points) per word that you'd expect needs to be coined
> to be able to fulfil it, from a condition where the game language is

That sounds about right.

> There's also the question of who should be allowed to score a
> constraint: just the first player to it, or everyone once (as long as
> it's in a different enough fashion to previous approaches), or a
> combination (everyone with a bonus for first)? I suppose this
> interacts with whether there are private constraints; if there are,
> then perhaps everyone else should get a crack at them after their
> original holder fulfils them, to mitigate the potential unfair effects
> of being dealt constraints of low value. Either that, or once you
> fulfil a private constraint you draw to replace it.

What about this:

At the start of the game, everyone is dealt a private constraint card,
and there are also N (maybe around 3-5?) constraint cards dealt
face-up in the center. Whenever someone fulfils a private constraint,
they get 150% of its point value, and that card goes to the center of
the table and other players get a crack at fulfulling it. There may
also be (probably rare) cards in the main deck that discard and
replace one of the common constraints, like the existing cards that
discard and replace one's private translation challenge sentence.

> Make a sentence with rhymes (one point per rhyming pair of words?).
> With alliteration (one point per word after the first with the same
> initial?).

One point per *stressed syllable* with the same initial consonant.
And for rhymes, I'd like to make it a little more restrictive, say a
rhyming couplet in any consistent meter, or with equal numbers of
syllables in each rhymed part if there are no stress cards played yet.

> Variations on number of phonemes used: make a pangram, or a lipogram;
> presumably to make the latter contentful for short sentences it should
> be "use at most N different phonemes". Or these could be split out
> into vowel and consonant versions of each (e.g. for the latter, use
> only two Vs, resp. only 4 Cs?)

Lipograms might be tricky to constrain when we don't know how many
phonemes the game-language will have. Perhaps "use no more than 50%
of the phonemes in play, in a sentence containing at least N
syllables"?

> as far. There's always "use as many morphemes / morphological
> processes in a word as you can" (people love the
> _Avrupa-lı-laş-tır-ama-dık-lar-ımız-dan_ 'one of those whom we could
> not Europeanize' thing).

That's good. I'd probably say to score that one with only half a
point per morpheme or morphological process, because it's especially
good for piggybacking on coinages by other players.

> Append a sentence to a collectively-written story. This would work
> best if players were encouraged to do it more than once, so the
> scoring should be specially adjusted here.
> In a same fashion, append a reply to a collectively-written dialogue.

Those are good. It might be good to have bonus points scored at the
end based on how complete the story or dialogue seems when the game
ends.

--
Jim Henry
http://www.pobox.com/~jimhenry/
http://www.jimhenrymedicaltrust.org