ARCHIVE: A possible new conlang art project

19 views

Skip to first unread message

Sai Emrys

unread,

Sep 9, 2008, 1:51:40 PM9/9/08

to conlang-...@googlegroups.com

Forwarded conversation
Subject: A possible new conlang art project
------------------------

From: Sai Emrys <s...@saizai.com>
Date: Thu, Aug 7, 2008 at 10:41 PM
To: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>

Paul -

First off, let me give you, yet again, a big thanks for your help with
the pin design and LCS seal. The pins have been ordered, so you'll get
to see a real physical fruit thereof.

Second, I'd like to ask whether you'd be interested in designing a
deck, for Glossotechnia -
http://bellsouthpwp.net/j/i/jimhenry1973/conlang/glossotechnia.html;
possibly in conjunction with Matthew Haupt's variant -
http://makealang.blogspot.com/search/label/card%20game

This would be for production into an actual card deck, poker size. Jim
and Matt, while no doubt great conlangers, are not great at visual
design.

The scope of your participation in this of course would be entirely up
to you, from simply composing a template for the front/back of a card
to helping refine the rules into something that could be made into a
simple and comprehensible rulebook such that it could be bundled the
cards, making an entirely self-contained and
normal-person-comprehensible set.

The timeframe for this is to finalize it for sending to printers by February.

I don't know yet whether this will happen, but you're the first person
I thought of when reading Jim's tentative proposal to make (by hand)
such a deck for LCC3.

Thanks,
- Sai

----------
From: Paul Schleitwiler, FCM <pjschlei...@gmail.com>
Date: Sat, Aug 9, 2008 at 11:27 AM
To: Sai Emrys <s...@saizai.com>

Sounds interesting. Designing a graphic for the cards should not be
hard. Designing a rulebook and play aids is more, but may be easier
since they have already done a lot of development. The card examples
in Matt Haupt's posting are good, clean, graphic text. Since the cards
display what to do and give an example, the rulebook does not have to
be large or complicated.
Have Jim and Matt contact me directly and we can start a discussion.

BTW, I am also working with another designer and the author on a
children's book about butterflies.
Dio lo benedice sempre, tutte le maniere,
Paul

----------
From: Sai Emrys <s...@saizai.com>
Date: Sat, Aug 9, 2008 at 12:33 PM
To: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>

*nod* Glad to hear it, will do.
Hee! /me wanna see

- Sai

----------
From: Sai Emrys <s...@saizai.com>
Date: Sat, Aug 9, 2008 at 12:35 PM
To: Jim Henry <jimhen...@gmail.com>, Matthew Haupt <film...@hotmail.com>

Jim, Matt -

Please contact Paul per below to collaborate on the art for the project.

FYI, he's the one who made the revised version of the LCS Seal &
membership pins; he definitely knows his stuff.

- Sai

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Sun, Aug 10, 2008 at 5:14 PM
To: Paul Schleitwiler <pjschlei...@gmail.com>
Cc: Matthew Haupt <film...@hotmail.com>, Sai Emrys <s...@saizai.com>

The rules for the advanced version can probably, paradoxically,
be shorter than the rules for the basic version, since we can assume
knowledge of linguistics on the part of the readers. In the basic
version we'd have to explain things at greater length and give more
copious examples, I think. In any case, we're going to do the advanced
version first because it's been more thoroughly playtested.
I am fixing to work on an expanded, more detailed
version of the deck description document. When I finish a draft,
I will send a copy to you as well as Matthew.
Yes, I'm pretty well satisfied with the rules for the advanced
version, I mean the Platonic ideal of the rules implied by
the actual gameplay; but the text describing those rules
could stand improvement for clarity.

--
Jim Henry
http://www.pobox.com/~jimhenry/conlang/fluency-survey.html
Conlang fluency survey -- there's still time to participate before
I analyze the results and write the article

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Mon, Aug 11, 2008 at 11:30 AM
To: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, Matthew Haupt <film...@hotmail.com>

On Mon, Aug 11, 2008 at 1:59 PM, Paul Schleitwiler, FCM
<pjschlei...@gmail.com> wrote:
> Have you any ideas about whether and what you might want for graphics,
> e.g. the card backs?

Not really. I suppose the backs of the cards should have
the name of the game on them and some suitable image;
what, I'm not sure. Maye the Conlang Flag logo?

As for the front of the cards, the IPA symbols on the Phoneme
cards need to be in a very readable font that clearly distinguishes
capital I from lowercase l, etc. And it might be nice to have
little pictures illustrating the example phrases and sentences
on the syntax cards, but it's not critical. I'm not sure what kind
of font(s) the rest of the text on the cards should be in; anything
that's clearly readable should be fine.

Someone suggested color-coding the phoneme descriptions
on the phoneme cards to match the phoneme type symbols on the
syllable cards. E.g., on the cards for /m/ and /n/ the word
"nasal" would appear in blue, and on the syllable cards like
FN (for fricative + nasal onset) and VN (for vowel + nasal rime)
the large letter N would also be in blue, and so on for the other
manners of articulation: fricative, plosive, approximant and so forth
each having their own color...

Or maybe in the lower left corner of the cards for nasal
consonants there would be a blue N a bit smaller than
the N that appears in large print on the syllable onset/rime cards,
while the phonemic description text like "bilabial nasal"
would be in black or whatever the normal foreground text color
is going to be.

> Do you want proof reading or editing help with the rules besides layout?

If you could read through them and point out any passages that seem
to you unclear, that would help.

----------
From: Sai Emrys <s...@saizai.com>
Date: Mon, Aug 11, 2008 at 11:53 AM
To: Jim Henry <jimhen...@gmail.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Alex Fink <000...@gmail.com>

Good default. One could also make a more custom logo putting more
emphasize on some sort of game or construction visuals, but that's
more work. :-P
Possibly a good idea, possibly confusing. You can't have more than ~5
color categories before it's too much for normal people to process in
a beneficial way. So make sure it's limited to helping specify
critical, broad categories.

Another thing is that cards of different types should have distinctive
design - e.g. a background texture, border color, or logo - so that
it's immediately obvious what type of card it is. (Think, e.g., Uno.)
This might be better put explicitly if it can be done in the space
limits. Code letters are confusing and take extra processing to
remember.

BTW, I'd suggest that the rules regarding charades and target
sentences be worked out better.

Current problems AFAICT:

* charade round, despite being the majority of gameplay, is only very
weakly tied to the mechanic
* target sentences could be remembered; there should be some good way
to generate them instead of using (even a large set of) pre-set
targets
* good strategy is to just relentlessly go for your target words to
get as fast a win as possible. It would be nice to balance this with
complexity, e.g. by awarding more points for a more constrained or
complex grammar - i.e. ensure that the "competition *and* cooperation"
aim is baked in to the rules

- Sai

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Mon, Aug 11, 2008 at 12:23 PM
To: Sai Emrys <s...@saizai.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Alex Fink <000...@gmail.com>

Yes, if we do that it would probably be just for manners
of articulation of consonants, with another color for vowels,
maybe.
Yes, right.
At present the cards have these symbols in large print, plus
more detailed descriptions in smaller print. So you can get
an idea about it looking at it at a distance, or get the details
by looking at it up close. Similarly with the syntax cards,
having e.g. "SVO" in large print and then "subject - verb - object"
in smaller print below that.
The best way I've come up with so far to do so *within card-game
mechanics* is the mix-and-match subject and predicate decks.
If there are e.g. 100 of each, there's a myriad sentences
to be generated.

Or maybe there could be a sentence template deck and
several other decks with words of specific semantic/syntactic
categories; draw a template card like A N V A N
and then draw two Adj cards, two Noun cards and a Verb card
to fill in the blanks.... that could get complicated.

If we supplement the cards with entirely different means of generating
sentences -- e.g. a program you would run to generate a file of
sentences of roughly equal difficulty, print it out and cut it up and
let people draw one sentence each -- that might work better
in some ways, but qua card game it weakens the consistency;
not everyone who would want to play owns a computer or
has access to a computer on which they have the right to
install their own software.

If we tell players to generate their own challenge sentences
before each game or every several games, we need to
give detailed guidelines on how to make sure the sentences
are all of roughly equal difficulty, and even then there will
probably be more variation in difficulty than is ideal.

Ideally we would offer both -- there would be a set of challenge
sentences (or subjects and predicates to mix and match)
that come with the game, plus ideas and resources for making
more sentences when you've used up the ones that come
with the game.
Any ideas about how to do that?

--
Jim Henry

----------
From: Sai Emrys <s...@saizai.com>
Date: Mon, Aug 11, 2008 at 1:07 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Alex Fink <000...@gmail.com>

Could it, for instance, be generated by drawing normal cards in dual purpose?

The core cards already are ~160 in number, which is a very large
amount. Adding another hundred or two would take this out of the range
of 'easily transported spontaneous game' and more into the zone of
'Big Box of Impossible To Shuffle Cards' (a la trivial pursuit).

Possible mechanics:

1. Hidden draw normal cards until you have sufficient constraints.
Write down sentence on a scrap of paper, hidden. Put cards back,
reshuffle, repeat for each person, then play.
2. Add extra rules to the bottom of normal cards so that they double
as BOTH subject / predicate / meta cards for the rules phase, AND
normal gameplay cards. There are some games that do this, eg by having
a special line of text at the bottom of every card in a different
font. Then you can use some other generation rule, without having to
ship an extra couple hundred cards. May require putting cards back in
deck & writing it down if the removal of those cards would mar
gameplay.
3. Have a couple cards that purely have target-gen rules on them,
dictated eg by dice roll or similar mechanic. E.g. D&D magazine style
"elvish name generators" come to mind as a similar thing.
4. Have players make their own sentences, possibly under card-drawn constraints.
5. Have rules regarding the value of various sentence constraints,
such that you can make whatever sentence you like, but it's worth very
little unless you take several constraints. Think e.g. RPGs that use
'flaws', like http://en.wikipedia.org/wiki/GURPS#Advantages_and_Disadvantages

I like #5, especially combined with a cooperativity rule, in that it
could make for long or short games (and one can set a minimum target
difficulty to effectively dictate minimum gameplay time) and requires
essentially no hard-coding or extra materials.

Players can then strategize on what difficulty they want to aim for,
what others might be doing, etc, adding another level of interest. Or
you could have a rule variant that the difficulty level is fixed.
It could be good to have this strictly as supplement, but I'd suggest
that it must be possible to play the game just fine, repeatedly, with
a set of cards (+ whatever else is needed) that is small enough to be
easily carried around.
Possibly so. In any case, there should be at least a way for n00bs to
get easily started - though this could be as simple as a couple
handholdingy examples of how to make a target sentence per my mechanic
#5.
One simple idea: you get bonus points based on the number of
constraints on the result, that are not part of your original target.

Alternately, you get bonus based on how many of others' constraints
you've added, so that the closer they are, the more bonus for you is
involved. However, this could upset the scoring balance - it'd need to
be less valuable than the core scoring. It would allow for a
poker-type 'milking' strategy, where you go for more points at the
risk of losing it entirely.

One issue, though: a point-based mechanic implies multiple plays,
which in turn means that each playthrough should be relatively fast so
as to allow a half hour or hour of play to aggregate enough
playthroughs for the point total to be more relevant than an
individual win. That is, each playthrough would be essentially a
trick. [in the Hearts etc sense]

- Sai

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Mon, Aug 11, 2008 at 6:04 PM
To: Sai Emrys <s...@saizai.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Alex Fink <000...@gmail.com>

There's a lot to digest here and I'm not sure I've understood it all.
So start by drawing a series of cards from the main deck, which
would have some key text re: challenge sentence formation as
well as the main text re: Phoneme, Syllable, Syntax or whatever,
and then put the stuff from those cards together into a sentence
and write it down; then let the next player do the same...?
I like the general idea, but it's still a little nebulous.
OK, so maybe we have Subject and Predicate decks as at present,
but instead of fixed phrases printed on them, they have templates
with 2-3 slots to be filled in with a random adj, noun, verb, etc...?
Or maybe every card in the main deck has a challenge sentence
template with an associated point value at the bottom of the
card, but you look up the words selected by your die rolls
in tables at the back of the rule book?

Like subject "The [adj] [animate noun] ....." with two tables to pick
the specific adj and noun with a die roll, and predicate
"..... [verb]s [inanimate noun]s in [placename]", with three more such tables?

Designing the templates and tables so that all the sentences
generated would make at least a little bit of sense would be
harder than designing sets of subjects and predicates such
that each of them could go with any other. But it would give a
wider variety of sentence types, incl. some with dummy
subjects or inanimate subjects, so that's a plus.
Or make sentences for each other, each not knowing which other player
has said sentence; or each player makes two sentences and of the
N*2 sentences, N are drawn by the N players.
That makes sense. I have had vague plans to make a larger
more varied set of translation challenges marked with point
values based on an estimate of how hard they are.
How do we put point values on the constraints? Here's a wild
stab at it:

+1 point for each concrete word (fairly easy to draw, charade etc.)
+2 points for each abstract word (harder to draw, charade etc.)
+2 point for non-indicative mood
+2 for subordinate or relative clause

It would need a lot of playtesting with these changed rules
to make sure the point value system is really balanced.
And there would probably be arguments about whether
a particular word counts as abstract or concrete...
One of the optional rules (which I've never been able to use yet
for want of time) is that the game goes on until external
circumstances force an end to it, and the winner is whoever
has translated the most sentences by then. We could modify
this by taking into account the point value of the sentences
translated.
Yes, absolutely.
Details? Example?
I'm not sure how to make that work. The multiple plays in a
game should build the language cumulatively, anyway; but
I'm not sure how to make the game a lot faster than it already
is. The games I've played with simple translation
challenges of 4-5 words lasted around an hour, and those
with mix-match subjects + predicates adding up to 5-7 words
around 1.5-2 hours. Maybe one could starting the game off with
everyone having very simple 2-3 word sentences (or phrases?)
to work on, and then each person draws/generates another
sentence as they finish their first, and when the group challenge
is completed another more complex group challenge is
drawn/generated?

Games with experienced players would probably go faster
than most of the playtesting games so far, all or most
of which involved at least some players who had never
played before. But I doubt playing with all experienced
players would cut the runtime in half or even by a third.

Another thing the point value mechanic might help with:
maybe when the group challenge sentence is finished,
everyone who contributed to it (by coining words or affixes
needed for it or specifying grammar rules needed for it)
gets one or more points depending on the scope of
their contribution?

All these things make the game more balanced, perhaps,
and more open-ended (so it won't suddenly become boring
when you run out of challenge sentences) but also more complex
and harder to describe, and would require a lot more playtesting
before we decide the new rules are well-balanced.

----------
From: Alex Fink <000...@gmail.com>
Date: Mon, Aug 11, 2008 at 8:55 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>

Talking to Sai I think I've just understood what it is he means to do
in his point #5 (it wasn't clear what a "constraint" was), and it
means radically, and radically, revising the goal of the game. But
now that I think I see it, it's actually quite exciting.

I suspect I'm still missing the point of some of Sai's other suggested
mechanics; they all seem a bit similar, and some of the things he's
suggesting making cards to do are better off part of the manual (like
extra rules of any sort, #2 and #3), it would seem.

First of all, there's this issue of card bloat. I don't think the
situation is quite as drastic as Sai sees it -- I know Jim carries
around all 160-odd main cards plus subj/predicate cards in a single
bag, and they could be boxed up perfectly portably in a box not much
larger than four standard decks of playing cards stacked. But it is a
largish set, yes. I earlier proposed (and still like) the idea of
obviating the need for dice by printing little indices 1 through 6
discreetly in one corner of the main cards, and I guess in suggesting
combining the challenge-sentence cards onto the main deck Sai is doing
something similar. It could work, modulo nonobtrusive visual design
etc. Not doing it would prevent having to write down your sentence,
though; there's something pleasant in the simplicity of just being
able to set two cards down in front of you and go.

So. The goal of Glossotechnia as we know it now is to be able to
translate given sentences, first a public challenge and then a private
one. This naturally extends to not stopping the game after one
sentence but making multiple and getting a point / difficulty level
many points for each, or perhaps sharing the points proportionally for
the common sentence. (Sai's "trick" but I think the word is inapt --
it suggests way way shorter rounds.) But as you say that's never been
tested for want of time. I think it'd work quite well, anyhow.

If one wants the goal to remain of this form, then the problem becomes
generating sample sentences, and this is a problem that's been given a
lot of thought already. If the sentences are to come from cards, then
something like predicate and subject decks is probably optimal.

One could refine to even more templates, per Jim's "the [adj] [anim.
noun] ...", but this seems rather niggly, requiring either lots more
cards and in particular lots more kinds of cards (increasing the
mental load), or else lots of lookup tables in the manual. There's
also the design issues:

Jim> Designing the templates and tables so that all the sentences
If the goal was to remain of this form, my recommendation would be
much like yours. In particular:

* ship the game with a moderately ample set of sentence cards,
probably in subj/predicate form or something else not too complicated,
perhaps enough for ~eight unrepeating games

* suggest in the manual the variant that the players make their own
challenge sentences, and give a set of guidelines for achieving
appropriate difficulty, likely some elaboration on

Jim> +1 point for each concrete word (fairly easy to draw, charade etc.)
* also include in the manual a link to a webpage (for ease of
accessibility) with a script that generates difficulty-tailored goal
sentences, and can in particular generate a list of them to print and
cut out

Now, we could totally change the goal. You may recall that, once, I
suggested a goal almost the flip of the current one, where instead of
trying to translate a given sentence you were trying to provide an
interpretation to a given phonetic string. This has a bit of that
flavour, but a bit of many others as well.

I'll proceed to sketch a possible (underspecified) ruleset in some
detail. Of course, this is only a starting point, and certainly may
be improved by some variations. The mechanics of the ruleset are
based on a short-lived pencil-and-paper graph game my IMO team
invented during a training camp in 2001, called 7.9. I'm sure the
general idea is attested out there elsewhere in the game design world,
but my expertise with this world is shallow.

So. The goal of this game is not to translate a single fixed
sentence, but to produce sentences which fulfill a number of smaller
goals.

Each player would have a scoring pool of sentences which they will
have considered in the scoring at the end of the game. They may add a
sentence to the scoring pool on any of their turns, by composing it
just as they do in standard Glossotechnia. Each player may have
multiple sentences in the pool. To prevent spamming the pool, there
should be a rule that every sentence added to the pool must be
sufficiently distinct from every previous pool sentence, where perhaps
"sufficiently distinct from" might mean 'contains at least three words
/ roots not found in'. (Maybe it should also be possible to retract
your sentences from the pool; we can adjust the fine mechanics later.)

The smaller goals, in this variant, would be chosen freely by the
players at the start of the game. This ruleset unabashedly assumes a
complete set of familiar players at the start of the game; I haven't
thought about how to accommodate newbies yet.

In the setup phase, every player would come up with a number of goals
(or "constraints"), maybe five. Each goal would be a property that
sentences in the scoring pool may or may not have. Goals could be
boolean -- you're trying to make sentences that have them, that's all
-- or they could be scalar -- you're trying to make sentences that
have the most of some particular feature, compared globally to all
sentences of all players in the scoring pool.

Constraints could be

_phonological_:
* use as many velars as possible
* use a phoneme from each point of articulation in the phonology
* avoid fricatives
* use the initial phoneme of each player's name as the initial phoneme
of one of your words
* use only words that observe front-back vowel harmony
* use at least two five-phoneme syllables
* use at least four two-syllable words
* use the longest palindromic substring possible
* use as many words that alliterate with other words in the sentence as possible
* make a sentence that scans with iambic meter
* make a sentence that rhymes with some other sentence in the pool
...

_morphosyntactic_:
* use an equal number of prefixes as suffixes
* use a word that contains as many morphemes as possible
* use a word that shows a sandhi effect at morpheme juncture (this one
would require an appropriate sound change card)
* use as many cases as possible
* use as many noun classes as possible
* use all the possible persons of pronouns or verb agreement
* use an irrealis verb or modal
* use an embedded question
* use a ditransitive verb
* make a completely left-branching sentence of at least four words
* make a sentence whose syntax tree exactly matches another syntax
tree in the pool
...

_semantic_:
* make a sentence whose subject is "My father's dentist"
* make a sentence about cooking
* make a sentence which explains how the clown lost his keys
* make a sentence which explains why you deserve to win this game of
Glossotechnia
* make a sentence with a seasonal reference (per the kigo constraint in haiku)
* make a sentence which displays semantic parallelism (per the poetic
form in the Hebrew OT)
* make a sentence which translates a well-known song title
* make a sentence which snarkily rebuts a previous contribution to the pool
...
I imagine that the judging on whether given sentences fulfill certain
semantic constraints might have to be put to popular vote if they're
borderline cases, or (especially) if they're scalar. Also, complete
goal sentences can be regarded as very specific special cases of
semantic goals.

Maybe among the say five goals each player creates, there has to be at
least one phonological, at least one morphosyntactic, and at least one
semantic.

Each player would share the goals they created. The other players
would somehow vote on them to assign them ratings, corresponding to
the difficulty of achieving them. The ratings should probably be from
one point for easy goals to maybe three or four points for hard ones.

Then the players would play Glossotechnia as usual, except that they'd
be trying to put sentences in the pool which fulfill the goals. All
players were subject to all of the goals. At the end of the game,
which should probably either be when external circumstances force an
end or after a prescribed time limit (turn limit?), the pool is
scored.
For every boolean constraint, every player who has a sentence in the
pool satisfying it gains its difficulty rating in points.
For every scalar constrains, every player who has a sentence in the
pool maximising it gains its difficulty rating in points. (i.e. if
there's a tie, all players get it.)
Each constraint can be scored by each player at most once.
To encourage cleverness, i.e. achieving multiple goals in the same
sentence, a player could get a bonus point for each goal beyond the
first that it scores for, or something like that.

The player(s) with the most points after this finishes have won, of course.

With the multitude of different goals, this should do a fair bit to
prevent the problem associated to the fact that racing to be the first
to translate your challenge sentence and having nothing else in mind
leads to a rather degenerate game. And it makes every mechanic of the
game potentially pointful -- the game can be won or lost on a single
sound change, or semantic change, or what have you.

That said, a play of this variant probably takes even longer to
complete than a good multi-sentence game of Glossotechnia as it is
now, and we've never completed one of those. As I briefly mentioned
above, it's not really friendly to the linguistically-uninformed as it
stands here. And, of course, anything like this will need extensive
playtesting. Grist for the mill.

Alex

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Tue, Aug 12, 2008 at 10:14 AM
To: Alex Fink <000...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

<snip long explanation>
Another idea I thought of was to have each
subject and each predicate card have six
phrases on it, of which one would select a
particular one with a die roll (or by drawing
a card from the main deck and looking at
the number 1-6 in the lower right corner
or whatever).
*lîfyŋ-bâm-ŋa gǒ.
[Wow, what a cool new winner-algorithm!]

I like this a lot. Some of the earlier playtesters suggested trying
to add incentives to use the game-language more than just for
translating the challenge sentences; this achieves that goal
in spades.
Or maybe you only get points for satisfying a particular constraint
once, no matter how many sentences you've written that satisfy
said constraint?
I don't think there should be a penalty for
composing too many sentences in the game
language, even if some of the sentences are
less original than others. The more you say/write
in the language, the better handle you and
other players get on how it works, even if
some of the sentences produced don't
directly help you win.
I reckon new players would be better off starting with the
older form of the game, with fixed challenge sentences
or mix-and-match subjects and predicates, and graduating
to this form after getting familiar with the mechanics of the game.
In this form of the game, what is there that corresponds to
the group challenge sentence in the older form?
What about:

:Sound change
:Sandhi
:Specify a rule for how specified sequences
of phonemes formed by affixation will change.

That seems a bit too vague... needs work.
That's a big plus.
Variations on this idea -- each player contributing fewer than
five constraints at the beginning -- could lead to shorter games
than the five-constraints-from-each-player version you outline.
A version with each player contributing only one or two constraints
might not take any longer than the existing version; we'd have
to try and see.

And maybe new players / non-linguistically sophisticated players
could use a version of this ruleset, simplified by (say) omitting
phonological constraints and having simpler, all-boolean
grammatical and semantic constraints...? E.g., constraints
suitable for a basic game might be like:

* ask a yes-no question
* form a sentence using at least three adjectives
* say something about another player
* say something about the host's cat

Or you could include phonological constraints, but probably
keep them on the prosodic/poetic level rather than the phonemic
level -- e.g., must fit a certain meter, must rhyme or alliterate.
Whatever all the players feel comfortable with.
(Maybe, with a group of players of varying levels of
linguistic expertise, you could let the less experienced
players veto one or more constraints proposed by more
experienced players.)

On another subject -- Alex, I don't think you were in the cc: list
yet when I wrote this; I'd like to get your feedback on it as well:

> In the "Art is when someone says 'Now' -- or is it?" thread, I wrote:

>> This reminds me that a friend once asked me if I thought
>> we could adapt Glossotechnia in such as way as to combine
>> it with a role-playing game. We still haven't figured out a way
>> to do it yet. The languages created by the players in Glossotechnia
>> games tend to lack concultural context; maybe it would be
>> interesting to add optional cards and rules covering
>> the language/culture interaction....? And then the players could,
>> in addition to working on coining words to translate their
>> challenge sentences, also be roleplaying as persons living
>> in the as-yet-underspecified conculture that speaks the
>> language being created by the game?

> A little later I thought of another possibility: on each
> turn, in addition to drawing a card, playing a card,
> and coining a word, a player can make some statement
> about the culture or context of the language -- anything
> goes as long as it doesn't contradict what other players
> have said about it so far. And then in the deck there
> might be Culture Change cards that let you change
> aspects of the culture already defined, preferably telling
> the other players a story about how the change happens...?
> I might try that out the next time I test the simplified
> deck with gamer friends. Do y'all have any suggestions
> for specific subtypes of Culture Change cards or
> other ways to guide collaborative conculture development
> with card-game mechanics?

That could tie in with the constraint-based version of
Glossotechnia: some of the semantic constraints could have to
do with the culture of the people speaking the game-language.
E.g., constraints like

* make a sentence using two or more kinship terms
* say something about how weird foreigners are
* ask a question about how/why a certain custom originated

etc....? And players get bonus points if the information
about the conculture they contribute on their turn is stated partly
or wholly in the game-language, over and above whether
such phrases or sentences satisfy one of the overall
constraints for that particular game.

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Aug 12, 2008 at 4:53 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Alex Fink <000...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

Just to briefly explain the parts of my proposal that I think Alex's
recap didn't quite get (I don't have the time now to go over any of
the other replies since then):

This is a variant of my proposal #5 from my last email.

1. Setup is nonexistent. You just start playing. There is no fixed
target sentence for anyone, EXCEPT that there is a group sentence per
standard rules.

2. Every turn, if you can make a sentence that follows all extant
rules, including respecting as canon all previously created sentences
[modulo explicit changes like phoneme or semantic shifts played using
cards], you get points. All sentences are written down w/ interlinear.

- Any new word or grammar rule should go through charade phase,
wherein the group determines what its meaning is. Majority rules,
person to left of charader breaks ties.

3. Points are awarded as follows:

* BASE VALUE: "newness", i.e. minimum difference from any previously
existing sentence (a.k.a., you can make trivial modifications of
others' work, but it's worth little to do so)
* bonus points:
- fulfilling any of a fixed set of constraints; points awarded are
directly based on how hard it is to do ingame, or how hard it is to do
a priori - e.g. like Scrabble awards lots of points for using Q and X
- accuracy and speed with which people guess your charade
- fulfilling meta goals, e.g. being the first to make a sentence
that's left-branching to depth X, or the first to translate any
specific verse of the Babel Text
- eliciting a group laugh / applause or the like
- translating the current group target,which should be either a)
fairly difficult to reach, or b) part of a series of targets

Fulfilling zero constraints is valid, but worth very little.
Fulfilling multiple constraints simultaneously, especially when they
are difficult to do together, awards extra bonus points.

Option 1: Group target is the Babel Text, and points will only be
awarded for completing some portion thereof. (Sentences can of course
be broken down in non-English fashion, and possibly distorting the
story significantly to fit sociolinguistics.)

Option 2: Sentences created MUST meet a minimum value to be played.

Option 3: Add [school] constraints, such that the language is pushed
towards being [school]istic. School here ca be, e.g. "naturalism"
(where such a constraint might be e.g. negative points for having only
one click, or positive points for vowel harmony, etc), or "alien"
(where it'd be the reverse), or something more complex like "logical".
School is fixed at the start of the game, and can be decided by group
consensus - e.g. say that the language of this game is going to be
"spoken by humanoids who have no voicebox, are highly technologically
advanced, and have a very strong cultural belief in time being
circular". +1 point for having consonance, -1 for having dissonance, 0
for sentences that are enh with respect to this constraint.

Option 4: Charades must meet additional constraints, e.g. no alluding
to English / other shared languages; no drawing; no gesturing (= audio
charades, though you're not allowed to use any extant language and
thus have to either babble or use the glossolang); etc

4. Play continues in this fashion until some end is reached - which
could be e.g. one player gets X points, or the group target is reached
(a la Mornington Crescent), the deck is exhausted, or whatnot

Benefits:
* much faster gameplay; more stuff happens, more granular scoring
(i.e. not all-or-none)
* no need for extra target cards, except possibly to specify the group
target- though for this, I think writing target sentences at the
bottom of every single card would work very well, such that the first
card drawn has on its bottom the group target (& how many points it's
worth & a citation)
* no fixed target and thus no dogged pursuit of a (boring with
repetition) goal; goal will morph with each turn, requiring you to
think more dynamically to maximize points
* lower bar for newbies, in that there's always going to be at least
something that scores (if not very high), a la Scrabble's three letter
words or adding -s
* no long, boring setup phase
* you come away with a more fleshed out language
* cooperation AND competition are both inherently rewarded
* if you flub your charade, you can still think quickly in order to
incorporate the group consensus of its meaning (which, per our
previous mod, is final decider thereof) into your created sentence

Cons:
* may take excessive time to determine point values; would need to be
done in a way that is extremely fast to look up (i.e. as fast as it is
to calculate a Scrabble word score)

- Sai

----------
From: Alex Fink <000...@gmail.com>
Date: Tue, Aug 12, 2008 at 4:57 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

Note that I started composing this before Sai's last message, which
I've not yet read.

2008/8/12 Jim Henry <jimhen...@gmail.com>:
Ah, yes, that's quite good for compactness. With this you could
probably get by for a good while with only ten or so of each sort of
card, and that's not much physical bloat at all.
I thought I said that several paragraphs later down. In any case, I
meant that to be the case.
Good point. What I mostly meant to rule out by that was the scenario
where, as soon as one player achieves one of the goals, everyone else
gets in on it by immediately submitting (degenerately) the same
sentence, or (less so) the same sentences with one of the
non-goal-critical words changed. Maybe there's a better way to
prevent this; or maybe it's not a problem.
Nothing, I suppose, as I presented it. Or, maybe, it's the individual
challenge sentences that nothing corresponds to, given that in this
form everyone shares all the goals.

Perhaps you could remind me what the purpose(s), game mechanics-wise,
of having both group challenges and hidden individual challenges
was/were.
I didn't mean to propose this as a _new_ sound change card. Just, for
instance, if there had been an earlier ordinary sound change that
palatalised velars before front vowels and an (earlier-coined?)
front-vowel initial suffix applicable to a velar-final root, this
would be a situation in which sandhi effects would arise, and you
could use it to satisfy this goal.
I had assumed that with so few constraints, many of which are likely
to be fairly weak, achieving them all might be too easy; that said,
with a suitably chosen time limit, that might not be a problem cause
there'd probably be nearly as much groundwork to do to get to the
point where you can even make a sentence as in the standard game. So
it would be worth a try.
Either of these would probably work.

Providing an incentive for making one's cultural statements in the
game language is a natural idea for any form of Glossotechnia, I
think, quite parallel to providing an incentive for defining one's new
word or morpheme in terms of the game language. And given that we
want to encourage lots of composition of sentences in the game
language, this is probably a good way to give the right amount of
consolidation to them -- not so monotonous as "everything you say must
be about the host's cat", but with some structure. For comparison it
was quite good fun once the idea caught on that the Kalusa speakers
revered A'Tuin.

Your extension to the constraint game seems to be the natural one as well.

If it weren't for the easy tie-in that making statements has to the
game language, though, I might've be inclined to regard this as moving
towards scope bloat. It's still the case that (especially with the
original challenge-sentence rules) the cultural component and the
linguistic one are fairly weakly bound together, all the more so in
the phases during which the gamelang isn't powerful enough to say
much. Eh, maybe it's just that I find something faintly silly about
saying "okay, the culture believes that the river god demands regular
human sacrifice to prevent him from overflowing his banks and
destroying all the crops, oh, and, now you can use /f/".

Anyway, one could also use this for role-playing, or story composing,
with a more traditional focus, outside the framework of a conculture:
though in this case, to retain interest, I suppose the rule would be
that only sentences composed in the gamelang would be valid. (One
could call this the Malia and Kuana game.)

Lastly, in terms of game balance, I don't have a good idea whether
awarding a point for a sentence in the gamelang (in those variants
with points) is balanced or not.

2008/8/12 Jim Henry <jimhen...@gmail.com>:
It would work with the current mix -- that was part of the intent --
but I think a greater ratio sound and grammar changes would be more
interesting.

Related to this is that, as currently implemented, there are certain
kinds of sound and grammar change cards such that, even once you've
got them in hand, you still may have to wait to also acquire a target
phoneme or word order card. This makes these changes harder to use
than cards that don't require hangers-on, and probably makes their
frequency of occurrence disproportionately lower than their
representation on cards would predicts.

For that matter, to focus on sound changes, I'd say that those which
affect whole natural classes of sounds at once, such as "voiceless
sounds voice intervocalically", are a good measure more common in the
wild than those affecting single sounds, like "/f/ voices
intervocalically". But these former aren't provided for in the
current deck at all. I'd like them to be, but I don't have a good
implementation yet.

Alex

----------
From: Alex Fink <000...@gmail.com>
Date: Tue, Aug 12, 2008 at 5:31 PM
To: Sai Emrys <s...@saizai.com>
Cc: Jim Henry <jimhen...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

2008/8/12 Sai Emrys <s...@saizai.com>:
> recap didn't quite get:
Where then do the constraints and meta goals come from, that in my
game are part of the setup phase? ("Bonus points", items #1 and #3.)
Is there any substance to the restriction "respects as canon
previously created sentences"? Afaict it follows from the more
fundamental rule that all sentences must be grammatical in the game
grammar. This is not Kalusa, remember, the grammar _is_ explicit.
Okay, here I think is one of the greater divergences, from which
several of the others follow. In your game, every sentence is scored
immediately; in mine, all scoring waits till the end. Yours generates
points for every sentence, whereas mine (and I read that Jim agrees,
by his suggestion) makes each goal scorable on only once and this the
only primary source of points.

I like the aspect of mine that there can be explicitly competitive
goals, my "scalar" ones from earlier where you have to have done at
least as well as everyone else playing on a certain metric to score on
them. Yours I don't think have that.

(Scalar goals of this form are also better for maintaining interest
than "first to X" goals; those become irrelevant once someone has
them.)
A reasonable idea, except not applicable when it's not the case that
all sentences score immediately.

> * bonus points:
I don't like that. This is first and foremost a game about
linguistics, not charades. I know you were worried about the charade
mechanic being irrelevant, but this seems too far in the other
direction.
Hm, that's a nice idea for one long game, what with playing to a
conlangerly standard text. But it's not replayable.
[...]
"School" here seems really to mean "conculture", and I think Jim's
handling of this is more involving -- allow players to declare
concultural dicta on each turn, not all at once initially. (Also this
would be a setup phase, which you claim to lack.)
Seems pretty orthogonal to the substantial stuff. And there are rules
in place to discourage using English already, no?

> Benefits:
Oh, maybe, given the cool factor in translating real sentences from
real works. But I still don't think your dire concern over game size
is all that merited.
Nor any variant proposed yet, I think. Or do you mean to suggest that
thinking up target goals is long and boring?

Alex

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Aug 12, 2008 at 6:16 PM
To: Alex Fink <000...@gmail.com>

1. Rulebook
2. Optional rule: everyone adds one constraint/goal as a one-round setup
Inasmuch as previous sentences perfectly match grammar, this is indeed
redundant. But I think it's a more natural way to think of the
requirement.
Which, in turn, makes the scoring round tedious and eliminates most of
sense of the concurrent competition.

I can't think of any games that have only-at-the-end-of-the-game
scoring for which that feature of them is not a major drawback.

Each goal may be scorable once or multiple times, depending on the
kind of goal. I would suggest that some (eg metas) be only once,
It could, in that you could "steal" goals from others by making
something scalarly better than the current goal-holder. That's a
fairly simple addition if you feel that scalar is really a good thing,
and would if anything add to the immediate competitiveness of it.

It would however not be possible except at the end to scale stuff
that's inter-sentence; all scoring in my system is based on the
current sentence.
Sure.
Granted. But one could grant conlangery bonuses to charades, e.g. +1
for each time you exclusively use the glossolang in it.
Sure it is, if there is sufficient deviation in the
goals/constraints/conculture/school stuff, and minding that I
encourage sometimes dramatically sociolingly varying interpretations.
School here means both conculture OR a more Bangsian sense, really.
Anything that strongly influences / dictates the overall goals, shape,
design, opinion of the language.
It's not purely orthogonal, when some of these constraints could be
tied to the glossolang per above.
*shrug* IME professionally, people vastly underestimate the degree to
which minimal increases in difficulty of doing things decrease
someone's likelihood to do it.

In this case, I would suggest that all things being equal, people will
be more likely to carry with them at all times, and thus be able to
randomly start playing, a game that is no bigger than a double poker
deck than one that is double that size.
If it has to be explicit and takes time, and could be done within the
flow of normal rounds instead, then yes.

~

----------
From: Alex Fink <000...@gmail.com>
Date: Tue, Aug 12, 2008 at 6:54 PM
To: Sai Emrys <s...@saizai.com>
Cc: Jim Henry <jimhen...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

Full text of Sai's message subjoined, since he didn't cc it to anyone.
Okay. I think it's more interesting if they're not constant, anyhow,
which is your (2) or my variant.
Unless "natural" has a tricky meaning, like 'in accord with the way
people learn a natural language', I disagree. Yours takes a two-step
process: extant sentences => rules of grammar => new sentence. Mine
takes just the second step of this.
Only if the aspects of the gamestate being scored are somehow
non-cumulative. That's not the case with my rules.

For my boolean goals, the scoring-at-the-end is only a feature of the
way I presented the rules (which I did for explanatory simplicity).
You can, and it's natural to, view them as concurrent instead. Once
you achieve a boolean goal, those points will necessarily be yours
come the end of the game, and so you may as well account them as yours
the moment you achieve it. I imagine this is the way any sensible
player will think of it, that they have earned those points then.
Likewise, your recasting of my scalar goals as stealing goals is
precisely isomorphic to mine, and is simply what you get if you
project the current situation to the end of the game for the
accounting.
Sure.
[...]
Mm, I forgot you had a philosophy regarding being nonliteral to the
Babel text. Fine, then. Still seems odd game-design-wise to so
enshrine one text in the rules, regardless of its status in the
community, though.
True, as far as high-level goals go. Probably something which must be
agreed on beforehand whatever the rules, if the players care.
I still don't think that makes it boring. Surely playing the main
phase of Glossotechnia isn't the only non-boring activity that
exists...

Alex

----

Sai's message:

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Thu, Aug 14, 2008 at 5:37 AM
To: Alex Fink <000...@gmail.com>, Sai Emrys <s...@saizai.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

There's not that much setup in the existing version, except shuffling
the deck and dealing cards. And I think more often than not,
lately, I've used the optional rule of laying out a few random
phonemes and syllable cards before play starts. Alex's constraint
version, as I understand it, would take a little more time, but
not nearly as much time as explaining the rules to new players
takes, and probably no more time (with experienced players,
at least) than the setup phase of 1000 Blank Cards takes.
That's an improvement that should be incorporated even into the
basic form of the game. I have it in at present in the optional rules
section; I think I'll move into to the main part of the doc.
Or maybe the rulebook could suggest several short texts
that a group might work on during a long game, and give web
resources for them -- the Babel Text, the North Wind and the Sun,
and so forth...?

> Option 2: Sentences created MUST meet a minimum value to be played.

Hmm. As I said, I'd like to see maximum incentives for using
the game language. Any time a player can create a sentence
that's nontrivially different from another sentence already played,
there should be at least a slight benefit to doing so, in a point-based
game, I think. Maybe only one or two points (on a scale where
satisfying even a moderately difficult constraint scores 10-20
points) but something.
Good.
No alluding to English in the form of "sounds like" homophones,
yes. See below. The others, I'm not keen on. It seems that any
nonverbal means of getting meaning across, plus use of the
gamelang, should be allowed in any game. -- Although maybe
we could add an Action card that lets you limit another player's
charade/drawing/etc. options on their next turn, like the Action
card that lets you specify the part of speech or semantic category
of their next word.

In the current version of the rule I have this text:

===

<h3>A note on the use of the word "charade"</h3>

<p>
In a traditional game of charades, one is trying to get one's fellow
players to guess at an English word or phrase, or perhaps a proper
name. One strategy is to break it up into syllables or other
sub-parts which are, or sound like, other words, and do successive
charades for each syllable. However, this strategy doesn't fit very
well with the spirit of Glossotechnia; here, you are trying to get
your fellow players to figure out the <em>meaning</em> of your newly
coined word, not the English word or words that could be used to
translate it. That is the whole point of using pictures, charades and
so forth instead of simply defining your words in English.

===

Can you suggest any way to improve it? Maybe it should give
specific examples of charades based on English homonyms
that are against the spirit of the game. And maybe there should
be specific penalties (discard a random card, reducing the size
of your hand; or lose points; or both) for violating this; but it
probably shouldn't be strictly enforced against new players.
Probably not playing till the deck is exhausted; there are action
cards that can cause the discard pile to be reshuffled into the deck.
Maybe play till the Nth group challenge is translated, N=1 to 3
or so based on how much time is available.
That's probably good.
I'm not sure how you mean "goal will morph each turn".
Do you mean someone might aim at satisfying one constraint
for a while, and then pick another constraint to try to satisfy
once they've come up with a good sentence satisfying the
first constraint, etc...? Or that, with sound change and
grammar change cards being played, the attainability
of certain phonological and morphosyntactic goals
could change from turn to turn, so someone might
(temporarily) give up on a goal they were working toward
and pick another that looks more achievable at
the moment?
In contrast to what other version of the game?
Yes, any rule change that gives incentives to compose more
text in the language would help with that.
Good.
> to calculate a Scrabble word score)

Yes, which is why I suggested, for beginners, using clear and boolean
constraints so that the players won't have to debate and vote on whether
each sentence satisfies them or how well.
Hmm. I see. Maybe you get points for sentences based
partly on how well they satisfy constraints, and partly on
how much they differ from other sentences made earlier?
So if one player satisfies a rhyme/meter constraint with
a certain sentence, and another player contributes
a sentence that's identical but for one nonrhyme word
changed to another of the same length, they would get
a minimum number of points for it. Or even zero.
But not, I think, an actual negative penalty.
Yes, in your version everyone shares all the goals set by
each player; but they're all working competitively to satisfy
those goals with their own sentences.
The point was to make the game more cooperative as well
as competitive; you can't necessarily win just by focusing
on your own challenge, because the group challenge has
to be translated first. There would be some element of
cooperation in any form of G'technia; whatever contributions
you make to the language make it more powerful and easier
for other players to express things in. But it seems there
should be more cooperative elements than just that.

-- Although, degenerately, you could try to win with a strategy
of free-riding off other players' work on the words and
grammar of the group challenge until you've got all the
lexical and grammatical material you need for your own;
then pitch in to help with the group sentence if it still
needs work. I'm not sure if anyone's ever done that;
I've usually randomly switched from my own sentence
to the group challenge and back every turn or three.
OK, good. It's still possible we need more sound change cards,
and more specific ones, to make this sort of thing happen more
often.
......
Hmmm, maybe so. But statements made in English about the
conculture early on can provide context for statements made
in the gamelang later on, and maybe context for interpreting
charades to codify the meaning of new words vis-a-vis the
conculture.
What about this: all sentences in the game language that aren't
essentially identical to earlier sentences earn points, based on
how different they are from earlier sentences. But sentences that
satisfy constraints earn points on a much larger scale. E.g.,
a random 5-word sentence that differs by 3 words from the most
similar previous sentence gets 3 points, while a sentence that
satisfies any constraint of even moderate difficulty would get
something in the range of 10-50 points (depending on the difficulty
of the constraint, etc.). -- And maybe the originality points
are multiplied by the constraint-satisfaction points, so of two
sentences that satisfy the same constraint, but are very similar,
the earlier one would get far more points because it was
more original. A sentence that just tweaks a previous
sentence slightly gets only 1 point if it satisfies no constraints;
if it satisfies a constraint, it gets 1 x the number of points
the constraint is worth, whereas a more original sentence
might get 5 or 10 x the value of the constraint.

.....But maybe that's getting into too complex a scoring rule.

> 2008/8/12 Jim Henry <jimhen...@gmail.com>:
Maybe there should be an option for using the Phonemic
Contrast cards that way? Either in and of themselves, at any
time, or in combination with certain sound change cards?
Probably the former. And maybe there need to be more
Phonemic Contrast cards in the deck relative to Phoneme
cards etc.

So, for instance, when playing a Voicing card, you wouldn't
just add voiced/unvoiced variants of all the fricatives in play,
but would specify a contextual voicing sound change affecting
fricatives in existing words.
In addition to the explicit grammar specified by the Typology and
Syntax cards in play and the affixes or mutations in the lexicon, there
is liable to be implicit grammar implied by the way players have
composed their sentences up to now. What would be a good way to
say this in the rules, preferably with concrete examples?
I'm not sure why it wouldn't work with endgame-scoring, as long
as there's a clear record of what order sentences were made in.
I agree.
Though I don't see why you couldn't have both -- some of the
initial constraints contributed by players during setup could
involve the conculture, and then further specification of the
conculture would go on throughout the game.

But there needs to be a way to resolve outright conflict
or contradiction between the constraints set during the
setup phase. With most of the kinds of constraints Alex
mentioned that's not a problem; if they're contradictory,
it just means a single sentence can't satisfy both, but
any player could satisfy both with different sentences.
But re: conculture or school-of-conlanging, there could
be contradictions if one player says "we're making
an unambiguous engelang" and another player says
"we're making a naturalistic artlang spoken by
tundra-dwellers". Maybe pick a random player
to start with, then go round in a circle offering
constraints, anything that doesn't *globally*
contradict an earlier constraint, until everyone has
contributed N constraints. -- And as each player makes
a constraint, the other players decide how many points
it's worth?
Hmm.... Sai, do you mean a fixed set of constraints would be in the
rulebook? Or that the rulebook offers a large set of potential
constraints or ideas for kinds of constraints and the players go
through and pick a subset they like? Or a subset is picked for
them randomly by die rolls or drawing cards with numbers
discreetly printed in the corners...? A combination of that plus
Alex's version might be good, starting out with M constraints
randomly chosen from a list in the rulebook plus N * P
constraints where each of P players makes up N constraints.
But unless the rulebook has a fixed set of constraints to
be used in every game -- which seems like a bad idea except
maybe as a starter set for beginners -- there's still going
to be a setup phase, and probably not much shorter when
picking a subset of constraints from a rulebook list than
when players make up their own from scratch.
That makes sense.
OK, so another variant perhaps.... there are no constraints
at first, or perhaps only a few (one randomly picked from the
rulebook, one contributed by each player). More constraints
are added during play, perhaps whenever certain Action
cards come up, or perhaps on each turn in the first N rounds....?

Here's another variant that makes more use of card game
mechanics: possibly replacing the current Action cards
that let you swap or rotate challenge sentence cards,
there could be Action cards that let you add, delete,
alter or replace a constraint during the course of
the game.
Something that satisfies the constraint better but is also original
enough; yes.
What do you mean by "inter-sentence"? Points for originality
re: some metric of difference from previous sentences?
So constraints that specify a location within the Gnoli Triangle,
or say we're trying to make the sound reminiscent of a specified
natlang, or specify a semantic field the speakers consider
important and will make lots of fine distinctions in...?
How would "no drawing" or "no gesturing" or maybe "no pointing
at examples of NOUN in the environment" tie into the glossolang?
That's probably true. Do you have any ideas for how to reduce the size
of the existing deck? The only things I can think of are:

* eliminate the C onset and V rime cards, and say you can use CV
syllables any time (unless some phonotactic constraint card forbids it?)

* reduce the number of Phoneme cards and compensate by increasing
(but not by so many) the number of Phonemic Contrast Cards.
E.g., we could probably drop all the front rounded and back unrounded
vowel cards and add an additional instance or two of the Rounding
contrast card, or eliminate all the voiced plosives and fricatives
and unvoiced nasals and add another instance or three of the
Voicing contrast card.

* Radically change the phonology portion of the deck so there are
no Phoneme cards, only Point of Articulation and Manner of Articulation
and Phonemic Contrast cards. That could create schematically
boring phoneme inventories unless there were a few more cards
to restrict things in interesting ways, though, and would be a major
enough change that it would require far more additional playtesting
than the other options.

Here are a couple of other points I don't think I've brought
up yet in this dicussion.

1. Have I described the simplified version of the game yet, the
one I've only playtested once (with younger cousins at this summer's
family reunion)? It eliminates the phoneme and syllable cards
entirely; players are entirely unconstrainted in the phonology
and phonotactics. The Sound Change cards are altered
so they don't refer to Phoneme or Syllable cards in play, but
instead say things like "Pick a sound that occurs in one
or more existing words, and specify another sound it will
change into". Syntax, Typology, Action and Grammar Change
cards are pretty much the same as in the advanced game,
except I've tried in some areas to avoid linguistic terminology
that the cards in the advanced game use.

2. A while ago I threw out this idea for feedback and didn't
get much; I haven't tried implementing it yet. What do y'all
think? Sai will probably say it increases the size of the game
far too much, and I would probably agree, but sometime or
other I'd like to try this at least once:

On Tue, Feb 12, 2008 at 12:31 PM, Jim Henry <jimhen...@gmail.com> wrote:
> An idea has occurred to me for a possible variation or refinement of
> Glossotechnia.
> In addition to the main deck and the translation challenge deck, I
> might add a deck
> containing the 60-odd words of the Natural Semantic Metalanguage. Each turn
> a player would coin one word on their own initiative, using charades,
> drawings, etc
> to define it, and then would draw the top card from the NSM deck and coin a word
> for that concept as well.
>
> I'm not sure if I would want to include the entire NSM vocabulary, or
> just a subset
> of those words that are harder than average to define with charades
> and pictures.
> Or I might add those cards to the main deck; when you get one of those cards
> in your hand, you can play it to coin a word for that concept without having to
> go through charades etc.
>
> With a separate NSM deck from which a word is drawn every turn, the
> vocabulary of
> the language would grow about twice as fast, and players might sooner be able
> to start using words of the game-language to define their new words.

----------
From: Alex Fink <000...@gmail.com>
Date: Fri, Aug 15, 2008 at 2:55 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

2008/8/14 Jim Henry <jimhen...@gmail.com>:
That we could do. Any text of the right length would serve just as
well, of course, and I'd say as much; if all your players are Shelley
fans, why then go translate Ozymandias. This would be a list of
suggestions, of texts which have come to significance in linguistics
or conlanging.
Yeah. See the scoring discussion, below.
[...]
That would help to drive the point home, sure.
That's a little too punitive for my tastes. Something like that might
prove necessary if people otherwise wilfully ignored the rule; but I'd
prefer just to leave it at against the rules, and declare charades in
which it happens void.

(Given the adoption of the "majority rules" resolution, if I were in a
game with someone who kept doing this I might, in a playful spirit,
just be obtuse and try to force misinterpretations -- see "three
syllables, first syllable 'pot'", stir up support for "pot".
Vigilante justice?)

We'd also want to try to make sure players don't make use of English
homonymy and (less egregiously) slippages within semantic space
peculiar to English down the line. You know, if a word is charaded
and guessed perfectly properly as meaning 'bank of a river' but
recorded just as "bank" and then someone uses it in a sentence for
'financial institution', that's no good.

> The other [restrictions on charades], I'm not keen on. It seems that any
One advantage of the charades mechanic is that it's modular. If
people decide they want to play Glossotechnia entirely drawn -- say,
because they're playing across the internet and have a site which
supports real-time drawing (like iSketch etc.) but not video or
anything else -- they should be able to. This, to me, speaks against
trying to hook into the details of the charades mechanic in any ways
other than general high-level ones.
Right, good.
Yeah, this is probably something best addressed in the scoring scheme.
Right. That's the principal reason I included scalar goals. For
common goals, one might introduce bonus points for achieving them
first.
It is the obvious degenerate strategy. I haven't noticed it
happening, but I think the possibility of it is quite apparent.
What's to be inferred is probably that I haven't played in any game
with anyone else who valued winning much higher than they valued
making an interesting gamelang.
Perhaps. It does rely on the precipitating sound change being
conditioned, and of the existing cards only Phoneme split, Eliminate
cluster, Syncope, and Metathesis are explicitly described as being
able to achieve this. That's five out of thirteen.
Right, my objection about incohesion was mostly a hypothetical one.
It is a solid idea.
So, you're proposing this?

Say that the base value of a sentence (per Sai) is a measure of
distinctness from everything that's come before, perhaps by counting
words -- what ever it is, it should be quickly computable. Then, if a
sentence achieves no (applicable) goal constraints, it scores its base
value, whereas if it does achieve an applicable goal constraint, it
scores (base value) * (difficulty rating of the constraint) * N, for
some suitable constant N.

I'm not sure whether that's too complex, but I hope it's not. I think
it addresses many of our concerns.

Setting N will be a delicate matter. 10 sounds the right order of
magnitude, but I don't know. (It would make the scoring
calculationally simpler if instead of invoking the N explicitly we
pull some gambit like having "small points" and "big points" (sickles
and knuts, whatever) which are related by a factor of N. Then again
if N happens to be 10 or something round this might not be necessary.)

Other considerations:
* Any goal constraint should be scorable for just once per player. So
if you later re-achieve a goal but with a more original sentence, you
should be able to score it for the constraint and adjust the score for
the earlier one down accordingly.
* To increase the competition, maybe each boolean constraint carries a
bonus for the first to achieve it.
* For concurrence we'd think of scalar constraints as stealing
constraints; that would mean some potential rescoring too.
* If we allow points for flair, like Sai's making the audience laugh,
they should just be an additive bonus.
That's a good idea.

Perhaps all the sound change cards should be able to affect phonemic
contrasts as well, analogously to how they affect phonemes. For
instance, this might let me play a chain shift, replacing a Voicing
card by an Aspiration card, and declaring that the voiced and unvoiced
series of stops are now unaspirated and aspirated, respectively.

On the other hand, you do have Eliminate contrast and Phoneme merge as
distinct cards at present... (And do you mean to allow Eliminate
contrast to eliminate a contrast in any fashion -- not just by
dropping it, but for instance by merging all labialised phones into
existing labial phones?)
[...]
Hm, true. So we should probably encourage (mandate?) players to
follow implicit trends in the example sentences.

Should we require it? There are plenty of quirks in natlangs on this
scale: to give one example there's English preposition choice ("on the
train / bus" but "in the car", etc.). Perhaps if someone points out a
deviation of a sentence you've just made from an existing sentence in
one of these unregulated points, you'd have the choice of fixing it to
conform or justifying the difference. But maybe that's too much fuss.
Whoops, I was confused, that certainly works. What I meant to say is
it (manifestly) doesn't work if making an ordinary undistinguished
sentence doesn't score you points.
Yes, like that. Something to give a foundation, set a direction, for
the language to be crafted.
Certainly you could.
Right, you're discussing two different mechanics here. For my sort of
constraints -- which I usually managed to call "goals" --
contradiction is no problem.

As for school-level constraints, I think leaving it fairly open and
unregulated how the players decide on these is best, just to avoid
e.g. a player who really doesn't like engelangs being dragged into an
engelang game, or to allow players who just want to wing it to do so.

And I didn't think school-level constraints were the sort of object
that were worth points.
[...]
If we want to add the constraints (school-level or goal constraints)
during the game, then one each turn for the first N rounds is probably
the best way to go. Since these form (one of) the ultimate
objective(s) / determine the fundamental flavour of the game, it seems
unwise to have them first specified later than the very early game.

I still don't see the advantage to this proposal, though.
Right. It's unspecified how to apply those Action cards in the
goal-based variant anyhow. Since AFAICT at this point we envision two
or three variants -- one with challenge sentences, one with goal
constraints, and perhaps one with both -- probably the best thing to
do is to make those Action cards operate on challenge sentences or
goal constraints, depending on which your game contains, or your
choice of either if it contains both. (This would be for
goal-constraints, not school-level constraints, which should be set in
stone for the whole game.)

>> Sai's message:
[...]
You've done this, yes?
This would help with size, certainly. But you're right that the
second needs more thought to prevent the schematically boring aspect
of it.

One should probably review the phonology cards at some point, anyway,
whether or not we do this. A few thoughts as I was first looking
through the long deck page:
* I had forgotten there was no /G/; it's common enough it should be
added, I'd say (if we stick with separate voiced/unvoiced cards).
* In terms of natlang frequency it makes much more sense to have two
/a/s than two /&/s. (I suppose this had something to do with
English?)
* There are two click cards? That over-represents clicks.
Conversely, for the languages that do use them, you're bound to
under-represent the click systems with any reasonable number of cards.
I'd drop these and make "Clicks" a card like the phonemic contrast
cards. Perhaps:
"While in play, all clicks whose release is the same as some extant
stop forward of velar, and whose accompaniment is the same as some
extant velar or uvular phoneme, are available."
(or the same better worded)
I think I saw that on your webpage, even if you didn't describe it
here or on-list.

So, how did it go? Did the younger cousins and everyone catch on?
How much linguistic explanation was necessary?
Modulo clutter in design, you could always cram those onto the main
deck as well...

Or, and this assumes you're putting challenge sentences on the cards
and imposes a constraint on their choice, you could emphasise one word
(morpheme) in each challenge sentence, to serve at the card's NSM
word. Bonus, this provides an example context of usage.
Or an exponent of some nature. In the more synthetic typologies many
of these would be inflection or derivation instead of words.
Right, some number of these ('above', 'below', 'big', 'small', ...)
seem easy to charade. Conversely it would make a lot of sense to move
beyond the NSM in some places where their hypothesis has led them to
be parsimonious (e.g. they've got 'can' but no other modals, alethic
or otherwise).
Or, if you consolidate the decks, include an Action card: "(Instead
of|In addition to) defining a word through charades, look at (some
constant number) of cards from the deck and coin the NSM words on
them".
Yes, that it would quite likely achieve, I think.

Alex

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Fri, Aug 15, 2008 at 6:05 PM
To: Alex Fink <000...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>, Emerson Knapp <enze...@gmail.com>

I've added Carrie Schutrick and Emerson Knapp, who've playtested
Glossotechnia and said they were interested in the rules change
discussion, to the cc: list.
Yes. That's a good example; I'll probably incorporate it into the
next rules doc edit.

I think people have done that kind of thing two or three times in
the games I've played, but not as often as I might have feared.
And perhaps people who are interested in winning more than
in making an interesting gamelang are unlikely to be interested
in Glossotechnia in the first place...?
So what could we add to make sandhi more common?

[Scoring:]
Yes, that's what I had in mind. Say you have two sentences with
no words in common; then the newer sentence gets as many
points as it has words. Say the two sentences share N words;
then the newer sentence gets its length minus the number of
words it shares with the most similar previous sentence as
its point value. -- But that's imperfect because a sentence
with the same words in completely different order should maybe get
more points than a sentence that plugs in a couple of different
words while having the same overall structure as an earlier
sentence... E.g. if "Dog bites man" is already in the corpus,
should "Man bites dog" or "Dog bites child" get more
points?
We could try this in a few playtesting games and see how it goes.
It seems to be the best we've come up with so far.
Why only boolean constraints?
All that makes sense, I think.
That's a good way to use a chain shift. How to say that in a reasonably
short text that will fit on the card without too small a typeface, though?
It's already one of the most verbose cards in the deck.
I reckon I could allow the Eliminate Contrast card to be used
that way as well; the text on the card might get too long and
small-print and hard to read though. -- Maybe these alternate
uses of those cards should be covered in the rulebook rather
than in the text of the cards themselves?
Probably encourage rather than mandate. On the scale of
the size of gamelangs created by a typical-length game,
almost any variation can be justified as an irregularity.
That makes sense.
Right; at least I don't see how to make them worth points.
That might work. Again, though, we'd have to be careful that the
longish, conditional text on the cards doesn't get too small-print
and confusing.

What about:

New challenge <> New constraint
Swap challenges <> Drop constraint
Rotate challenges <> Replace constraint
Yes. It didn't help as much as I thought; when I was going through
the advanced deck to write up the detailed description document,
I found I had far fewer C and V cards than I thought. The deck
is still larger than ideal after this; and there are onset and
rime combinations that are common enough in natlangs but
aren't represented in the deck, even so. (PP- and -VPP for
instance, as in Greek "ptera" and English "snapped".)
.....
.....
I thought I had two /a/s or two /A/s -- I must have been thinking
of another deck. I wouldn't have added a second /&/ if I realized
I only had one /a/ and one /A/.
Yes, that's a good idea but the Click contrast card needs more work.
The simplest would be "while in play, click versions of all plosives
are available"; it's inaccurate in that clicks aren't possible at all
plosive points of articulation, though.
It went about as well as the first couple of games in March and
July 2007 with the original alpha-test deck. Eliminating the phoneme
and syllable cards didn't make it as much easier as I thought;
the syntax cards and sound change cards still required a fair
bit of explanation. We played with everyone's challenge cards
face-up, so the more experienced players could help out
the less experienced players with suggestions. And I think
one player came up with a novel use for the Syncope card
(I'm not sure what I was thinking; before I use that deck
again it will be retitled "delete sound" or something, and ditto
Metathesis > "swap sounds", "Epenthesis" > "insert sound")
to form an inflectional mutation -- a certain consonant
dropped out of words to mark excited attitude. None of the
players were as young as the small children who played
(with the same deck we used at LCC2!) at the 2007
family reunion; I think the age range, not counting
my brother and I, was about 15-25.
That's clever. We'd have to be careful to have roughly equal
numbers of instances of each NSM word, though.
Yes, that's clearer.
Right.
OK.

On the amount of text we can safely try squeezing onto a card:
I just pulled out my Chrononauts deck
( http://www.wunderland.com/LooneyLabs/Chrononauts/ )
and looked at some of the text-densest cards, the Mission cards
and Identity cards. They have around 40-60 words of text
on them, but with different formatting I think you could fit
almost 100 words on cards of that size and still have room
for border and caption. Some of the most verbose cards
in the current Glossotechnia deck have around 40-50 words;
so I don't guess we have to worry too much about making
some of them slightly more verbose and still having room
for challenge sentences / NSM words at the bottom.

The Chrononauts deck has 136 cards, and it's about
as big as you would want to carry around routinely. The index
cards I've been using for my handmade decks are thinner;
my 170+ card advanced deck is less thick than the Chrononauts
deck although bulkier because the cards are longer and wider.
We definitely need to figure out a good way to reduce the size
of the deck.

Looking at my advanced deck description, eliminating
clicks and voiced plosives and fricatives would gain
9 cards; eliminating front rounded and back unrounded
vowels and the extra instance of /&/ gets us 5 more.
If we add /q/ and a Click contrast card and
an extra Rounding contrast card, that's a net reduction
of only 11 cards. Eliminating some or all the extra
instances of common vowels and consonants like
/i/, /u/, /s/, /t/, .... could reduce it by up to 13 cards more;
but I like the idea of having those phonemes most
common in the world's languages occur more than
once. Eliminating the extra instances of common
onset or rime patterns would only reduce it by
3 more cards. Eliminating the extra instances
of the more common word orders (SOV, SVO, VSO)
would only get 3 more, and might seriously
throw off the balance of Syntax cards vs. other card
types as well.

Also, I think there ought to be two Postpositional
cards as well as two Prepositional.

--
Jim Henry

----------
From: Alex Fink <000...@gmail.com>
Date: Fri, Aug 15, 2008 at 9:03 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>, Emerson Knapp <enze...@gmail.com>

2008/8/15 Jim Henry <jimhen...@gmail.com>:
Well, depends how we pitch it, I'd think, whether to conlangers as a
fun diversion, or to gamers as something which merely has language as
its mechanic. In the latter group I could see that ordering of
interests.

(Sai probably has something to say about this.)
Ah, a category of sound change which we lack is redistribution, where
a phone (or class thereof) falls in with another one (or class) but
only conditionally, so that the inventory doesn't change. E.g.
intervocalic voicing, where voiced sounds already exist
intervocalically; or the pen-pin merger before nasals. This is a
pretty common kind of change: if we weren't tightening the belts, I'd
say add two.

Maybe also a special Coalescence card: change a particular sequence,
or class of sequence (per Metathesis), to some single phone (e.g. /t?/
to /t_>/, or /aj/ to /E/). (Again, a change like stop-plus-/?/ to
ejective is more natural on all voiceless stops than on /t/ alone.
All good, though, if the target is the Ejective contrast card.)

On other sound change matters, Syncope should probably just be called
Deletion; "syncope" is the special case of an unstressed vowel
dropping.

> [Scoring:]
No upper bound? All else being equal I'd think base value should have
an upper bound, for balance's sake.

If my score is linear in sentence length I'm equally encouraged to
make a big long sentence even if I do it by having a big long
conjunction of one word and a second and a third and a fourth and a
fifth and a sixth and ... Or, similarly, if the gamelang has a
sentential conjunction "and", why not just and together all previous
sentences? This does have most of its words different from any single
other sentence.
You could get around this by not looking just as the unordered set of
words a sentence contains but considering its structure, and reckoning
the number of substitutions-in-place one needs to get from the old
sentence to the new.
Yes, this is worth testing once we have the base value etc. nailed down.

>> Other considerations:
I was thinking that every sentence satisfies a scalar constraint to
some degree, even if only to the worst possible one. To apply this to
scalars you'd need to introduce some minimum cutoff.

I meant also to add another bullet, which both Sai and I had before:
* Sentences which satisfy two or more constraints at once should have
their constraint value set higher than the sum of the constituent
constraint values, to reward cleverness.
[...]
Probably. They can be considered "advanced" uses of the cards, if we
want to justify this to ourselves.

>> Right. It's unspecified how to apply those Action cards [that act on challenge sentences] in the
Hm. Not a very satisfying pairing (and maybe a little jarring if
you're in a variant where you may do both), but then none really are.
Still, anything would be fine if we can make the card texts fit.
I wonder whether we might further reduce the number of phonotactics
cards by making them more modular. Modularity and the accordant
orthogonality would be a good thing to have anyhow, to capture the
phenomenon that e.g. a language like English that has /p/ and /pr/ and
/sp/ onsets is very likely to have /spr/. One might capture this by
separately allowing (i.e. on two cards, not that I mean to suggest
these precise cards) "preinitial" /s/ before some Cs and "medial"
liquids after obstruents so that the syllable structure becomes
(s)P(L). (And I don't know whether that terminology is standard; I've
been reading some Sino-Tibetan material that uses it.)

I'm not sure what the best implementation of this is in G'nia, but we
might hope to avoid providing for each of FF- and FP- and PF- and
hypothetical PP- and MF- and MP- and so on separately by such means.
Right. The thing about clicks is that, in I think every language that
has them except for a few Bantu ones where they're borrowed, they vary
along two dimensions, point of articulation at the front and what sort
of release is going on at the back. To take an extreme example, !Xoo
allows each of labial, dental, lateral, alveolar, and palatal clicks
coarticulated with each of [k_h k k?) g kx) N_0 N_0_<_h N ?N) q q_>
G\] and clusters [gk_h gkx) k_>q_> gq_> G\x]. This is what I was
trying to capture with the two-dimensionality of my proposal.

[simplified Glossotechnia]
Right, that's about what I would have expected.
Aha, that's a cute idea. (Something like the way some Algonquian
languages, IIRC, mark several things -- diminutive, speech of Raven,
speech of a person with a sight defect, etc. -- by processes like
"shift all alveolars to lateral".)

Anyway, I think the existence of this simplified version is mostly
orthogonal to anything we do in the redesign, provided we leave in the
option to play with challenge sentences. To the extent it wants the
card texts to be different it'll be hard to support in one deck
whatever we do, though.
Easiest addressed, I think, by making the number of instances of each
small (1 or maybe 2) and expanding beyond the NSM wordset.
Ah, that's good to know.
I was expecting you to want to put in an extra Voicing to balance out
dropping all the voiced obstruents.

Oh, btw, there's also no vowel length at present.
Likewise.
The following suggestion would probably throw the balance off even
more, but you might have only SOV, SVO, VSO cards provided for and
some sort of modifier card which reverses the position of subject and
object.
Oh yes, a question about those cards. Is there in fact,
typologically, a strong enough correlation that it makes sense to only
have cards implementing the orders "books on shelf" and "shelf on
books" when the position of the adpositional phrase wrt the noun and
the position of the adposition's object could vary independently
(giving "on shelf books" and "books shelf on" as well)? It is true
that the two you have are concordant with consistent modifier-head
order.

For that matter, you could drop Post- and Prepositional and let them
follow from head vs. modifier order cards, of course allowing a
secondary word order card to vary things up. But this might not even
save us any cards, given that we might need to add another secondary
work order or extra head vs. modifier cards to balance it / allow the
variability we lose by dropping it, and besides you lose the chance to
specify the correlations with order of S and O and V.

Alex

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Sat, Aug 16, 2008 at 5:36 AM
To: Alex Fink <000...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>, Emerson Knapp <enze...@gmail.com>

......

So what do we call said card? Conditional Phoneme Merge, or Redistribution,
or what?
.....
OK.
Yes, that makes sense. OK, maybe a sentence gets points based on the
length of the most similar previous sentence minus the number of words
it (structurally) has in common with said sentence. So if e.g. you
had a sentence 20 words long, with no particular points in common
with previous sentences, then it would get points based on the
length of the longest previous sentence...? Or perhaps there needs
to be a hard upper bound, as you say; maybe no more than 10 points
no matter how long the sentence is.
That makes sense. We need a clear explanation of it with a couple
of good examples for the rulebook.
Or maybe the bonus points are in proportion to the difficulty of the
constraint and the degree to which the sentences satisfies it? If the
degree to which a sentence satisfies a scalar constraint is very
slight, then it gets fewer bonus points for being the first to satisfy
it than it would if it satisfied it better. Say the bonus points for
first-to-satisfy are 10% of the points the sentence earns on other
criteria (originality and partly satisfying the constraint). So if
the sentence earns less than 10 points on other criteria it gets no
bonus points (unless the players want to keep track of fractional
scores).
How to quantify that? Maybe in addition to the base points for
originality * constraint-satisfaction, a sentence gets 5 points
for each constraint it satisfies beyond the first? Or maybe better,
it gets 2^N points for satisfiying N constraints at once.
So satisfying 2 constraints gets 4 bonus points, but 5 constraints
at once gets 32...

That's getting complicated again.
Hmm... maybe we want to introduce a certain wildcard element into each
of the syllable structure cards. So one might say "fricatives are
allowed in complex onsets [before/after] the phoneme(s) or phoneme
classes specified by the player playing this card", etc. That needs
to be simplified further. Or maybe we want to do some research on the
most common kinds of onset and rime clusters, reduce the fixed
onset/rime cards to that set (plus simpler rimes like VN and VC), and
add a few more syllable wildcards?
......
....

OK. That makes sense. I'm not sure how to word your card text better
without losing accuracy.

BTW, I just realized another universal that gzb violates:

>>>The Southern African Khoisan languages only utilize root-initial
clicks. Hadza, Sandawe, and several of the Bantu languages also allow
syllable-initial clicks within roots, but in no known language does a
click close a syllable or end a word.
<<< http://en.wikipedia.org/wiki/Click_consonant

gzb has word-final clicks when the pronouns are used as personal-marking
suffixes. But in gzb clicks are syllable nuclei, not onsets or codas;
that might violate a universal too.
Yes. The syntax and grammar change and action cards are about the same,
though their proportions within the deck are different; the sound
change cards all have different text from in the advanced version.
And the deck as a whole is much smaller, around 50 cards.
Oh. Yes, I meant to.
Hmm. I know it's in one of my decks; I thought it was in the advanced
deck. I'll add it.

Should it have a wildcard element like the Tone card, where the player
specifies how many vowel lengths are distinguished (e.g. 3 as in
Estonian, or 2 in most other languages; 4+ to be really evil)?
I'm not sure. That was my impression from reading Greenberg et al.
on word order universals. My impression is that some prepositional
languages allow prepositional phrases before their head nouns for
emphasis e.g. but none that I know of have that as their sole or
primary order, and vice versa re: postpositional languages. But I
haven't studied it enough to have more than an impression.

If we didn't want to enforce that, how else would we structure the
cards? Let the prep/postp. cards say whether the lang uses
prepositions or postpositions, but remain silent on whether
prep/postp. phrases follow or precede their heads, and have that
property specified by the H/M or M/H cards in play?
Yes, drop the three prep/postp. cards, add two H/M and M/H cards
and a Secondary Word Order card, and there's no net gain.

----------
From: Alex Fink <000...@gmail.com>
Date: Sat, Aug 16, 2008 at 11:58 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>, Emerson Knapp <enze...@gmail.com>

2008/8/16 Jim Henry <jimhen...@gmail.com>:
"Redistribution" seems the closest parallel to "split" and "merge",
but I've never seen the category presented that way in e.g. taxonomies
of sound changes in introductory texts, which would probably call such
a thing a conditional merger. On the other hand for parallelism I'd
expect you to insert a word like "phoneme" in there: but these changes
aren't redistributions of phonemes, rather of some of their
occurrences or perhaps some of their allophones, so "Phoneme
redistribution" doesn't quite hit home. "Allophone redistribution"?
All of these are somewhat cumbrous, unfortunately.

Also on sandhi, what are the implications of having the Fusional
typology in play, or for that matter the Agglutinative one? Sandhi
phenomena are the first step on the path from agglutinative to
fusional, as it were. Is the substance of the distinction between
these two cards that e.g. in fusional you might coin an accusative
plural, whereas in agglutinative you'd have to take case and number
separately -- or is it also meant to touch on phonological matters? I
note that in fusional you can coin a mutation -- does this mean you'd
also be allowed to coin an affix with its own sandhi ab initio?

Also on typology cards: when playing Isolating, you may eliminate any
number of existing inflections / derivations. So, when playing
certain other typology cards, might you not have the option to
grammaticalise some number of current periphrastic markings into
inflections / derivations? (Make cases out of adpositions, e.g.?)
I don't have a good solution for this at the moment. Leaving aside
the matter of length, we fundamentally want to score only that
material in new sentences which hasn't been lifted templatically from
other sentences. But given that we want to prevent this conjunction
of all previous sentences from scoring highly, the obvious mechanisms
I can think of also prevent "dog bites man" from scoring highly in the
presence of "dog is a bad thing to read inside" and "flea bites larger
flea" and "man behaves inhumanely to man". This is probably not
desirable? Maybe we need to address the granularity of the copied
matter? -- but that starts to get nigglesome again.

Whatever we do with this, I think that the particular place where
having an upper bound on base value is sensible is on sentences that
achieve goal constraints. These are already being scored on their
compliance to constraints which may by their natures be synergistic or
antagonistic or orthogonal to being long; it seems peculiar to insist
on multiplying scores on all of these by length, which is a bit like a
second constraint. That said, we do still want the value of these to
drop if they're unoriginal.
Unless I'm missing something this is subject to the same treatment:
the very first sentence played is the very first to satisfy it, if
only 10% so, and so this first sentencer gets their meager 10% bonus,
and then that's that. Maybe we could unequivocally agree that
satisfying a constraint 0% shouldn't count as "satisfaction" and
should therefore leave the bonus to be claimed. But not every scalar
constraint can even be satisfied as poorly as 0%. On the other end,
not every scalar constraint has an upper bound which it makes sense to
mark out as the 100% level. As an example of both (though a bad
example of a constraint for other reasons) you might have the scalar
constraint "uses as many words as possible" -- certainly every
sentence uses some words, but there's no upper bound.

My objections would be addressed if we impose the rule that the player
creating a sentence has to explicitly call which constraints they mean
to have it scored against.
My original conception was that this be a modifier applied to the
constraint difficulty (meeting two constraints together is probably
more difficult than meeting them separately). So, perhaps, if there
are n constraints that a given sentence satisfies of intrinsic
difficulty ratings d_1, ..., d_n, the combined difficulty rating might
be d_1 + ... + d_n + f(n) for a suitable function f, maybe as simple
as f(n) = n-1.
[...]
I note your message on the Conlang list about this.

Do you think that sample card is in fact too complicated? How so? If
it's the ability to choose the phoneme classes, we could provide a
default for players who don't want to think up their own, and if
necessary banish certain options for use to the rulebook. If it's the
complexity of the resulting syllable structure, it might help to use
something like a system of numbered slots, like the way verbs in
highly synthetic langs are sometimes presented: perhaps the original
C and V are in slots 5 and 10, and you might draw a card letting you
introduce a liquid in slot 8 (in certain environments, etc.).
It could... but three degrees of length is IME pretty rare, comparably
rare to three degrees of rounding (Swedish high front vowels, some
analyses) or of nasalisation (Google says Chinantec). So should all
of these allow it? Maybe another "advanced" rule?

>> Oh yes, a question about [Pre-/Postpositional]. Is there in fact,
If you've read Greenberg you're better read than I am on this point, anyhow.
Yeah, that's the natural thing, I think, given that adpositional
phrases can exist in other circumstances than modifying nouns ("we
danced _in the streets_", "smoking prohibited except _in designated
areas_"). Anyway I'd expect them to behave in natlangs like other
modifiers, especially whatever type of modifier is suggested by the
diachronic origin of the adposition class, if that's not hugely
ancient. (But maybe this is taking too non-innatist a position on the
typological question...?)

I'm still musing on the possibility of drastically reducing the set of
phonology cards by moving entirely to feature / contrast cards. Two
high-level ways to combat the unnatural orthogonality this engenders
that have come to mind are
* allow the player of each new feature card, if they choose, to
specify some gaps, phonemes that are producible from the array of
features on the board but which you're barred from using anyhow
* encourage sound changes to mess up the orthogonality

But there's something unsatisfying about moving away from phoneme
cards, too. If we did anything like the above against orthogonality,
we lose the property that the current phonology is easy to determine
from what cards are on the table (granted, this already isn't _quite_
true given sound changes &c). One would probably have to keep
separate track of which phonemes really exist, and this requires
separate equipment and is potentially bothersome. (And this property
that we're losing is the one that IIRC excites Matt, that the array of
cards in play is itself much like a first take at a short grammar
sketch already.) Playing without explicit phonemes could be too
abstract for many players, as well.

Alex

----------
From: Emerson Knapp <enze...@gmail.com>
Date: Sun, Aug 17, 2008 at 7:40 PM
To: Alex Fink <000...@gmail.com>
Cc: Jim Henry <jimhen...@gmail.com>, Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

This isn't really relevant to the rules discussion, but I've come up with an idea for a card back. I used the standard trading game card proportions, beginning with a Magic: The Gathering card image for a basis. The central image is the conlang flag. I think I still want to modify the "Glossotechnia" text, but I'm pretty satisfied with the rest. This JPEG version is slightly lower quality than the original, I can create a higher-resolution copy if need be. I'd like suggestions for improvement, or you can completely reject it if that's what suits you. The image can be found here if you can't see it below.

Highest regards,

Emerson Knapp

----------
From: Sai Emrys <s...@saizai.com>
Date: Sun, Aug 17, 2008 at 9:32 PM
To: Emerson Knapp <enze...@gmail.com>
Cc: Alex Fink <000...@gmail.com>, Jim Henry <jimhen...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

I haven't the memory span to reply to everything in detail (stack
overflow), but hopefully this covers all the points.

* Fixed vs random / chosen constraints

I would say that one should have a core set of constraints (aka bonus
conditions) that are encoded in the rules.

On top of that, one can suggest in the rules certain other
constraints, and players are to determine which of these (or any
'house rule' ones they care to add) to adopt before gameplay starts.

Constraints that affect the goal of the language - e.g. 'school'
target - must not be contradictory. A simple mechanism is: draw cards
to see who goes first (highest # wins, redraw breaks ties), first come
first serve, no contradicting previous players.

* NSML deck

Yeah, I dislike this for its volume and relatively low usefulness. One
could, however, add the NSML inventory into the rulebook as a
suggestion for what to make first.

This however relates to:

* Standard / known targets vs charades

I think that to the extent that players know what target(s) a player
may be aiming at, their charade becomes pointless because it has too
much context to take on its own. I.e. if you know someone is trying to
go for Babel 1, "language" won't get interpreted as "eat" except
through someone being intentionally obtuse.

I don't know how to handle this, but I feel like it should be fixed.

* Individual target sentences

One could have these, or not. I'm ambivalent.

If we go with simply writing a unique sentence at the bottom of each
card, that should be sufficient in number to last a good while without
repeat. Just draw one at the beginning of the game, keep it face down.

However, I think that this will tend to make people a bit too fixed in
mindset in trying to make that to the exclusion of coming up with
something possibly more creative.

Perhaps it could be simply worth a one-time bonus?

* Novelty value

What I intended was that the base score = novelty... and novelty meant
a combination of new words, new grammatical usage, etc.

Perhaps this could be simplified significantly into just:
1 pt - at least 1 new word OR 1 new grammar usage (including new
affixes etc - no discrimination against non-isolating langs :-P)
2 pt - at least one of both
4 pt - at least two of both

... or something like that 'cept better thought out.

Basically, something that is very simple and doesn't require counting.

* Phoneme vs POA/MOA cards

I'm ambivalent.

Phoneme pro: they're prettier and look better when played. They're
simpler to explain.
... con: There's an awful lot of them, and our choice thereof dictates
the style of language (implicitly towards naturalism). Plus we don't
(IIRC) have tone cards.

POA + MOA cards could lift this. You'd have a multiple of each, so
adding any particular sound won't be as hard. Each card could have a
slice of the IPA on it; the combination of two will isolate one. You
could have 'combo' plays, where e.g. you play 1 POA with a couple MOA
cards at once, or add-on plays, where e.g. under some restriction you
can play a POA on someone else's played cards, thus expanding it to
that.

But then you have to track what they each mean (like wild cards in
other games). And you can't easily give pronunciation guides on each.

And it's not enough... POA, MOA, voicing (a boolean card? player
fiat?), tone, ...

I don't know how to fix this.

* Pitch / Target audience(s)

From my POV, the potential audiences & their desires are:

* Conlangers - should be, essentially, a fun analogue of the real
thing to some extent; should be able to make whatever their conlang
actually is (else it'll feel constraining); should have fast-paced
expansion of the grammar/vocab/etc. Bonus points if it has some
in-jokes. Should be aux/enge/art compatible. Should result in
something (towards the end of a game) that feels like a grammar
sketch. (Perhaps use DM's outline of questions to ask for ideas here?)

* Gamer geeks / Party game - can be analogous to other language-based
games, e.g. Scrabble, Apples to Apples, etc. No doubt
http://www.boardgamegeek.com/browser.php?itemtype=game&sortby=rank&categoryid=2
contains at least some similarities. Needs to have easily understood
rules for non-linguists, a clearish push (i.e. so you don't sit
thinking "what am I supposed to *do*?"), some good scoring mechanism,
etc. Should be able to handle 3-7 players. Should be attractively
designed. Should have high replay value.

* Teachers - Should be a way to teach students linguistics concepts in
a fun way, or [creative writing class] teach them how to just play
with language in an innovative way. Should be atheoretical, i.e.
compatible with whatever pro/anti-Chomskian/whateverian creed they
teach. Should allow for small groups to mostly take care of themselves
for half an hour with the teacher wandering in and out to help. Should
have concrete applicability to teaching specific concepts (so it can
be justified to the higher-ups).

- Sai

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Mon, Aug 18, 2008 at 12:25 PM
To: Emerson Knapp <enze...@gmail.com>
Cc: Alex Fink <000...@gmail.com>, Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

It looks pretty good to me; maybe Paul or Matthew will have detailed
suggestions about the art.

On the website I describe it as "the language invention card game";
I'm open to changing the epithet depending on what y'all think about
what sounds better or would sound better to conlangers / linguistics teachers
/ gamers. To me "invention" fits better with the term "glosso*technia*"
(which should probably be "glossotechne", with an eta on the end,
but that doesn't sound right in English); "language creation" with the
more exalted "glosso*poeia*".

What might also be relevant is that someone who knows only a little
bit about conlanging and hasn't encountered the community and
its terminology yet is maybe more likely to do a web search for terms
like "invented language" or "language invention" than "constructed language"
or perhaps "language creation".

In an earlier thread I wrote,

On Mon, Mar 17, 2008 at 9:46 PM, Jim Henry <jimhen...@gmail.com> wrote:
> Conlanging is an art; playing Glossotechnia is not quite the same
> thing, any more than playing Nano Fictionary is the same as the art
> of writing short stories or novels.

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Mon, Aug 18, 2008 at 12:38 PM
To: Sai Emrys <s...@saizai.com>, Alex Fink <000...@gmail.com>
Cc: Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

That was the idea, yes. --- The rules should probably give examples
of mutations as well as affixes.
That's a good idea; let's do it.
Maybe a sentence doesn't get penalized, at least not much, for
sharing a single word with a previous other sentence, only for
sharing phrases of 2+ words? Maybe one point off for each
2-word phrase shared with a previous sentence, two points
off for each 3-word phrase, etc. -- but what base value are
we subtracting from?
Hmm. I think I misunderstood you; I was thinking we don't want to
give much of a bonus to a sentence that doesn't satisfy its constraint
very well, and you were worried about wasting the one-time bonus
on a first sentence that doesn't satisfy its constraint very well. I think
my idea would work for both if we work out the granularity well enough;
i.e. a sentence that satisfies its constraint poorly enough would get
no bonus points because the calculated bonus would be less than a whole
point and so the bonus would still be available for a later sentence
satisfying that constraint better.
......
Providing a default is good. Something like

"Fricatives are allowed in complex onsets before nasals
(or another phoneme class specified by the player of this card)."
I'm not sure that would help. It seems that all the ideas we've tried
out for reducing the size of the deck -- except the fairly unsatisfactory
one of omitting phonology cards altogether and letting players coin
words free-form -- would make the resulting game harder to understand
and play.
Maybe the card itself can just say "Length: Place above the vowel array. While
in play, long and short versions of all vowels are available" -- but
in the rulebook's
advanced section we could say something about the possibility of 3+ length
grades (or three degrees of rounding for that matter).
......
I haven't read him recently -- certainly hadn't read him recently when
I made up the original Glossotechnia deck, although I did Google for
his name and "postposition" and found a page that summarized the
book or article on the subject I'd read many years ago.
Either of those would be good. We've already had some of this
re: sound changes modifying some of the phonemes created by
a Phonemic Contrast card in our CNN Center game; the rulebook
needs to say something about those possibilities.
Yes, probably another paper or two besides the lexicon, to keep
track of the phoneme inventory. After a lot of scratching out
and inserting it might need to be rewritten from scratch once
or twice per game, like our lexicon during the CNN Center
game after the monster chain shift. :)
Yes, not as hard as an Optimality Theory deck, but harder to
use than the existing deck. Maybe harder to use to teach beginning
linguistics students, too.
Hmm. That seems to me only somewhat less constraining (so to speak)
than a fixed set of translation challenge sentences. I would suggest,
rather, that there's a default set of constraints for beginning players
to use, along with a larger set of example constraints to choose from
and suggestions for making up one's own, with optional/advanced rules for
setting constraints in rotation or by randomly choosing from a rulebook
table or both.
Yes, if there are tight semantic or content constraints that could
make the charades, pictures etc. too easy to guess. In the
existing version that's ameliorated by the fact that you don't
know whether someone is going for a word in the group challenge
or in their own challenge. Probably in the constraint version it
could be ameliorated by saying that if there are any semantic
or content constraints, there should be multiple ones.
Or better, draw one, copy the sentence onto a piece of paper, and
put the card back in the deck. We don't want the only Polysynthetic
or the only Metathesis card being taken out of circulation.
Fixed translation challenges could be the sole/primary goal
for a short game, or one goal among several in a longer game.
We do have a tone card. It got played during the CNN Center game,
if I remember right.

Yes, I don't see a good way around this. Having a reasonably
large choice of phonemes in the deck makes it large, verging
on too large to be easily portable. Using POA/MOA cards
would reduce the deck to a manageable size, but at the cost
of making the game almost impossible for non-linguist players
and harder even for conlangers/linguists.
Let's say 12 unique POA cards and 9 MOA cards for consonants,
some of them (bilabial, alveolar, velar POA and fricative, plosive, nasal MOA?)
occurring more than once. Then maybe three backness and five height
cards for vowels, plus rounding, length, tone and nasalization contrast cards.
Some of the height/backness cards would have to occur more than once;
not sure which ones or how many. That would give us about 40 phonology
cards (not counting Phonemic Contrast, which would be about the same as
at present except maybe with extra instances of Voicing and one or two others)
as opposed to 67 phoneme cards in my present advanced deck description
(which, as Alex pointed out, needs some work even if we keep the same basic
Phoneme + Phonemic Contrast deck structure).

What about a compromise? Omit a few of the less usual phonemes -- the
clicks, as Alex suggested, but also the trills and some less common
approximants and fricatives. Make sure we have a plosive or fricative
or both at every possible POA. Omit the voiced plosives and fricative and
add at least one, maybe several more instaneces of the Voicing card.
Then add a few MOA cards, similar to the present Phonemic Contrast cards,
for clicks, trills, and taps -- maybe approximants and lateral approximants
as well?

Or maybe just reduce the phoneme set to 30-40 phonemes most common
in the world's languages and add about 10 phoneme wildcards?
E.g., the translation challenge set should include
"My God! there's an axe in my head!", "I can eat glass, it does
not hurt me," "Verbing weirds language", etc.?
For this version I would go with either

1. the extremely simplified version I'm playtesting now, with no phonology cards
except Sound Change, or

2. a version with simpler phonology cards, basically the phonemes of English
and a handful of front rounded vowels etc. some Americans would be familiar
with from French and German classes, with no Phonemic Contrast cards
or MOA cards, and probably a simpler set of syllables like CV, CVC, CCVC, etc.
Also the phoneme cards might be called "Sound" cards, and use
pseudo-English orthography instead of / as well as IPA symbols.

And probably a fixed set of translation challenges or mix-and-match Subjects
and Predicates; I don't see how to explain the constraint-based version briefly
for non-conlanger/linguist players.
I don't know enough about theoretical controversies in linguistics
to know for sure how to avoid whatever landmines there might be
in that area. I think the only theoretical objection someone has
made to the existing game was that the SVO/VSO/.... cards were
too rigid and the terminology old-fashioned; the game should
(1) have Agent-Verb-Patient instead of Subject-verb-Object
terminology, and (2) allow for topic/comment type pragmatic
word ordering. I've implemented #2 but #1 seemed to require
more explanation for beginning players than SVO..., which use
terms most people have encountered in school.
If we get the Sound Change cards right, probably teaching historical
linguistics and sound change processes would probably be the easiest
sell for linguistics teachers.

----------
From: Alex Fink <000...@gmail.com>
Date: Mon, Aug 18, 2008 at 10:15 PM
To: Jim Henry <jimhen...@gmail.com>, Emerson Knapp <enze...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

Emerson: the card design looks good to me, but then I've never had an
especially fine eye for such things. Cloud effect is neat -- nearly
missed it on first viewing. (Though, is this one of those places
where we have an incentive to keep the total number of colours small
and finite?) Perhaps the only substantive suggestion I have is that
fitting the "Glossotechnia" text to a curve by moving the baselines of
the letters individually looks a little, hm, ragged, but you've noted
this yourself.

Also -- you're in Concord? Do you often have occasion to travel to
the inner Bay Area? That would be useful if we wanted to have some
playtesting sessions hereabouts.

2008/8/18 Jim Henry <jimhen...@gmail.com>:
>> fusional, as it were. [...] -- does this mean you'd
This means that, then, given a sufficiently long period of time during
which no Typology cards are played but several Sound changes are, the
gamelang might become fusional even as the Typology card in force is
Agglutinative. Maybe this is no trouble, since this is an unlikely
scenario.
[...]
[...]
Yes, seemingly one of us misunderstood the other, at least.
It may. I find that an indirect way to achieve the desired effect,
though, and while that's perhaps no criticism on its own it renders
the mechanism nonrobust -- if we subsequently tweak other aspects of
the scoring, we'll likely have to revisit the granularity.

Plus, computing with percentages would slow some people down. On the
whole I'm more inclined at this point to go for something like Sai's
for base values, which is lean and not too far off what we're after.
Or, better, one plus these numbers, given that we want every sentence
to score something. I'm probably missing why Sai didn't let lexical
variation and grammatical variation both vary from 0 to 2
independently, but that seems like it would be simpler yet and just as
good.
Yes, good. (A strange default, though, even be this just an example.)
I meant it to address the problem of whether, say, if you can have a
plosive before a fricative and a fricative before a plosive in the
onset, you're then allowed to use onsets like /ststststs/. This is a
minor issue, though, and my solution is probably unduly complicated.

If you've read through Tom Chappell's latest response to your clusters
inquiry on Conlang, might you summarise it? Too many links for me to
have had time to follow and pick through.

[Three vowel lengths, etc.:]
Yes, that's what I was trying to suggest.
Ah, okay. In any case, hardly important, given that we can't reduce
the deck size by it.
Which gets tedious -- don't forget whoever it was guessing "playing an
evil chain shift"!

I had an off-the-wall idea re this several days back (which I markedly
don't mean to propose as a recommendation for the basic rules):
suppose you recorded all the forms in the gamelang lexicon in (to take
the easiest case) an alphabetic conscript, and notated the phonemic
value of each letter. Then evil chain shifts are constant time in the
lexicon size to apply: just rewrite the values of some of the letters!
I spoke to Sai IRL on this point a bit. As I understood the rules he
means to canonise as "core constraints" are not constraints in the
thread's established sense: he gave examples such as points for
defining one's word entirely in the glossolang, for translating a
sentence of the target text (if using one), for getting a laugh from
the assembled players. He insists that there may be some constraints
of our form that he'd think should be canonised, but couldn't name an
example.

Anyway, I agree with you here; it may turn out that I agree with Sai as well.
In particular that individual challenges are kept obscured from the
other players.
Good idea. Semantic constraints that have these problems might turn
out not to be so problematic, anyhow. If one of the constraints is to
make a sentence about cats, even if this gives the guy who's coining
the word for "cat" a free ride, that's a one-time happening (barring
semantic shift, loss, etc. -- hm, can a lexical item even be lost?).
We have a tone card now; I don't remember whether it was a later
addition than that game. The tone that got into our CNN Center
language was a knock-on effect of a phoneme loss, I think.

It's more flexible to have a tone card than tone cards, I'd say, in
that there're several kinds of tone system, and they might not mix
well.

And, heh, to say that our choice of phonemes by itself is forcing
phonologies towards naturalism is a little rich, at least as the rules
stand ... besides, what would say an engelangly collection of phonemes
even look like?
Indeed, this is a difficult problem. Of the several solutions you
listed below (which I've snup) -- all phonemes, all features,
something in between -- none's cleanly better.
Potentially. Would a community of new players know what to do with so
many phoneme wildcards? I guess that, at the least, this would have
to go along with an IPA chart and some explanation of articulatory
phonetics in the rulebook (in which we may as well include a link to
some stable "hear the sounds of the IPA" website).

[moved from above]
I think, so long as we're keeping an eye towards publishing a polished
version, usability issues (like this, and deck size) trump any worries
we might have about naturality of the resulting sorts of phonologies.

What forces our hands particularly is that we don't really have
freedom to postulate variants that require separate decks on an equal
footing, given that we're going to be publishing a game with one deck.
Without that desideratum I'd be inclined to leave both as variants on
equal footing at this point.

>> * Pitch / Target audience(s)
[...]
*grin*

>> * Gamer geeks / Party game

which are two quite different communities yet. The rules-complexity
tolerance and esteem of strategic depth are much higher for the first,
the capability to support the rollicking free-flowing fun atmosphere
for the second.
"Language-based games"? Going by your examples this doesn't seem a
useful category. With Scrabble the closest commonality would seem to
be the skill of getting words with letters / phonemes in the right
place (in G'nia for potential phonological constraints), but the
requisite skills are vastly different, Scrabble needing good recall /
memorisation of the English lexicon, G'nia overall letting you put
them there yourself. The closest Apples to Apples seems to come to
this is that both deal with, er, concepts, that occur as meaning of
words...
(see above, re freedom to posit deck variants)

I'd find losing the rich phonological potential, however it's done, to
be a misfortune. I guess this turns on how much we can get
non-linguist players to buy into it as "flavour".
The concept I'd expect to not give gamers difficulty, only the points
of linguistics on which the constraints may depend.
Yes. Or more generally diachronic change in various aspects of
language. We have a start on the typology in there, for instance.

But getting this _really_ right -- as right as having game-generated
phonologies reflect a naturalistic distribution of phonologies -- is
probably just as difficult as that has been proven to be.

Alex

----------
From: Sai Emrys <s...@saizai.com>
Date: Mon, Aug 18, 2008 at 11:11 PM
To: Alex Fink <000...@gmail.com>
Cc: Jim Henry <jimhen...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

Jim:
That seems like a very good idea, actually. Perhaps have the extra
ones just mentioned in the rules.

One could even have rules that make it easier to make whole classes;
e.g. if you already have t, k, p then getting d, g, b could be a
single move ("add voicing to existing series" ish). Again, this would
push it towards naturalism.

Re. the scoring rubric, the only reason I would have it be one list
rather than two separate variables is that adding is bad. The more
adding, the more things to track, the more time to do secretarial work
and less time playing. (Alex, you don't count for this, since you're
insanely good on this particular point. It does take time for
normals.)

Generally, the scoring should be uber uber simple. No math more
complicated than (base #) + (1* X) or so. It just gets in the way and
the extra finesse is probably unnecessary.

Re the card image, my only comment is that it's off-center and either
"Glossotechnia", or the sun, or both, should be moved upwards more so
that it's balanced.

- Sai

----------
From: Emerson Knapp <enze...@gmail.com>
Date: Tue, Aug 19, 2008 at 2:10 AM
To: Sai Emrys <s...@saizai.com>
Cc: Alex Fink <000...@gmail.com>, Jim Henry <jimhen...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

Re. scoring and "constraints": I have an idea, from a gaming perspective rather than a language invention perspective. Some of this probably overlaps with what's been said already, but I just can't process all of what's been discussed into one coherent set of rules (so correct me if I'm wrong, or am repeating any of you).

At the beginning of the game, a set of goals (rather than "constraints", which sound like things restricting you rather than things you must meet) are picked, either from an index in the book, or from cards, the how is irrelevant right now. These goals specify certain types or qualities of sentences that the players must try to compose. As a very simple example, one goal could be "Create a sentence that includes 10 morphemes."

The goals sit on the table, or are at least kept in mind. Some goals are boolean, you can just achieve them, and some are scalar, there are different degrees of achieving them, to use the terminology that you've already used.

The boolean goals, once a player achieves them, give points to that player (point number specified on card or in the index), and become null, so no other players can achieve them (think Yahtzee, but with everybody playing together). At this point, another goal could be drawn, unless the object is to eliminate all goals.

The scalar goals, alternatively, could be what I jokingly call "latent points" when playing games. The goal could be "10 points to the player with the longest sentence on record", so each time a player creates a longer sentence, the points would be lost from the previous player and given to the creator of the longest sentence. The concept for this comes from the "Largest Army" and "Longest Road" achievements in the game Settlers of Catan. So the card for that goal would be passed around to whoever currently had the longest sentence on record.

All sentences could still be given points based on the rubric proposed (which I don't quite understand yet), but the players would be vying to achieve the goals on the board, for some trying to be first, and for some trying to be best.

The final goal could be either a total amount of points, or a certain number of goals achieved. Perhaps this is exactly what you were thinking, perhaps you'll hate the idea, it's hard to tell. What do you think?

Also, regarding the number of cards: I don't know if 160+ cards is really that big of a deal. The Munchkin card game is the first thing that comes to mind, it has 168 cards in the initial deck, and 112 more in the first expansion, and it's never seemed very cumbersome. If the deck needs to be 200-250 cards, it's really not that bad, especially if there is a play deck and a goal/constraint/sample/whatever deck, so that it's not all one huge pile. 160 cards on index card paper or whatever you're using takes a lot more space than once it's printed as playing cards.

At your mercy,
Emerson Knapp

----------
From: Alex Fink <000...@gmail.com>
Date: Tue, Aug 19, 2008 at 10:40 PM
To: Emerson Knapp <enze...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, Jim Henry <jimhen...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

2008/8/18 Sai Emrys <s...@saizai.com>:
Were you not aware of the phonemic contrast cards? They do this,
assuming we keep them under this proposal.

If there are many phoneme wildcards, we might consider a "wildcard
recycling" rule (I've seen a Scrabble variant do this): if a wildcard
has been played as say /t/ and you have a real /t/ in your hand, you
could swap it for the wildcard.
*phbbt*
I understand that we want simplicity, but not at the cost of having an
evident imbalance or even exploit, which has been the general level of
the discussion so far. It's not finesse to not want everyone making
all their sentences interminable lists to be the optimal strategy...

I think my refinement of your scheme is at least as simple: it's just
counting new items up to a limit of two for each all together, whereas
yours is doing that but then imposing some funny thresholding where if
you have two of one kind and one of the other you still only count
two.

Here's ... I'll reserve judgment and call it a property of your
scoring scheme. Counting arbitrary "innovations" in morphosyntax is
difficult and probably subject to opinion; the cleanest-seeming way to
do this is only to count innovations which have been explicitly
introduced in the non-sentential phase of a turn. But, in that case,
there is only one point (two, with my revision) of amortised base
value available on most rounds, since only one word or affix is coined
per round; on the uncommon occasions that a morphosyntactic card is
played there are two (three). If you make sure to use the word you
coin that turn, as you certainly would by default, you automatically
get this amortised value. The only way you can hope to exceed this,
once things have stabilised and we're out of the early game, is by a
previous player coining a word and leaving it unused, so that you can
use both it and your own coinage. This is probably undesirable, that
to get more points than the norm you have to rely on slipups /
benevolence of your opponents.

I meant but forgot to say in my last response that it seems a good
idea to integrate Jim's "sharing phrases of 2+ words" criterion into a
scheme of this nature. E.g. one might add to your scheme one more
cumulatory point of base value for "structural innovation", i.e.
sharing no 2-word substring with a previous sentence.

The precise criterion for "structural innovation" may need tweaking,
at the least to not favor either isolating or highly synthetic langs.

2008/8/19 Emerson Knapp <enze...@gmail.com>:
The ruleset you've described is, to my eye, certainly within or very
close to the cloud of rulesets our discussion has been floating among.
We haven't really achieved coherence yet.
This is new TTBOMK, that boolean goals go away once one player
achieves them. We'd been working on the premise that they'd stay
around, but there would be a bonus for being first to them. (And then
debating extending the bonus to scalar goals.)

Your version might be cleaner. It would be more competitive, and more
dynamic, but perhaps over the line toward cutthroatness: plenty of
opportunities to be nearly at a goal only to have it snatched away for
good. It would burn through boolean goals very quickly, too. In my
first proposal the goals came from the players at the start of the
game, so there was only a small finite pool of them, and they
consequently wouldn't be expendable the way this ruleset requires.
Providing this many goals is a materialological obstacle however we do
it, I think, just as providing target sentences was/is.
Yeah, it's those very achievements in Settlers which suggested to me
that kind of handling of scalar constraints (though I tweaked the
statefulness a little). I think we're on the same page here.
Oh, those are natural ending conditions, aren't they. We'd been
carrying over current G'nia's secondary end condition of (volitionally
or otherwise imposed) time limit.

Pretty close to what we were thinking, on the whole.
Nor I. Sai?

Alex

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Aug 19, 2008 at 11:30 PM
To: Emerson Knapp <enze...@gmail.com>

See http://www.playingcardsindia.com/packing-options1.html for some background.

110 cards is a standard double tuck box. More than that requires
custom boxes. That makes the cost go up significantly.

Also, I would point out that I know of nobody who ever brings their
Munchkin *boxes* around with them. They're simply too big to do unless
you've got something in mind. Provisioning for an classroomful of
people has similar problems - consider how a dozen sets would fit in
the teacher's backpack.

The thing is: yes, it's *possible*. It would be more expensive, and it
would be less likely to be something people carry around with them
(thus less prone to spontaneous play), both of which reduce how well
it will sell. The cost is especially significant because we don't have
the money or audience to take advantage of volume discounts such that
this levels out; to us, the marginal cost increase will be that much
more. And our audience is pretty "cost-sensitive", as the jargon goes.

IME, constraints are a good thing. They help you trim the fat off and
make for something that's ultimately better.

If it turns out that, after trying to make it slimmer, we determine
that it really does play better with more cards, then okay. It's my
problem to run the numbers and figure out what we can pragmatically do
financially, and I'll come up with something.

First pass though, let's try to see if this can be done in a better
way with fewer cards. I believe the answer is yes.

- Sai

P.S. Just FWIW: I work at an online game company (zynga.com). Our
market's not the same, and our games are very different, but it does
seem to be universally accepted - including by those people coming
from design backgrounds in more traditional game companies - that
simplicity rules unless overwhelmingly demonstrated otherwise.

I would point out that I cannot think of any example where a game has
many cards, and that bulk of cards is not because of a large amount of
*creative* content [e.g. Munchkin, Magic, Trivial Pursuit - all those
cards are semiunique and are essentially new content with up to 5-7
different underlying templates] - RATHER than because of a large
amount of *game mechanic* content - which is what we've been
discussing here.

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Aug 19, 2008 at 11:31 PM
To: Alex Fink <000...@gmail.com>
Cc: Emerson Knapp <enze...@gmail.com>, Jim Henry <jimhen...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Wed, Aug 20, 2008 at 9:20 AM
To: Alex Fink <000...@gmail.com>, Emerson Knapp <enze...@gmail.com>, Sai Emrys <s...@saizai.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

Yes, you would have to have several Add Inflection cards played
while there is no typology specified, creating several inflections,
then sound changes that affect those particular inflections in some
contexts, fusing them, and then someone plays an Agglutinative
card. Not likely.

That reminds me: ever since I added the Typology cards I've thought
off and on about removing the Add Inflection cards. I haven't yet
because my understanding is that even many "isolating" languages have
a certain amount of derivational morphology, and in practice the Add
Inflection cards are used as often as not to add derivations as well
as inflections. If I keep them I need to change the text to cover
derivations as well as inflections somehow, and maybe say "if
Isolating typology card is in play, this can only be used for
derivational affixes/mutations" --- but given we're trying to reduce
the size of the deck, maybe we don't want to keep them.
That's probably good.
Yes, just off the top of my head. I reckon each syllable wildcard
would have a different fixed part and a different variable part
with a default, except one or two that would be true wide-open
wildcards that can do anything. So e.g.

"Plosives (or another phoneme class specified by the player
of this card) are allowed in complex onsets before fricatives."

"Nasals (or another phoneme class specified by the player
of this card) are allowed in complex onsets before approximants."

"Nasals are allowed in complex onsets before voiced plosives
....etc.? Details to be revisited after I digest the "Most common
consonant clusters" thread and the monster list of links Tom Chappell
posted.
.......
Yes. Anything that produces card text of more than ~40 words, or
requires you to refer to the rulebook to interpret what a card's text
means, should probably be reexamined carefully before we implement
it.
I've hardly followed up any of his links either, yet.
That would be a neat option. I've thought off and on about ways
to incorporate that into the rules in some way, but haven't come up
with anything yet. Except maybe allow the player who plays a new
phoneme card to come up with a conscript letter for it, unless he's
doing a phoneme split in which case maybe the spelling of
existing words doesn't change, only their pronunciation...?
Definitely something for the optional / advanced rules section.
Some of the funniest moments of any Glossotechnia game, I think, I
think, have been people doing absurd charades that it turned out no
one should guess. Should we give bonus points for that? Otherwise I
think Sai's basic idea of core constraints that always get points
whatever other constraints are in play is basically a good one, but a
game with no other constraints than those might be impoverished.
Someone could use "extend meaning" to make the word for
"cat" mean "vertebrate", or "restrict meaning" to make it mean
"Maine Coon" -- either of which would make it less useful for
talking about cats at just the right level of specificity.
Schematic and symmetrical, I reckon. For each point of articulation
used in the lang there is a phoneme for every manner of articulation,
and vice versa. If voicing is distinctive, there are voiced and
unvoiced versions of everything except maybe the approximants. The
number of phonemes is 2^N for some N.
Probably so. Ultimately we need to think about at least two
different versions, though -- a version simple enough for gamers
with no linguistics background at all might be too simple to be
interesting for conlangers or useful for linguistics teachers.
For instance, maybe the Polysynthetic card is only playable
when Fusional or Agglutinating is in play, and so forth --
a language is unlikely to go from isolating to polysynthetic
without an intermediate stage. Similarly maybe OSV
can't be played when e.g. SVO is in play, -- we could generalize
it by saying that a word order card being played has to be
one transposition away from the existing word order, maybe?

On Tue, Aug 19, 2008 at 2:11 AM, Sai Emrys <s...@saizai.com> wrote:
> Jim:
So keep the existing phonemic contrast cards (or even increase the
number of voicing cards), reduce the Phoneme card list to the cardinal
vowels, common unvoiced fricatives and plosives, voiced nasals and
approximants, and add several more phoneme wildcards and/or Fill Gap
cards? What does that get us in terms of number of cards?

The problem is that I think the Phonemic Contrasts are one of the
hardest parts of the existing game to explain to non-linguist players
-- harder than explaining the pronunciation of specific non-English
phonemes, or how the sound changes work, or what a postposition is,
etc. So a version that heavily leans on Phonemic Contrasts (or Manner
of Articulation cards) to get a naturalistic phonology probably isn't
suitable for non-linguist gamers.
That's probably a better term, less potentially misleading.
The rubrics proposed have been various and a bit vague; we're
still trying to come up with something simple enough to require
very little play-time arithmetic but powerful enough to be fair
in rewarding various achievements in proportion to their
difficulty.
Maybe if one had a balanced set of boolean goals that were all about
equally difficult, aiming for the most goals achieved would be a lot
simpler (less arithmetic do to) than aiming to get the most points.
That limits the players' freedom to come up with arbitrary goals for
use in any given game, though. More probably there would be a set of
sample goals in the rulebook with associated numbers of points, and
just before a game players could come up with other goals which the
other players would assign a number of points to.
Longer and wider, yes, but as I think I mentioned earlier, my handmade
index-card deck of 160-odd cards is thinner than the 136-card
Chrononauts deck. Maybe Chrononauts is printed on thicker
cardstock than Munchkins?
That might be a good idea, but wouldn't it be an even better idea
when there are few wildcards than when there are many? I feel like
I'm missing something.
Which they might do, I suppose, if the other players couldn't guess
their charade and it wound up being assigned a meaning they couldn't
think of a use for right away, and they had to improvise a sentence
using previously existing lexical material. Otherwise, as you say,
it's hardly likely.
Maybe we would give points for innovating in any area of grammar not
explicitly covered by the syntax cards? I'm not sure why we would
give people points for playing a syntax card and then being the first
to use the new syntax rule in a sentence, unless it's a syntax
wildcard or a secondary word order and they're doing something
creative with it.
.....
.....
All right. Our discussions so far about changing the syllable and
phoneme cards have reduced the deck size from about 160 to about 130
cards, I think -- we've come up with several ways (none of them
without drawbacks) to reduce those cards' number by about 20-30:

1. eliminate some less common phonemes incl. all voiced plosives and
fricatives and add some more Phonemic Contrasts incl more copies of
Voicing;

2. eliminate all Phoneme cards and replace with Point and Manner of
Articulation cards;

3. eliminate all but the ~30 most common phonemes of the world's
languages and add 8 more Phoneme Wildcards.

Where can we trim another 20 or so cards? Omitting the Add Inflection
cards as I mentioned earlier gains us 2; but we probably want to keep
the Drop Inflection card.

We could drop the "Draw 4" card (keeping "Draw 3"), drop the four
swap/new/rotate challenge cards and replace them with a couple of
Add and Drop Goal cards, drop the Back-Formation card and add
something to the rules suggesting that you can do this anytime on your
word-coining turn: net gain, 4 cards, or 6 if we don't add the
Add/Drop Goal cards.

Is there some alternate mechanic we could use to obviate the need
for the six Typology cards?

I don't see a way to reduce the Suprasegmental cards without
eliminating them entirely.

We could reduce the number or even the number of kinds of Sound Change
cards, but I'd rather reduce almost any other part of the deck than
that. Eliminating all the extra copies of different kinds of Sound
Change card and having just one of each kind would gain us 5 cards.

Eliminating the extra copies of all the syntax cards so there's one of
each kind (probably a bad idea for game balance, but hypothetically)
woud gain us 7 more.

Net gain: 18 cards, or down to just over our target, I reckon. But I
fear this drastic reduction would imbalance things.

My original alpha-test deck was about 90 cards, and was playable
though definitely sub-optimal. It had too few phonemes and syllables
relative to other card types, and had no Suprasegmental, Phonemic
Contrast, or Typology cards, and fewer/less varied Sound Change,
Grammar Change and Action cards.

BTW, I've revised the text of the Suprasegmental cards -- does
this make sense/use correct terminology?

Initial stress: While in play, the first syllable of each word gets
the primary stress. The player playing this card specifies whether
the language uses stress or pitch accent.

----------
From: Matthew Haupt <film...@hotmail.com>
Date: Wed, Aug 20, 2008 at 12:12 PM
To: Sai Emrys <s...@saizai.com>, Alex Fink <000...@gmail.com>
Cc: Emerson Knapp <enze...@gmail.com>, Jim Henry <jimhen...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

Hey everybody, I know I'm new to the conversation; its been a very busy week with my business and my wife is going to be having a baby very soon, so I'm sorry I haven't contributed much. But I saw this post from Sai and wanted to chime in and emphasize that you want to be BRUTAL with cutting things for simplicity, at least at first. Make the game work at the simplest level you can, the fewest cards you can, and only when you've reached that point, should you start thinking about adding complexity. I saw that there was A LOT of discussion about challenge cards and some about phoneme changing cards. I nixed these out of Make A Lang pretty early on, and I would suggest the same here. Everyone can make up their own challenge sentence; put an arbitrary measurement on it, like any sentence as long as it is ten words long. Write it down on an index card, or tell it to the player on your right, to keep everyone honest.

~Matt Haupt

> Date: Tue, 19 Aug 2008 23:31:32 -0700
> From: s...@saizai.com
> To: 000...@gmail.com> CC: enze...@gmail.com; jimhen...@gmail.com; pjschlei...@gmail.com; film...@hotmail.com; mrpark...@gmail.com; jati....@gmail.com; cschu...@yahoo.com

Get thousands of games on your PC, your mobile phone, and the web with Windows®. Game with Windows

----------
From: Matthew Haupt <film...@hotmail.com>
Date: Wed, Aug 20, 2008 at 12:17 PM
To: Sai Emrys <s...@saizai.com>

Hey, I know I said I'd email you some files in a few days - Its been really busy all last week and this week might be the same. I did try to give some input just now, basically echoing what you said. I have been doing some finishing work on Make A Lang and I haven't been able to read all the emails flying around. Tell me again - the purpose is to create a solid language creating card game that you can sell at the next LLC, correct?

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Wed, Aug 20, 2008 at 2:17 PM
To: Matthew Haupt <film...@hotmail.com>
Cc: Sai Emrys <s...@saizai.com>, Alex Fink <000...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

Whoo hoo! Good for y'all!
We might easily drop the fixed set of challenge sentences, and instead
come up with a system for determining the point value of sentences you
make up yourself. But the sound change cards are pretty central to
the feel of the game -- even the non-conlanger players who've given me
feedback on it have said those are one of the most fun aspects of the
game and they'd like to see relatively more sound change cards in the
deck. In the simplified verison for non-conlangers I left out the
Phonemes and Syllables and Phonemic Contrasts altogether but I still
have Sound Change cards, and maybe a higher proportion of them relative to
Syntax and Action etc. cards than in the advanced deck. We might
simplify the sound changes, especially for the non-conlanger gamers'
deck -- leave out the chain shift and metathesis maybe -- but not
leave out the sound changes altogether.
How much playtesting have you done in that mode? How has
it worked?

What seems to me a little better is for everyone make up two
sentences, shuffle them together, and everyone draw one at random
plus one more face-up for the group. That way there's the element of
external challenge, without having a fixed set of sentences that come
with the game and which you're sooner or later going to run out of --
soon if they're a small enough set to be easily portable, eventually
if they're a monster Trivial Pursuit-size box of cards. There should
be an objective or fairly objective way to score them, which is done
when you are ready to translate your sentence into the gamelang -- in
the simpler variant where everyone is aiming at a translation of a
specific sentence instead of having a variety of goals to work toward
that various sentences could satisfy to greater or lesser degrees.
There would be guidance about making up sentences, the rules for
scoring them plus a ballpark point value range for a particular game,
e.g. all sentences have to be 6-7 points for a short game, or 10-12
points for a longer game, etc., where a sentence gets 1 point for each
concrete word, 2 points for each abstract word, 2 points for relative
or subordinate clause or non-indicative mood... something like that.
The only problem with this method I see is that players might not
agree at scoring-time whether a particular word counts as concrete or
abstract.

--
Jim Henry

----------
From: Alex Fink <000...@gmail.com>
Date: Fri, Aug 22, 2008 at 12:51 AM
To: Jim Henry <jimhen...@gmail.com>
Cc: Emerson Knapp <enze...@gmail.com>, Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

2008/8/20 Jim Henry <jimhen...@gmail.com>:
Not _quite_ as unlikely as all that. I meant to describe the scenario
in which an Agglutinative card was played as the first thing, meaning
the inflections can be added freely without needing any Add inflection
card. But, still, unlikely.
> a certain amount of derivational morphology [...]

If Add Inflection were dropped, gamelangs with Isolating in force but
that do have some derivations would still be generable, if these
derivations were created in an earlier typology and not removed by the
player of Isolating. Given this possibility I'm inclined to regard
Add Inflection as redundant at this point.
Yes, all reasonable examples.

Coda clusters might be a little more delicate, given that in the
situation where no cards permitting coda consonants have been played
there's nothing to condition a before or after environment on. But I
suppose that a reasonable handling would be to say e.g. (omitting the
variability)
"Nasals are allowed in codas following an approximant, or as the only
member of a coda."
Indeed. I'd agree with (the suggestion of) not respelling on splits.
For that matter, players could also hold off respelling on other
changes as well, and eventually get nice fun etabnannimous systems,
and then promulgate spelling reform...
"... that no one should guess" why, or by what standard?
If you include challenge sentences among the constrains you're
excluding, then I'd think so. Otherwise it's much like the extant
version, though.
Right. But neither is outright loss (alternatively, replacement),
which is a process natlangs certainly get. This might be worth a
card.
Yes, I suppose so. I've made a few phonologies that render
bit-strings. Anyway, the symmetry isn't a distinguishing factor
between these and nat-phonologies by itself (coarsely; differences of
degree probably exist).
I have the same suspicions. Someone who knows these things: is having
variants with multiple decks viable, production-wise / economically /
mindshare-wise / etc.?
For instance. My knowledge of typological diachronic syntax is scant,
though. I couldn't say, for instance, whether the most typical word
order change is because of a marked information-structural pattern
(topic fronting e.g.) becoming unmarked, or influence from a
neighbouring language, or a periphrastic construction (with different
word order) becoming neutral, or something else yet. And each of
these will have a distinct transition matrix between the various word
orders.
Well, factoring out as many constrasts as possible and getting all the
common phonemes aren't quite the same goal. I think the commonest
voiced stops, at least, would certainly get included under the latter
but not the former.

And again, "cardinal vowel" might not mean quite what you want -- in
the usual use both rounded and unrounded variants of each represented
(height, frontness) are cardinal vowels. (It's the first eight that
go round the vowel trapezoid with the usual unmarked roundings.)
And I presume you've done more explaining than any of us, so I'll
believe this, despite it not seeming _that_ hard naively.
Recasting in terms of number of goals achieved alone would be
scrapping the whole base value scheme, and with it the discouragements
against making minor variants of previous sentences -- which I suppose
we don't need quite so desperately if goals go away once achieved, but
still seems a shame to lose. Granting that loss, something like Jim's
suggestion would be fine.
Hm, I can't remember why I was thinking that'd be suited to many
wildcards. You may be right.
Point.
I was worried that this might be too much a matter of opinion. Maybe
it's okay; unclear cases might be rare. On the other hand it might
lead to players overstating the "innovations" in their sentences. On
the gripping hand that might be good too, in the mode of revealing all
the niggly wrinkles in the grammar that are often glossed over.
[...]
Perhaps. Though the dual of my argument would apply here: you can
already drop inflections by switching to Isolating.
All good.
Haven't thought of a user-friendly one yet. A non-user-friendly
option goes something like this: the rulebook contains a scale of
typology types with their effects on gameplay, and ther are cards to
move up and down the scale.
One could leave the Wildcard suprasegmental card while discarding the rest.

I've (counterproductively) thought of another class of stress pattern
that's probably common enough to represent on a card: these are the
weight-sensitive patterns. (Ultimate if that's a heavy syllable,
otherwise penultimate, for example.)
Likewise, I'd rather not reduce these. If necessary we could strike
out one or two of the uncommon types, and perhaps compensate with some
nature of sound change wildcard.
Again, wildcards might help with proportions at least; probably not so
much with game balance. E.g. one might reduce the basic clausal word
order cards to one each of VSO, SVO, SOV, and a basic clause order
wildcard.
Well, testing should provide evidence...
Reads fine to me.

What should happen in-game when a suprasegmental card and the Tone
card interact? I suppose you can have a stress-accent system in
conjunction with constrastive tone, but I don't know how common in
natlangs that is; it may be that more often tonal languages lack any
good equivalent to stress accent at the domain of the word. Anyway,
there are conceivable cases of actual conflict -- what happens in
these?

Alex

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Fri, Aug 22, 2008 at 8:24 AM
To: Alex Fink <000...@gmail.com>
Cc: Emerson Knapp <enze...@gmail.com>, Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

Should be "...that no one *could* guess." Weird lexicotypo.
I've thought of adding more "Meaning Change" cards like
Archaicize (which would be much like loss, except the word
is still there to be used if you don't mind sounding overformal),
Amelioration, Pejoration.... but haven't mentioned them before
on this thread because we've been talking about trying to reduce
the deck size.
......
Yes; and if we're trying to make all phonemes possible with base
phonemes plus contrast or MOA cards we'd include some relatively rare
things like /q/ because we'd need at least one base phoneme at each
POA. But if we include enough wildcards maybe we don't need to do
that. And maybe about half of the wildcards should be of the "Fill
Gap" variety, i.e. constraining the player to argue for his phoneme as
filling a gap in the existing inventory.
OK, two more here.
One of each? Two of each? The first means they would rarely come up
and the second only gains us two cards.
Yes, someone on ZBB suggested something like that. Maybe there are
multiple suprasegmental wildcards, one constraining the player to
specify a fixed stress position, one allowing them to create a
weigh-sensitive system, one unconstrained? That doesn't reduce the
deck size much though.
Do you reckon you'd be able to test an advanced deck of this reduced
type at a Bay Area conlang meet in the next few months? The only
playtesting opportunities I'm likely to have before LCC3 are with
non-conlanger gamers. (One this coming Sunday, I hope.)
The later-played card overrides the earlier-played card, I reckon.

----------
From: Sai Emrys <s...@saizai.com>
Date: Mon, Aug 25, 2008 at 11:07 PM
To: Matthew Haupt <film...@hotmail.com>

The purpose is to create a solid conlang card game, yes.

Whether it's ready by next LCC or not is less important than ensuring that it's high quality. Yes, suitable for commercial use, though to me this is more a quality marker than a financial issue. I don't honestly give it more than 10% chance of making more than $100 or so net, which is a relatively insignificant amount.

The point is simply to get something good out - not just a rough game suitable for hardcore geeks only, and not something watered down such that it's not enjoyable for conlangers. It's a balance.

- Sai

----------
From: Sai Emrys <s...@saizai.com>
Date: Mon, Aug 25, 2008 at 11:17 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Alex Fink <000...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

> variants with multiple decks viable, production-wise / economically /0
> mindshare-wise / etc.?

Only marginally.

It's possible to release either two different sets, or one set plus an expansion set, but the costs rise accordingly and the net sales may well be less (because splitting it up means that each has less volume, hence less cost savings).

I would prefer not to go there however primarily because I'm yet quite unconvinced that it is in fact necessary, i.e. that we couldn't have a deck perfectly capable of handling both basic and advanced play only by tweaking the rules.

- Sai

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Aug 26, 2008 at 8:28 AM
To: Jim Henry <jimhen...@gmail.com>
Cc: Alex Fink <000...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

Jim on Conlang:
> The game ran between 3.5 and 4 hours, and lasted 33 turns, if I'm counting correctly.

7 minutes per turn? Why so long?

- Sai

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Tue, Aug 26, 2008 at 10:42 AM
To: Sai Emrys <s...@saizai.com>
Cc: Alex Fink <000...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

.....
The phonology portion of the deck is the main issue here, I think.
Non-conlangers look in bewilderment on a set of cards with
IPA symbols on them, even though they all have example words
in English and other languages on them; and conlangers and
linguists would (rightly, I think) turn their noses up at a deck
where the phoneme cards (or "Sound" cards) have the sounds
represented by English fauxnetics.

At some point I want to try making a deck with English fauxnetic
orthography "Sound" cards in place of the "Phoneme" cards
and see if:

1. it's significantly harder for non-conlangers than the current
simplified version with no phonology cards at all;

2. it creates significantly more interesting languages than the
freeform word-coining of the current simplified version.

If not, one could create a simplified deck by simply removing the
Phoneme, Syllable Onset/Rime, Phonemic Contrast and
Suprasegmental cards from an advanced deck, and having
stuff in the rulebook about how the text of the Sound Change
cards is to be interpreted when you aren't using Phoneme
etc. cards (or you could leave out Sound Change cards too);
but the resulting simplified deck would (as last Sunday's experience
showed) be unbalanced with too many syntax
and sound change cards relative to other kinds of cards.

The playtesters this weekend told me they definitely think
I need to either focus on language geeks as the target
audience, or else make two different versions of the game,
and they all liked this version better than the previous versions
they'd played, though they had some criticisms about deck
balance, which I outlined in my last email.

Anyway, the advanced version is liable to be an easier sell,
with a ready target market (conlangers) and another target
market that shouldn't be too hard to persuade (linguistics
teachers). The simplified version for non-linguist gamers
would be harder to sell. We can save it for later, and maybe
use the established community of conlanger/linguist players
as a medium to start marketing the simplified version
a few years later: "here's a version of Glossotechnia you
can play with your gamer friends who don't know a phoneme
from a letter".

Mainly the concultural material, some conversations on how
the rules, card text, goals set at the beginning, etc. should be
interpreted, some conversation about how a sound change
or syntax change would work, pauses as someone composed a
sentence, read it out, and then the rest of us tried to
parse it, etc. There were a small number
of charades it took us several minutes to guess, but I don't
think they were a major contributor to the unusual length
of the game.

The largest single factor was probably the conculture
development; but it made the game so much more fun
that I wouldn't want to dispense with it, at least not with
groups of players like these.

And I'm not perfectly sure how long the game lasted; I didn't look a
clock the moment we stopped talking about playing and
started shuffling the deck, dealing, explaining rules, and
playing, and I didn't look at a clock the moment we
stopped playing and starting talking about what worked
and what didn't and so forth. It might have been as little
as three hours, maybe.

----------
From: Alex Fink <000...@gmail.com>
Date: Tue, Aug 26, 2008 at 2:29 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

2008/8/22 Jim Henry <jimhen...@gmail.com>:
Ah. Well, in that case, I don't know. Best left to the players'
(majority?) discretion, if indeed we have the rule at all.
Right, all of those are plausible, as well as Concretisation,
Abstractification, Metaphor. But (to restate the obvious) none of
these are outright loss. Do you have something against such a
(hypothetical) card?
[...]
Yeah, I'd say wildcards are the way to go with the "30ish most common"
scheme. Trying to represent everything won't achieve much reduction
over the current deck at all.
Independent of our pressures toward a smaller deck I'd say the latter.
This also allows the typology to swing about on the scale more: if
theye was only one of each card they'd probably occur in play in
something very close to alternating sequence, so that the gamelang
typology wouldn't wander very far over reasonable lengths of game.

What would the scale look like, for this?
It doesn't, but it would be a little more representative of
naturalistic patterns.

Has there been recent discussion of G'nia on ZBB? Could you point me
to the thread?
The next few months, very likely.

On the playtest with the simplified deck:

> When I played a
> sound shift card to merge /U:/ into /u/, a couple of players were
> puzzled because they'd been representing both sounds with the same
> grapheme, probably "oo".

How do we encourage nonlinguistic players to spell words precisely in
their lexicon, without imposing some system like IPA on them? How do
we get them to catch the fact that [U] and [u] are different beasts?
This is necessary if we want to avoid the gradual introduction of
discrepancies in people's copies of the lexicon, I think.

> I was surprised at the way some players used the sound change cards.
> In this deck, where there are no Phoneme cards, the Sound Shift cards
> have text like this:
>
> "Specify a sound that occurs in one or more words, another sound
> that will replace it, and optionally a limiting context."

We now have several pieces of evidence that Neogrammarian-style sound
change isn't very transparent to linguistically naive players: they're
not likely to understand the conditioning factors part, or do anything
very plausible. The obvious to the simplified sound change card text
would insert the word "phontic" somewhere around "limiting context",
but otoh a case could be made for just dropping the limiting context
outright (or relegating it to the rulebook): these are the sort of
players who don't much care for the details of sound change after all.

(And there's perhaps some natlang precedent for ridiculous changes
such as /ma/ > /kEl/, if not phonetically motivated ones. I was
recently reading about a case of taboo avoidance among the Rapanui
where they'd avoid (parts of) the king's name even as substrings of
something else. They had one long-lived kind Tu:-something, who
effectively accomplished by taboo the unconditioned sound change /tu:/
> /ta?i/.)

> Everyone said the syntax cards and sound change cards come up too
> often relative to other cards in this deck, and the resulting
> too-frequent changes in syntax and lexicon make it harder to follow
> what's going on.

I admit it hadn't actually occurred to me that people who wanted to
_learn_ the gamelang -- as opposed to those who wanted some good
hearty diachrony -- would be annoyed by frequent changes in it, but
that's perfectly sensible. So, yeah, I suppose a drastically reduced
set of change cards is good for the nonlinguists' variant.

Sai was proposing this morning that we might unify _all_ the foo
change cards (sound change, syntax change, culture change, others?)
into a very lean set of change cards. This is something I'd hate to
have to do in the conlangers' variant, but on reflection it might be
appropriate for the nonlinguists' one (maybe culture change excepted,
'cause those they like), modulo that we don't want the two variants to
have incompatible decks.

> They suggested several
> ways to expand the deck and balance the proportions:

(Of course, that's the wrong way about the problem, but never mind.)

The polysemous word-coining cards are a good idea, though in an ideal
world we wouldn't have the bias that guessing charades introduces
against polysemy. Constraint cards are fine, and in an ideal world we
might as well have more. The free pass cards I'm less keen on: one
can get at most adpositions/conjunctions/abstract stuff by contrasting
pairs of phrases with and without the element in question, or simply
by giving several examples, and so maybe we should guide the players
more towards that. Two-related-word cards don't appeal to me enough
that I'd find adding them justified, given a player can always use two
turns.

And I guess this suggests revisiting the NSM words card. Just having
one of these, say, and listing all the NSM words in the rulebook for
players to choose (randomly or at will) a selection from to coin is
not a bad idea. I forget why we put it aside, other than the gain in
notional complexity.

> Alternatively, there could be a second mini-deck with just three cards
> marked Subject, Object, and Verb, which are laid out in a random order
> to start with. In the main deck there would be "change word order"
> cards to swap the order of any two adjacent constituents. (Maybe this
> mini-deck could have cards for Indirect Object and Locative/Temporal
> Complement, etc., as well...?) But that doesn't allow for secondary
> word orders for questions, subordinate clauses, etc., except with a
> syntax wildcard perhaps.

That's neat, except that secondary decks are a hassle, and I wouldn't
want to lose secondary word orders. (The other way to retain these,
of course, is to have several sets of S, O, and V in this minideck;
but that's bloated.)

> All the players liked the free-formedness of word-coining without
> reference to a set of phoneme and syllable structure cards in play.
> One said, "I feel like our language showed our personality more than
> when we had to put sounds together."

To my mind that settles it: drop the phonology cards outright for the
nonlinguistically inclined. Free-formedness isn't a quality that'll
be affected by whether we use IPA or (*shudder*) fauxnetic
approximations.

In other words: the experiment here
doesn't seem particularly worth it, at least on point (2), and
accordingly I suspect the way to go is:
with most of the sound change cards left out, perhaps one or two
retained, usable entirely free-form.
This is if we can't finagle things such that the simplified deck is an
honest subset of the sophisticated deck, and therewith release both
rulesets with the single deck from the beginning. (Of course even if
we manage this, the correct marketing strategy may be as you
describe.)

Re turn length:

> [...] pauses as someone composed a
If there's a single factor that you might want to bear down on with
all your streamlining zeal, Sai, this I think is it: it comes up
whether playing with conlangers or newbies, whether with the
concultural rules or without.

Alex

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Tue, Aug 26, 2008 at 3:48 PM
To: Alex Fink <000...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

No, not really, except that we're trying to reduce the deck
size.
......
I guess you could have

Isolating > Agglutinating > Fusional > Polysynthetic

although that oversimplifies the variation in real languages,
since agglutinating / fusional and isolating / polysynthetic
are distinct scales. The problem is expressing or implementing
fine grains of those orthogonal scales in game mechanics.

Maybe we could have a three-way scale, affixes forbidden /
allowed / required, and a two-way scale, affixes must be
agglutinative / affixes can be fusional, and two sets of
cards (min. 5) to place the game on those scales? That only
reduces the deck by one card, and makes it a bit more
complicated as well.

The Polysynthetic card has never been played in any
game I've been in, so I'm not yet sure how it will work
in practice.
There was, but it's already aged off and been deleted.
Darn ephemeral ZBB!

Last I looked at it (and I think I had it marked to email me
when it was updated, so I probably saw all the few posts
there were) it started with someone (Jeff Burke?) suggesting
a particular stress system card, and then someone else
asking about natlangs where said system occurred and
how they work, etc.
I don't see a good way around it, yet. When we played with
the older deck with phoneme cards and had a single copy
of the lexicon with words written in IPA, players complained
that this made it hard for them to use the language, both
because only one player could look at the lexicon at a time
and because they had a hard time reading the IPA transcription.

I guess we could suggest that a player coining a word should
not only say it aloud, but write it down (on a whiteboard if
available, on a paper to pass around to other players if
necessary). That might let the players work out a standard
orthography by consensus. But if that orthography is
heavily influenced by English, as it will be for nonlinguistic
players, it will still tend to be ambiguous; one player could
say /wU:p/ (rising tone) and write "woop", and another could
say /kulmlE/ (level tone) and write "koolmleh", and nobody
notice any ambiguity.
.....
Yes, I reckon I'll probably drop the "context" modifying phrases
from the sound change cards in this version. And reduce
the number absolutely, not just relatively.
Several players said the game tends to come to an end
just as it's becoming possible to say interesting things in
the gamelang, and suggested ways to build vocabulary faster,
which I noted in my last email. Yes, they definitely like
learning and using the gamelang; a little diachronic change
is interesting for them, too much is annoying.
How would that work? A few change wildcards that let
you do a sound change, syntax change or meaning
change at your option?
Given that the simplified deck is only about 50 cards,
expanding it is probably not a problem, if we end up
making two distinct versions. If we try to design the
advanced deck so you can make a simplified deck
by leaving out certain cards, then yes, we don't need
to add anything else, but how to manage it so that
you still have a well-proportioned deck when you
take out all the phonology cards?

> ..... .. The free pass cards I'm less keen on: one
Maybe the rulebook should have examples of specific charades
and contrastive sentences that could be used for conveying
abstractions?
The idea was to develop vocabulary faster so the gamelang
becomes usable/expressive faster. But I'm not sure it would
work; you might coin more words per turn but turns would
take longer, so maybe no net gain. Or maybe you would gain
a little with less turn-transition overhead?

On the other hand charading/drawing two related words is
liable to take less than twice as long as charading/drawing
one word, because you'll be e.g. pointing to constrasting
examples. (My brother used this technique Sunday evening
coining a word for domestic animals, saying /ZwAnje/ and
pointing to some miniatures of horses, sheep, pigs, and
then saying the same while shaking his head as he pointed
to miniatures of snakes, tigers, etc. If he'd had a two-related-
words card he could have said a different word pointing
to the wild animals and coined two words with no increase
in turn length. Games with conlangers tend to be more
interesting linguistically, but my brother's house is the
ideal game environment for nonverbally conveying the
meaning of almost any concrete noun you care to invent. :)
.......
So far I don't see a way to do it. In our talk about how to
reduce the advanced deck to a size that's economically
feasible to mass-produce, we've barely managed to
get it down to 110 cards with drastic sacrifices, and that
won't include culture change or NSM cards or an
expanded number of constraint cards that we probably
want in the simplified version.
We could encourage people to compose sentences before
their turn comes; but then you might have someone not
paying attention to someone else's charade or sound change
or conculture embellishment, while they're composing a
sentence they want to play on their next turn.

It was nice when we had a whiteboard available for the
afternoon game at LCC2; maybe we could encourage
people to write their sentences in large-print with
a thick black marker and hold them up as they say them
aloud. That might help a bit, with the sentence-composing
player having to repeat themseves less.

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Aug 26, 2008 at 11:36 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Alex Fink <000...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

1. re Alex on streamlining the guess-the-sentence phase:

This is true. I don't have any idea right now on how to do it; I imagine that it's something I'd need to play with on the fly in a test session.

Ideally turns should be <30 seconds or so in length; certainly not 7 minutes! People should get their turn again on a very regular basis, otherwise they will tend to get bored because they aren't "doing anything".

It may be helpful for the next session to keep track of how much time is spent doing what exactly - e.g. discussing rules, waiting for someone to form a sentence, analyzing the sentence, looking up preexisting rules/phonemes, doing the charade itself, etc - so that we know what particularly the sticking points are. Most likely, possible resolutions would show themselves in the process.

2. Phonology cards etc

What of just ditching them altogether in favor of wildcard "make a sound" cards?

I agree that visually, seeing the IPA arranged there is nice. But in terms of mechanics, it's pretty bad as a constraint method and entails weird distributions.

Another method you could use is simply allotting each person a certain number of phonemes - either as chips to start or in the form of received cards - and instead of playing them as such, they are simply used up when making a new word.

I don't know how well that latter tweak would work honestly - my concern is that non-linguists especially won't even notice that they've just introduced a new sound - but it'd require testing.

- Sai

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Wed, Aug 27, 2008 at 8:07 AM
To: Sai Emrys <s...@saizai.com>
Cc: Alex Fink <000...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

In practice one is often "doing something" during another player's turn;
guessing at their charade, having conversation about their
concultural addition, editing your copy of the lexicon after their
sound change, etc. I didn't notice anyone getting bored between
turns last Sunday. I agree that 7 minutes per turn is too long;
undoubtedly playing again with the same players and a stable
set of rules, soon enough that they don't forget the rules and
have to have them re-explained, would reduce the time by
2-3 minutes at least. We might have to fiddle with the rules
to reduce the time much more than that. < 30 seconds average is an
unreasonable goal, however, IMO; even if you're doing
a totally obvious charade like holding up three fingers while
saying your word for the numeral "three" and another player
guesses it instantly, you first have to draw a card, figure out
whether to play the card you just drew or some card you already
had in your hand, explain what you're doing with the card
in many cases (secondary word order, sound change, extend
meaning, constraint...) and, if we're using the conculture
rules (they were VERY well received Sunday, and I suspect
would be well received by most conlangers as well),
tell the other players what detail you're adding to the culture
and answer their clarifying questions about it, if any.
30 seconds is a minimum turn length if you play a simple
card that doesn't require much thought or explanation
and do an obvious charade that is guessed instantly.
A more reasonable goal for *average* turn length would
be about 2-3 minutes, considering some early turns where
people are doing obvious charades for concrete nouns
and verbs and aren't composing sentences yet will
be shorter than that, and most of the later turns where
people are coining more abstract words and composing
sentences on their turn will be longer than that.
I'll try to keep more exact track of time from when we sit
down to play and start explaining the deck and rules, to
when we start playing, and to when the game ends; but
trying to keep track of the exact amount of time spent within
each turn on each element of play could slow the game
down as much as any other factor... I'll see what I can do.
And y'all do what you can to collect such data at the
next Bay Area conlang meetup game, maybe?
About how many relative to the size of the deck (I think 110 cards total
is our target)? Would it make sense to have some totally open wildcards
and some that are partially constrained, e.g. "add a fricative", "add
a plosive", "add an alveolar phoneme", etc.? Not in the simplified
deck, obviously, I mean in the advanced deck. If they're too few
relative to the size of the deck, it will be frustrating in early rounds
when few or no phonemes are available yet, and we'd need some
equivalent of the current quick-start method where you lay out
the first two consonants and first two vowels that appear before
you shuffle the deck again and deal. Maybe before the game
starts, go round once and every player contributes one phoneme
to the initial inventory, and after that you add phonemes only with
"add phoneme" cards?

Would we still have Phonemic Contrast cards in this version?
I'm not sure they wouldn't help.
So when you coin a word that contains a phoneme that hasn't
appeared in previous words, you expend a card/chip/token of some
kind, but if you coin a word containing only phonemes that already
appear in existing words, you don't expend any chips?

Maybe the phoneme inventory limit could be implemented by
rolling two or three dice, multiplying the result by the number of players,
and distributing that many phoneme tokens at the start of the
game? -- Probably you would add a certain figure to the die roll
before multiplying by the number of players.

How would syllable structures/phonotactic restrictions work in this
version? The same as at present, or would that also be managed
with tokens of some kind?
During our discussion after the game Sunday I suggested closing the
phonology after the first N rounds, so after that point you have to
coin words using sounds that already occur in existing words. The
playtesters nixed that idea, saying it would make it harder to coin
words, slowing things down and reducing spontaneity.
A couple said they were trying, after the first few rounds, to
make words that were consistent in sound with words already
coined, but didn't see a way to enforce that without impairing
playability.

--

----------
From: Alex Fink <000...@gmail.com>
Date: Sun, Aug 31, 2008 at 3:20 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

2008/8/26 Jim Henry <jimhen...@gmail.com>:
>> What would the [typology] scale look like, for this?
One could reduce the specificity of these cards too. "Move the game
typology to an adjacent point on the (/ a) typology scale", leaving
the direction to the player's choice.
Huh. I'll pay especial attention if it comes out in our games.
Did you retain the players' lexicon sheets from that game? I'd be
interested to know whether there was a lot of variation between them,
and what its nature was. My guess would be that what variation there
is would be partly asystematic: maybe someone consistently records [x]
as <kh> and someone else as <ch>, but there are also likely to be
cases where individual people vary among say <u> and <oo> for [u]
without a lot of principle. The lack of a standard orthography
needn't be a problem, I think, if all of the individual orthographies
work, though it would probably help keep people consistent and honest.

In addition there's the hazard that naive users of English-based
orthographies might feel the need to work around things that are
purely irregularities in English itself, giving another source of
imprecision. E.g. even if <oo> was consistently [u], one might have
compunctions at recording say [fut] as <foot> because that reads back
as [fUt], and so deviate from this; or conversely one might blithely
write <foot> and later read back [fUt].
I think that was the idea. I was expecting him to elaborate.
Good.
Only negligibly much, I'd think, if the two words were gestured
separately. But I think I missed the critical part of what was meant
by "related":
which is a very natural environment for defining two words. Enough so
that it might be worth considering allowing on every turn, not just
regulated by a card? Say, if after your main performance, you can
continue it / point at something else in the picture / etc. to convey
another word in some very short time limit (five seconds?) you're free
to make that word as well.
A definite plus! One to mention in the manual, I think.
I suppose that's true, as we currently stand. Ah well.
Right, I don't think this is something we should encourage too
strongly. Especially given that what goes on in other players' turns
has some engagement value; this seems like it might drive that
engagement out, as people start composing during spare moments and
player interaction decreases.
Yeah, that would help, for people who have such things to hand.

2008/8/27 Jim Henry <jimhen...@gmail.com>:
Yeah, we will. It would be useful to have a recording device of some
sort running alongside, if the technology is feasible and we have the
patience to go through it afterwards.
For the simplified set, no, I still favour ditching them altogether full stop.

For the advanced set this might be possible but I'm not sold on it.
We haven't yet tried the 30ish common phonemes plus wildcards option,
which may prove to eliminate the worst of our problems with weird
distributions. And I don't understand why you say individual phoneme
cards are bad as a constraint method.
Yes, this is a good patch for a variant where phoneme cards are scarce
but required.

Something else to watch out for in that circumstance is the limit it
puts on frequency of playability of certain classes of sound changes,
though, and this especially so if phoneme cards don't exist, or
contrast cards don't (unless sound changes work by a different
mechanism as well).
Yes, that's how I understand it; also one chip per phoneme, if there
is more than one new phoneme. Incrementally faster, I guess, even if
the version with cards is pretty similar in effects.
Two or three D6 per player? That's a larger limit than your other
dice rule, 2D6+8 per game, especially so if there are many players.
I suppose they could be managed with tokens, one for each expansion of
the syllable structure beyond CV, with the definition of one expansion
roughly as on the cards now. But this loses yet another thing that in
the current version you can see arrayed on the cards in front of you;
the syllable structure would have to be kept track of on paper too.

Other phonotactic constraints aren't really handled yet in the
card-based scheme, if I'm not mistaken. The trick with them is that
they're most naturally stated subtractively, as sequences which are
ruled out, which requires a bit of delicacy given that we're building
everything up additively. Abstractly, though, I see no reason you
couldn't be enabled to impose one, or pay for imposing one, with a
card or a token.

Phonotactic constraints often arise as the aftermath of sound changes
-- whatever was the triggering context for the change can no longer
occur, and if it would it gets corrected per the change, which can
stick around in the phonology for a while on some occasions. Is it
worth trying to emulate this?

Alex

----------
From: Sai Emrys <s...@saizai.com>
Date: Sun, Aug 31, 2008 at 4:55 PM
To: Alex Fink <000...@gmail.com>
Cc: Jim Henry <jimhen...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

I don't really have anything further to say on that...
What about the problem of L1 calquing? Do we care?
We should try, then.

Constraint: because essentially people will wind up with the "wrong"
phoneme cards, i.e. ones they're not interested in.

Do we really want the phoneme set to be decided by random distribution?

I would rather see it distributed in some more intentional fashion,
possibly with the sort of series-encouraging meta-phoneme-generation
cards I suggested (e.g. 'same as in play, plus velarized' or 'plus
vocalized' or whatever).

As is, there's no reason for someone to *not* play a phoneme card,
even if it results in something silly. Silliness is fine of course,
but it should be intentional.

- Sai

----------
From: Alex Fink <000...@gmail.com>
Date: Mon, Sep 1, 2008 at 1:58 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

So, we're aiming for a playtest this Saturday, as you may've seen.

We've yet to make a deck. What deck composition (phonology cards in
particular) would it be must useful for us to be testing?

Alex

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Tue, Sep 2, 2008 at 9:46 AM
To: Sai Emrys <s...@saizai.com>
Cc: Alex Fink <000...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

Maybe. And I've thought of some ways to graduate the agglutinative/
polysynthetic scale a bit more, though they're maybe too complicated;
e.g.

"If the previous two players both coined root words, you must coin
an affix."

"If no affix has been coined yet this round, you must do so."

etc...?
In another game with the simplified deck this weekend, the players
gave a spelling for their word after they said it aloud. I in addition
wrote an IPA transcription of each word in my own copy of the
lexicon. I noticed that even then, words tended to get variant
pronunciations based on ambiguity in the Englishesque orthography;
e.g. /tIsi/, written <ticcy>, was sometimes pronounced /tIki/
I kept only one of their lexica, the one that had the goals listed on it.
The original draft version of the rules allowed that; some people
objected to it as creating a greater imbalance between novice
and experienced players. I'll consider reintroducing it as a general
rule after 2-3 more games using the various "extra word" cards.
The game last Saturday ran about 50-55 minutes and was
interrupted when some players got tired and wanted to go
to bed; it lasted 15 turns. And there was a longish interruption
in the middle, too, so probably the average turn length was
under 3 minutes.
Yes, obviously if we make any extensive changes to the phonology
cards, we need to revisit the design of the sound change cards too.
It has a larger spread of possible values, the lower bound being lower
than in the old rule and the upper bound being much higher. The
rule probably needs to be sensitive to the number of players, e.g.
if you have more than 4 players roll 1 die, if 2-4 players roll 2...?

Or just roll 2d6+8, then round off to the nearest number that's a
multiple of the number of players?
Yes, we need some way to impose phonotactic constraints;
either with cards or tokens. (Could the same set of tokens be
used for both? That is, if you get 7 tokens at the start of the
game you could make 7 phonemes or 4 phonemes and 3
phonotactic constraints or whatever?
If we stick with cards, maybe you can play a phonotactic constraint
in tandem with a sound change card and draw another card at the
end of your turn...?

Probably either the 30 common phonemes plus about five general wildcards
and five Fill Gap cards, or the version with no phoneme cards and chips
distributed to each player giving them the right to add N phonemes at
various points in the game. -- And maybe Phonemic Contrast cards
as well, for either of those variants.
My experience of three games with the simplified deck, with nearly
disjoint sets of players, is that even monolingual players will sometimes
violate their L1's phonotactic constraints when making Glossotechnia
words, and (less often) introduce phonemes or phonemic distinctions
not made in their L1. And I don't see how a set of wildcard "add a sound"
cards would help any; the players could end up using a subset of their
L1's phoneme inventory.
So maybe the phoneme-adding tokens plus a set of Phonemic Contrast cards?
And similarly for syllable structures and phonotactic constraints?

--

----------
From: Alex Fink <000...@gmail.com>
Date: Fri, Sep 5, 2008 at 1:40 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

This is an informational message on the subject of most frequent
phonemes, which I was looking into for the purposes of compiling the
deck we're going to use for the Berkeley playtests. Thankfully the
UPSID survey exists and makes this much easier: for each phoneme it
records, it knows the proportion of its 451 languages that contain it.

For consonants it's got the irritating feature that dentals and
alveolars and unspecified dental/alveolars are all counted separately,
though. I've corrected for that by taking the unspecified counts and
multiplying those by 14/5, and discarding the other two sorts -- this
is indefensibly hacky, when I could've done the summation, but it was
quick. That gives the following top of the frequency list (warning,
monospace table ahead):

n .9935 g .5610 k_h .2284 dz) .1240
t .9436 N .5255 p_h .2239 G .1220
m .9424 ? .4789 r* .2234 c .1197
k .8936 tS) .4169 v .2106 B .1197
l .8445 S .4146 x .2084 q .1153
j .8381 f .3991 4 .1613 tS)_h .1131
s .8381 r .3167 ts)_h .1551 b_< .1086
p .8315 J .3126 t_> .1490 mb) .1064
w .7361 t_h .3041 K .1490 ts)_> .1056
b .6364 ts) .2794 k_> .1397 nd) .1056
h .6186 z .2669 Z .1353
d .5650 dZ) .2506 k_w .1330

and no other sounds in more than one language in ten. r* was glossed
in the list as "voiced dental/alveolar r-sound", whatever we make of
that. Also, several of these things are not atomic in the
Glossotechnia analysis: affricates, aspirates, implosives, ejectives,
prenasalised stops, ??voiced obstruents.

For vowels the parallel irritation is that e.g. /e/ and /E/ and
indifferent /e/~/E/ are counted separately; I've corrected (slightly
less undefensibly) by dropping the indifferents and multiplying the
others by 11/7, but special-cased /@/ and left it alone. This gives

i .8714 I .1641 a_": .0754
a_" .8692 U .1463 e: .0732
u .8182 1 .1353 e~ .0627
E .6481 E~ .1219
O .5645 O~ .1116
o .4565 o~ .0941
e .4320 M .0909
a_"~.1840 i: .0887
i~ .1818 & .0865
@ .1685 o: .0836
u~ .1641 u: .0798

and no other sounds in more than one language in sixteen. My /a_"/
was just written /a/ in the UPSID but they called it unambiguously a
low central vowel (this is a hole in the IPA more than anything).

A defensible set for a 30-ish common phonemes deck might thus be
/m n J N p b t d k g ? f v s z S x h l r j w/ (total 22)
/i 1 u I U e o @ E O a/ (total 11)
provided that has the right ratio of Cs to Vs.

Alex

----------
From: Alex Fink <000...@gmail.com>
Date: Fri, Sep 5, 2008 at 2:36 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

2008/9/2 Jim Henry <jimhen...@gmail.com>:
Probably too complicated, and for that matter I don't know what you'd
call them. "Kinda synthetic"?

Sai was arguing last night for the removal of the agglutinating /
fusional dimension, saying that fusional is a strictly weaker
restriction in play than agglutinating (you can still make a
sandhi-less affix), and that he wasn't sure that it had enough
influence on the flavour of a language to be worth modelling. At
first he thought it was something players wouldn't take advantage of
anyway -- and, if our guiding example is contact sandhi phenomena,
this is probably largely the case -- but I reminded him that things
like ablaut and Celtic-style mutation (and templatic morphology?) are
cases of fusionality, and then he wasn't so sure.
Ah, too bad.
>> regulated by a card? [...]
Point, assuming that being able to create two related words on your
turn often is that much of an imbalancing factor, and that it's
something inexperienced players wouldn't learn to do (in at least
simple cases). We may as well try this variant here, then, unless one
of us objects.
Good; that's more spirited.
In particular the mean is higher and significantly so except in a
two-player game, where it's nearly the same. And not many two-player
games have happened so far, though in theory they could.
The latter is best inasmuch as the phoneme limit isn't too dependent
on the number of players, I think. Maybe the initial roll should be a
little higher, though.
(I don't see why not.)
What would playing a phonotactic constraint consist of? Are you
envisioning the chips mechanic?

The most "natural" implementation is to say that, on some occasions
(perhaps at the player's discretion, perhaps at the cost of a chip?),
after a sound change is played the sequences that were subject to
changing remain phonotactically illegal (or are still repaired by the
rule, I don't know whether this makes a difference). The problems
with that are that one has to write down the phonotactic constraints
in force, since they're not represented on cards, and that there's no
evident way to get rid of one.
Per my other message I'm going for phoneme cards in this style; the
rest of the deck is prepared already, culture cards aside. Still
undecided whether we'll use them or chips. In either case we'll have
phonemic contrast cards.

Oh, by the way, there should be a phonemic contrast wildcard: there
are plenty of potential contrasts that aren't and shouldn't be
individually in the deck (prenasalisation, velarisation,
pharyngealisation, retroflexion on vowels, implosion, breathiness,
creakiness, stridency, ATR, ...)

Alex

----------
From: Sai Emrys <s...@saizai.com>
Date: Fri, Sep 5, 2008 at 9:47 PM
To: Jim Henry <jimhen...@gmail.com>

I would suggest that the game should strictly require people to stay
in one mode - be it speech, or writing, or signing - else this sort of
calque is all too likely. :-/
Or you can use one as a sound change - e.g. saving up for being able
to change stuff around later. However, Alex points out that this would
restrict sound changes over time, which seems a valid point... so
perhaps some other method would be appropriate. E.g., every (# players
+ 1) turns since the last sound change, one is allowed to make a sound
change (if one has remaining tokens).

One plausible way to do chips is to decide on an a sane range of
phoneme count, and fix the high end of that count as the number of
multi-purpose phonology tokens, which are distributed evenly.

There could be conditions that allow one to reclaim a token - e.g. if
another player ousts one of your phonemes via sound change.

- Sai

----------
From: Sai Emrys <s...@saizai.com>
Date: Fri, Sep 5, 2008 at 9:47 PM
To: Alex Fink <000...@gmail.com>
Cc: Jim Henry <jimhen...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

----------
From: Sai Emrys <s...@saizai.com>
Date: Sat, Sep 6, 2008 at 6:55 PM
To: Greg Shuflin <greg...@berkeley.edu>

----------
From: Paul Schleitwiler, FCM <pjschlei...@gmail.com>
Date: Sat, Aug 9, 2008 at 11:27 AM
To: Sai Emrys <s...@saizai.com>

----------
From: Sai Emrys <s...@saizai.com>
Date: Sat, Aug 9, 2008 at 12:33 PM
To: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>

----------
From: Sai Emrys <s...@saizai.com>
Date: Sat, Aug 9, 2008 at 12:35 PM
To: Jim Henry <jimhen...@gmail.com>, Matthew Haupt <film...@hotmail.com>

Jim, Matt -From: Jim Henry <jimhen...@gmail.com>
Date: Sun, Aug 10, 2008 at 5:14 PM
To: Paul Schleitwiler <pjschlei...@gmail.com>
Cc: Matthew Haupt <film...@hotmail.com>, Sai Emrys <s...@saizai.com>

--
Jim Henry
----------
From: Jim Henry <jimhen...@gmail.com>
Date: Mon, Aug 11, 2008 at 11:30 AM
To: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, Matthew Haupt <film...@hotmail.com>

----------
From: Sai Emrys <s...@saizai.com>
Date: Mon, Aug 11, 2008 at 11:53 AM
To: Jim Henry <jimhen...@gmail.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Alex Fink <000...@gmail.com>

- Sai

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Mon, Aug 11, 2008 at 12:23 PM
To: Sai Emrys <s...@saizai.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Alex Fink <000...@gmail.com>

Yes, right. more sentences when you've used up the ones that come
with the game.
Any ideas about how to do that?

--
Jim Henry

----------
From: Sai Emrys <s...@saizai.com>
Date: Mon, Aug 11, 2008 at 1:07 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Alex Fink <000...@gmail.com>

'flaws', like http://en.wikipedia.org/wiki/GURPS#Advantages_and_Disadvantages handholdingy examples of how to make a target sentence per my mechanic
#5. individual win. That is, each playthrough would be essentially a
trick. [in the Hearts etc sense]

- Sai

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Mon, Aug 11, 2008 at 6:04 PM
To: Sai Emrys <s...@saizai.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Alex Fink <000...@gmail.com>

And there would probably be arguments about whether
a particular word counts as abstract or concrete... Yes, absolutely.
Details? Example?

* make a sentence which explains why you deserve to win this game of
Glossotechnia

Alex

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Tue, Aug 12, 2008 at 10:14 AM
To: Alex Fink <000...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

<snip long explanation> translating the challenge sentences; this achieves that goal
in spades. other players get on how it works, even if
some of the sentences produced don't
directly help you win. That seems a bit too vague... needs work. >> it with a role-playing game. We still haven't figured out a way

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Tue, Aug 12, 2008 at 10:21 AM
To: Alex Fink <000...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Aug 12, 2008 at 4:53 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Alex Fink <000...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

card drawn has on its bottom the group target (& how many points it's
worth & a citation)
- Sai

----------
From: Alex Fink <000...@gmail.com>
Date: Tue, Aug 12, 2008 at 4:57 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

2008/8/12 Jim Henry <jimhen...@gmail.com>: meant that to be the case. of having both group challenges and hidden individual challenges
was/were. Your extension to the constraint game seems to be the natural one as well. 2008/8/12 Jim Henry <jimhen...@gmail.com>: current deck at all. I'd like them to be, but I don't have a good
implementation yet.

Alex

----------
From: Alex Fink <000...@gmail.com>
Date: Tue, Aug 12, 2008 at 5:31 PM
To: Sai Emrys <s...@saizai.com>
Cc: Jim Henry <jimhen...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

2008/8/12 Sai Emrys <s...@saizai.com>: > * bonus points: mechanic being irrelevant, but this seems too far in the other
direction. [...] > Benefits:
Alex

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Aug 12, 2008 at 6:16 PM
To: Alex Fink <000...@gmail.com>

1. Rulebook
2. Optional rule: everyone adds one constraint/goal as a one-round setup redundant. But I think it's a more natural way to think of the
requirement. kind of goal. I would suggest that some (eg metas) be only once, that's inter-sentence; all scoring in my system is based on the
current sentence.
Sure. design, opinion of the language. ~

----------
From: Alex Fink <000...@gmail.com>
Date: Tue, Aug 12, 2008 at 6:54 PM
To: Sai Emrys <s...@saizai.com>
Cc: Jim Henry <jimhen...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

non-cumulative. That's not the case with my rules. project the current situation to the end of the game for the
accounting.
Sure.
[...]----------
From: Jim Henry <jimhen...@gmail.com>
Date: Thu, Aug 14, 2008 at 5:37 AM
To: Alex Fink <000...@gmail.com>, Sai Emrys <s...@saizai.com>
Cc: "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

Good. yes. See below. The others, I'm not keen on. It seems that any that are against the spirit of the game. And maybe there should That's probably good. Good. ...... > 2008/8/12 Jim Henry <jimhen...@gmail.com>: I agree. when players make up their own from scratch.
That makes sense. far too much, and I would probably agree, but sometime or----------
From: Alex Fink <000...@gmail.com>
Date: Fri, Aug 15, 2008 at 2:55 PM
To: Jim Henry <jimhen...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>

2008/8/14 Jim Henry <jimhen...@gmail.com>: [...] Right, good. words -- what ever it is, it should be quickly computable. Then, if a * To increase the competition, maybe each boolean constraint carries a
bonus for the first to achieve it. they should just be an additive bonus.
That's a good idea. [...] Certainly you could. [...] Right. It's unspecified how to apply those Action cards in the >> Sai's message:
[...]
You've done this, yes? beyond the NSM in some places where their hypothesis has led them to
be parsimonious (e.g. they've got 'can' but no other modals, alethic
or otherwise).----------
From: Jim Henry <jimhen...@gmail.com>
Date: Fri, Aug 15, 2008 at 6:05 PM
To: Alex Fink <000...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>, Emerson Knapp <enze...@gmail.com>

So what could we add to make sandhi more common?

[Scoring:] points as it has words. Say the two sentences share N words; should "Man bites dog" or "Dog bites child" get more
points? It seems to be the best we've come up with so far.
Why only boolean constraints? It's already one of the most verbose cards in the deck. small-print and hard to read though. -- Maybe these alternate That makes sense. the less experienced players with suggestions. And I think Yes, that's clearer.
Right.
OK.

...

[Message clipped]
----------
From: Alex Fink <000...@gmail.com>
Date: Sun, Sep 7, 2008 at 1:19 AM
To: Sai Emrys <s...@saizai.com>
Cc: greg...@berkeley.edu, Jim Henry <jimhen...@gmail.com>, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

We've just finished a playtesting game of Glossotechnia in Berkeley,
with me and Sai and Emerson and Greg Shuflin (now on the cc list) and
Chip lnu. On the whole we were a linguistically well-versed group,
Chip perhaps somewhat less so than the others (he characterised
himself as "an auxlanger, not a conlanger"?). We used a ruleset and
deck incorporating several changes from the revision discussion going
on offlist, the deck perhaps 130 cards in total. All told the game
went well, though it began to drag at the end, and we didn't perhaps
have a good test of everything we meant to.

The deck had a slimmed-down phonology section, 33 common phonemes plus
five each of wildcards and fill gap cards. This worked pretty well
for naturalising the phonology: the consonant inventory we ended the
game with,
/m N p b t d k g s z K h l\/,
was aside from the missing /n/ nothing I'd bat an eye at finding in a
natlang. No-one used any of these cards on the vowel system, though,
and by consequence it ended up a little more off, /a E U 1/ with three
degrees of length.

We used free-form typology cards that let the player move the typology
one unit on a two-dimensional grid, one dimension running from
isolated through synthetic to polysynthetic, the other fusional vs.
non-fusional. We had thought to start isolating but after some
discussion changed this to non-fusional synthetic as a more neutral
starting point, neither the creation of words nor that of roots
regulated. The synthesis scale did its thing quite nicely: we moved
to polysynthetic about a third of the way through the game and then
dropped quickly to isolating in the very late game. Sai did this last
and proclaimed that whether each given affix was lost entirely or was
reinterpreted as an independent word was to the discretion of the
player who first tried to use it; this kind of thing is probably good
general procedure, in that it avoids the bottleneck of going through
all the affixes one might otherwise get when switching topologies.

By contrast the fusionality scale didn't seem to excite anyone's
interest, and it never moved, with only one fusional affix being added
and this by Chip's use of a Phonemic Contrast card (length for the
continuative), which use we didn't intend them to have.

As the game moved on a number of words were coined entirely in the
game language (this was worth points). In particular, when this
happened the coining and sentence-constructing phases of the turn were
naturally combined into one, which sped things up moderately, and in a
few cases led to a certain bleed-over from the "other players'
collective interpretation is binding" take on charades to new
morphosyntactic constructions (so there ended up being a null
causative, for instance). We didn't include the "coin two words"
card, on my oversight. A few borderline cases of coining two words on
a turn happened anyhow, one fresh coinage and one a derivation or a
compound.

Favourite lexical moments: the word /atatakaEkata/ evolving from 'see
something far away' to 'look forward to' to the future tense
auxiliary; the reanalysis of a set of existing forms for 'food in a
bowl', 'food on a plate' to a general derivational process (change all
vowels to /a~/) for 'food in an N'. Greg liked /l\ak1/ 'punch' enough
that he borrowed it as /raki/ for his own conlang, of name I didn't
catch.

Feeling that everyone having to make a culture change statement each
turn would be too much, we decided that after one statement each in
the initial round we'd not make any more unless they were the
sentences in the game language (worth points) or unless a Culture
Change card came up. This worked well, though I thought that it
really ought to be accompanied by a couple less tumultuous culture
cards than War and New Technology and the like, that just let you make
an innocent statement.

Highlights from the culture were the fact that the game language was a
minority language spoken under an oppressive and hated majority. Two
contrasting plural suffixes, for homogeneous and non-homogeneous sets,
took on fixed meanings respecting the "us vs. them" distinction.
Later, after an incidental comment about sheep, the oppressors quickly
became sheep, and punching up and disparagement of sheep became
standard sentence fodder. The reason for the oppression was
supposedly the speakers' exceptionally long lifespan of 5 years,
though not much was ever made of this. The "War" card was used to
introduce a secondary word order, OSV, putatively borrowed from a
secret military code to be used only in war-like contexts, and from
there generalised to be the word order used when angry.

We played that syntax cards could be used as secondary word orders in
every circumstance, no card required, which seemed to work fine to me,
although it meant we never actually lost our initial primary word
order completely (it calcified to a morpheme order in the verb at one
point). Though see below for a dissenting opinion.

Emerson left early, and Chip came late -- which is something the game
doesn't really accommodate comfortably, there being a lot of lexicon
and corpus and culture to take up all at a stroke. The game was in
progress for around 2 hours 45 minutes. That doesn't count time to
explain what was going on, except for small points we missed on first
presentation of the rules and a lightning explanation for Chip, but it
does count a couple of periods, maybe 30 minutes in all, where we were
discussing people's thoughts on the current rules and possible
changes. Judging by the lexicon there were 34 turns, meaning an
average turn length of about 4 minutes.

What was especially notable -- and noted -- by everyone about the turn
length was that it was essentially linear in the number of turns so
far. Sufficiently early in the game that sentences weren't yet being
composed, turns probably weren't more than a minute, meaning that in
the late game they must've been 7 minutes or so with regularity. This
we think was attributable mostly to composing sentences, especially to
having to find the words wanted in the lexicon and check for all the
appropriate syntax rules, both of which also increased linearly. We
really do want to find some way to trim this. It was suggested that
the lexicon might be kept more organisedly -- nouns here, verbs there,
etc., and within say the nouns food here, body parts there, etc. But
we expected the effect of that to be small.

As for the scoring. Following the revised ruleset, we came up with a
selection of goals at the beginning of play. People's attention
didn't seem especially to be on the goal-creating process, but I
suspect this is muchly because we'd just explained the rules of the
main game and everyone was eager to start. In particular we gave each
of the goals the same point value. We chose to use a much reduced
base value scheme as well, not having worked the kinks out of anything
fancier: one point, plus another if there was some syntactic
innovation.

That was in theory. In practice it became evident that no-one was
much interested in the score at all, aside from the understanding that
it was better to make a sentence than to not, and to do something
interesting than to not. Not having kept explicit running totals for
everyone may have been at fault for this (I had been thinking that I'd
just let notations next to each sentence on the corpus page stand in
for this). The selection of goals may have been partly at fault too.
We had:

* several phonological goals which may have been too niggly to pay
attention to without more command of the language (_longest
palindrome_, _most different phonemes_, _sentence-internal rhyme_) --
granted, some attention was paid to the palindrome one
* a syntactic goal with the same probable problem (_garden path sentence_)
* _completely head-final_, which had the difficulty that there wasn't
likely to be much variation in head-finality within a game language:
either most sentences would be or most wouldn't. As it happened our
basic word orer for most of the game was VSO so we didn't get this.
* _sandhi_, which I got
* _mention a season_ and _say something about the food_, which were both gotten.
plus the "standard" goals of playing one's word-coining turn in the
game language, and saying things about the conculture.

Everyone thought that a major thing this game was lacking was a single
preeminent goal, something to end the game by. It was proposed that
the achievement of every goal at least once bring about the end of the
game, but we never achieved several of them. No, what was most
approved of (and I think suggested independently of our earlier
discussion) was ending the game on translation of a passage, with each
of its sentences being a scoring goal. We did like the idea of the
other goals, so those could easily be retained as well, they'd just
differ in not being relevant to the ending condition.

It was suggested at one point that, for greater competitiveness, there
be private goals as well. This is really hearkening back to
Glossotechnia as it were, and we pointed out the known problems with
this... but?

(I hope I'm not misrepresenting anyone in what follows. Players, feel
free of course to provide your own comments.)

Greg liked the fact that, since the LCC2 games, the rules and deck
composition had changed to give more room for tinkering with the
language and doing interesting things with the morphosyntax. In
particular he didn't remember there being much opportunity for
creation of morphology at all in that game; this one was better in
that respect.

Chip took the opposite tack on this. His favourite part was the word
creation, and he found such things as secondary word orders to quickly
get distractingly and tediously detailed. (He reckons this a
difference between the conlangerly love of tinkering and his own of
vehicles for communication.)

The phonology he thought to be the easiest to understand, and do the
most for the cohesion of the language; he was surprised to hear that
your group found it probably the most constraining. On these grounds
it would be worth trying the phonology chips approach with your group,
in case that avoided the difficulty and sterility they perceive in
phoneme cards, if it weren't for the fact they'd still have to know
what a phoneme is...

There was some discussion of the viability of targeting the game
towards two (or more) different groups altogether. I'm increasingly
feeling this is unavoidable, unless we simply want to drop one of the
target audiences, but Chip thought that we'd better focus on one
group. In any case he did like the potential adjustibility of the
game by omitting subsets of cards as desired: so you might play with
no syntax cards if you're using G'nia as a teaching aid and haven't
gotten to syntax in the course yet, or if you don't like syntax
changes happening in game then dealing one card of each type from this
separate deck to be the syntax of the game language at the very
beginning of the game, and setting all the others out of the game.

He thought that having to rewrite all the old sentences on change of
syntax would be far too much effort -- that's probably right -- and
that discarding them would be a waste -- ditto -- but I didn't get his
reaction to just leaving them and claiming an archaic form of the
language, which seems a sensible thing to do.

And there's doubtless some of the discussion I've forgotten to mention here.

Alex

----------
From: Jim Henry <jimhen...@gmail.com>
Date: Tue, Sep 9, 2008 at 8:47 AM
To: Alex Fink <000...@gmail.com>
Cc: Sai Emrys <s...@saizai.com>, greg...@berkeley.edu, Emerson Knapp <enze...@gmail.com>, "Paul Schleitwiler, FCM" <pjschlei...@gmail.com>, Matthew Haupt <film...@hotmail.com>, Parker Bohn <mrpark...@gmail.com>, Brian Henry <jati....@gmail.com>, "Schutrick, Carrie A." <cschu...@yahoo.com>

>>> 2008/8/26 Jim Henry <jimhen...@gmail.com>:
It may be more of an imbalancing factor in the version with fixed
translation challenges than in the version with a variety of goals
decided on by the players at the beginning. That is, with translation
challenges having a specific set of words in the language is
a more critical prerequisite than it is for many of the other kinds
of goals one might be working toward; and in the more open-ended,
varied-goals version you are more likely to be able to find a use
for words the other players come up with in your own sentences
than in the fixed-challenge version. So rules that allow players to
come up with multiple words per turn, either any time or with
special cards, thus tending to grow the lexicon a bit faster,
are likely to benefit other players besides the ones who make
multiple words.

If there are cards that allow coining multiple words, there should
be enough of them in the deck that all players are likely to
come up with one at some point. I think I have about 4-5 of
them in my simplified deck, but they haven't been used yet.

>>> 2008/8/27 Jim Henry <jimhen...@gmail.com>:
Maybe 2d6+13 (15 to 25 phonemes), or 3d6+10 (13 to 28)?
.....
I said "if we stick with cards"; that is, if we use cards for
phonotactic constraints (probably some specific like "clusters must have
same voicing" or "clusters must be at same or neighboring point of
articulation", and some wildcards). But probably your idea of using
chips for phonotactic constraints as well as phonemes is a good one.
That makes sense.
Yes, definitely.
Are you serious? I doubt many players have the memory to
play a game of Glossotechnia without writing things down,
-- although it would be an interesting challenge to use G'technia
rules and deck to create a purely written language, still I think
the current mode of creating a spoken language with an ad-hoc
orthography (based on IPA or English according to the players'
knowledge/inclinations) is the easiest to work with, in spite of
problems like those I mentioned.
What about the idea of deciding how many chips are distributed
at the start of a given game with a dice roll, as Alex and I have
been talking about?
...Or maybe said tokens are garnered according to one's score
so far? Every N points one accumulates by making scoring
sentences can be used to get another token? In which case,
the number of chips distributed at the beginning might should be
lower than you would otherwise want.
I occasionally see that constrastive usage on the AUXLANG list
(or used to; I haven't been active there in a while), esp. re:
people who are interested in / promote one or more auxlangs but
don't devise conlangs (of any sort) themselves.
How many such were in the deck? How many came up in play?
Conculture embellishments on every turn do slow the game down.
I suppose if you're going to limit them on rounds after the first, you
would want a higher proportion of culture change cards in the deck,
most of them low-key and a few revolutionary.
What about allowing players to play syntax cards either additively
or substitutively on any turn? -- so for instance, one person could
play VSO as primary word order, then another plays SOV for
questions, then someone else with a SVO card could either add
a third conditional word order or substitute it for either of the other
two orders?

And if you make it easier to add secondary word orders, maybe
you should increase the number of "discard anything" cards
in the deck, or add specific "eliminate word order" cards.
That seems to fit my experience too, though I haven't timed the
turn lengths in any of my games.
Some players did that in the second game with the simplified deck.
I think those cards were introduced between the noon game
(which Greg was part of) and the evening game, by which time
I think he'd gone home.
One of the players in the second game with the simplified
deck had that reaction when we had four different word orders
in play at once. And in the third game (with a different set of
players) a couple of people had Secondary Word Order
cards and discarded them rather than use them.
I'll propose that next time we play.

I think I mentioned in earlier messages that in the first couple of
games with the simplified deck, players sometimes interpreted
the word "sound" in the sound change cards to refer to
a syllable rather than a phoneme. (In the third game the players
asked to leave the sound change cards out of the deck.)
Or save it for later.
That's what we've been doing so far, I think.

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Sep 9, 2008 at 10:47 AM
To: conlang-...@googlegroups.com

NOTE: I've switched this discussion over into a google group, 'cause
the CC list is getting out of hand.

Website: http://groups.google.com/group/conlang-card-game?lnk=gcamv

Email to group: conlang-...@googlegroups.com
What I meant is that the orthography, in a normal game, should have no
status whatsoever as canon - and preferably should be done in IPA.

I.e. one is not simultaneously creating *an orthography* (presumably a
Romanization) of the language together with the phonological form.
That's fine; just ensure the range is sane.
I think that this would overbalance the game in favor of good players
- where we want to ensure that it's fairly accessible to newbs (as is,
it'll be somewhat intimidating just due to content).
Make the low-key ones wildcards and the revolutionary ones specified?

Though IMO: most of the cultural stuff in our playtest at least was
more or less completely ignored other than as an occasional reference
joke. Is it worth putting a lot of cards into this?
That's the exact opposite of Chip's suggestion - namely, that too many
syntax cards is confusing and it should be simplified as possible.

I think that the default case might even be better off as "fix one
syntax, the end, play no more" and make using syntax cards an
advanced-play rule, because it causes complexities (and thus, in
addition to confusion, slowdowns).
Example of the syllable vs phoneme interpretation?

- Sai

----------
From: Sai Emrys <s...@saizai.com>
Date: Tue, Sep 9, 2008 at 10:50 AM
To: conlang-...@googlegroups.com