Zompist Bboard Again

bradrn wrote: ↑Wed Dec 16, 2020 1:56 am

• I think it should be stated that I see the primary aim of this game to be about having fun. In particular, I think this should be achieved by tryING to make the game go as quickly as possible, so we can prevent rounds from dragging on forever and let everyone have a turn at both guessing and presenting a language.

Yes, I agree. However, not everyone has the time to ask or answer questions several times a day, so the game shouldn't run too fast either. IMO the ideal duration for one round is something between two and four days.

• To that end, I think this game would run quite a bit faster if we were to allow people to guess language families and subgroupings. (I think a simple ‘is it Uralic?’ would have done wonders for guessing Moose-tache’s language.)

I also think that asking for language families should be allowed, but with restrictions. I'd suggest that the group as a whole gets one single shot at asking for a top-level family (maybe asking for a major subgroup of a large family, e.g. "Bodic" in Sino-Tibetan, would be OK too). So you should only ask for a family if you can't get a similarly useful answer with a different question. And if the language turns out to be too difficult to identify, the current quizmaster could decide to grant a second family question.

• Also, I think we should give people more guesses — as I said above, we very nearly did that anyway.
• A good compromise might be to give people 5 or so guesses (possibly more), but then count both languages and families as taking up one guess each. That way we still encourage more interesting questions and stop rounds from going too long, but also allow people to guess families if they’re confident.
• Perhaps, if everyone still runs out of guesses, we could then say that the winner is the person who guessed closest to the correct answer. (I happen to really like this approach, as I mentioned earlier… the really soul-sucking part of the old game, I think, was when we had to find the exact language within Chadic or Sino-Tibetan or whatnot. This way, we could just guess ‘Biu-Mandara’ etc., and be confident that this is good enough if no-one else guesses any closer.)

I like that approach too. So guessing a family or subfamily should be possible, and yes, it should count as a guess. But five guesses per person seems quite a lot. Especially if they can be asked right after one another, because then those people that have a lot of time could monopolise the game again. Maybe guesses should be spaced out in time a bit?

What about something like this: Everyone gets one guess in the first 48 hours, then one guess per day after that, and if the language hasn't been identified after three rounds of guessing, the winner is the person with the closest guess?


Another idea (not sure if it's any good, it might change the dynamics too much):
Maybe people may only make a guess when they've already had a question answered? Or even a question answered "yes"? (Both of these would have disqualified my own guess in the previous round though...) Or maybe you can gain the right to make an extra guess by asking a question that gets answered "yes"? Or by asking the best question of the day?

bradrn wrote:Finally: if we all agree on a permanent rule change, could we make a new thread please? (Call it ‘Name That Language! v2.0’ or something.) I just think that that would work better than having a rule change on page 35 of a pre-existing thread…

Agreed. I'll make a new thread when I post my language sample.

cedh wrote: ↑Wed Dec 16, 2020 4:32 am

bradrn wrote: ↑Wed Dec 16, 2020 1:56 am

• I think it should be stated that I see the primary aim of this game to be about having fun. In particular, I think this should be achieved by tryING to make the game go as quickly as possible, so we can prevent rounds from dragging on forever and let everyone have a turn at both guessing and presenting a language.

Yes, I agree. However, not everyone has the time to ask or answer questions several times a day, so the game shouldn't run too fast either. IMO the ideal duration for one round is something between two and four days.

A slightly longer duration that I would have chosen, but I agree that this is a good target nonetheless. (As it happens, I usually do have time to ask several questions a day, but I appreciate that others don’t.)

• To that end, I think this game would run quite a bit faster if we were to allow people to guess language families and subgroupings. (I think a simple ‘is it Uralic?’ would have done wonders for guessing Moose-tache’s language.)

I also think that asking for language families should be allowed, but with restrictions. I'd suggest that the group as a whole gets one single shot at asking for a top-level family (maybe asking for a major subgroup of a large family, e.g. "Bodic" in Sino-Tibetan, would be OK too). So you should only ask for a family if you can't get a similarly useful answer with a different question. And if the language turns out to be too difficult to identify, the current quizmaster could decide to grant a second family question.

Hmm… not a bad idea, but I can’t say I’m too enthused about it either. Possibly it would be better to give each person one chance to guess the family rather than giving it to the group as a whole?

• Also, I think we should give people more guesses — as I said above, we very nearly did that anyway.
• A good compromise might be to give people 5 or so guesses (possibly more), but then count both languages and families as taking up one guess each. That way we still encourage more interesting questions and stop rounds from going too long, but also allow people to guess families if they’re confident.
• Perhaps, if everyone still runs out of guesses, we could then say that the winner is the person who guessed closest to the correct answer. (I happen to really like this approach, as I mentioned earlier… the really soul-sucking part of the old game, I think, was when we had to find the exact language within Chadic or Sino-Tibetan or whatnot. This way, we could just guess ‘Biu-Mandara’ etc., and be confident that this is good enough if no-one else guesses any closer.)

I like that approach too. So guessing a family or subfamily should be possible, and yes, it should count as a guess. But five guesses per person seems quite a lot. Especially if they can be asked right after one another, because then those people that have a lot of time could monopolise the game again. Maybe guesses should be spaced out in time a bit?

True, I hadn’t thought of that consideration.

What about something like this: Everyone gets one guess in the first 48 hours, then one guess per day after that, and if the language hasn't been identified after three rounds of guessing, the winner is the person with the closest guess?

Good idea. Though I suspect 48 hours (=2 days) is a bit too long for one guess: that means it would take 6 days to exhaust all three guesses! I suspect 24 hours per guess would work better — it would take 3 days for three guesses, which fits in with your target of 2–4 days per round.

Finally: if we all agree on a permanent rule change, could we make a new thread please? (Call it ‘Name That Language! v2.0’ or something.) I just think that that would work better than having a rule change on page 35 of a pre-existing thread…

Agreed. I'll make a new thread when I post my language sample.

Thank you!

bradrn wrote: ↑Wed Dec 16, 2020 4:47 am

What about something like this: Everyone gets one guess in the first 48 hours, then one guess per day after that, and if the language hasn't been identified after three rounds of guessing, the winner is the person with the closest guess?

Good idea. Though I suspect 48 hours (=2 days) is a bit too long for one guess: that means it would take 6 days to exhaust all three guesses! I suspect 24 hours per guess would work better — it would take 3 days for three guesses, which fits in with your target of 2–4 days per round.

You misread: 48 hours is only for the first guess. Guesses 2 and 3 are granted after 24 hours. So it takes 4 days to use up all 3 guesses.

KathTheDragon wrote: ↑Wed Dec 16, 2020 5:42 am

bradrn wrote: ↑Wed Dec 16, 2020 4:47 am

What about something like this: Everyone gets one guess in the first 48 hours, then one guess per day after that, and if the language hasn't been identified after three rounds of guessing, the winner is the person with the closest guess?

Good idea. Though I suspect 48 hours (=2 days) is a bit too long for one guess: that means it would take 6 days to exhaust all three guesses! I suspect 24 hours per guess would work better — it would take 3 days for three guesses, which fits in with your target of 2–4 days per round.

You misread: 48 hours is only for the first guess. Guesses 2 and 3 are granted after 24 hours. So it takes 4 days to use up all 3 guesses.

Oops, it is indeed. Thanks for the correction!

Congrats, cedh!

This round has filled me with shame; Tungusic was in part my academic specialty and I have no excuse for having said "Oroch" instead of "Orok"... especially when I've spent decades, literally, with the nagging dread of confusing the names in public someday. I suck.

I like the new rules fine as they were this round. I'm very against expanding the number of phylogenetic guesses, having found that to be the absolute #1 interest-killer over the years this thread has run.

I also think speeding it up is not really helpful (effectively cutting non-online-24/7 folks out of full participation), and that adding the bookkeeping of how many questions of which kind each person vs. the group is allowed to ask, and when, creates a burden that won't pay for itself (and will likely become a distraction and source of contention in itself).

As far as the problem of the problem-poser being able to answer the 'more exotic' questions people will be asking, I see that as a potential feature-not-a-bug: it encourages us to draw on obscure languages we actually know a little about... or have to do some catch-up reading about, rather than just referring to a family tree to answer the genealogical questions that flood in. But, again, I think I'm short on the killer competitive instinct many of you have, and more of a 'the rapture of pursuit is the prize' kinda nerd...

I mostly prefer the new rules, too. Other than that, my two cents are that I agree with the non 24/7 hours sentiment.
If I were to make amendments to the latest rules, it would be that each participant would have one phylogenetical-OR-geographical yes/no question to ask every 48 hours and one is-it-this-specific-language question to ask every 24 hours. The time slots would be offset relative to the start of the current round, not per player; so if one really wanted, they could ask four such questions within minutes. Questions about linguistic topology, though, such as 'does the language have an X feature' would still not have any time limit.
The reasoning behind the specific cooldown periods is that phylogenetical and geographic questions are vastly more restrictive and elaborative than is-it-this-specific-language questions.

fusijui wrote: ↑Wed Dec 16, 2020 12:04 pm I also think speeding it up is not really helpful (effectively cutting non-online-24/7 folks out of full participation)

Hmm, OK. (Though by ‘speeding it up’, I meant more that one round shouldn’t take weeks to finish as they used to do, rather than that we should have multiple rounds a day or something.)

and that adding the bookkeeping of how many questions of which kind each person vs. the group is allowed to ask, and when, creates a burden that won't pay for itself (and will likely become a distraction and source of contention in itself).

I don’t see this as a problem, as long as the bookkeeping isn’t too complicated; this is one reason why my original suggestion was simply ‘you have 5 family/language guesses and no more’, since that’s easy enough to do.

As far as the problem of the problem-poser being able to answer the 'more exotic' questions people will be asking, I see that as a potential feature-not-a-bug: it encourages us to draw on obscure languages we actually know a little about... or have to do some catch-up reading about, rather than just referring to a family tree to answer the genealogical questions that flood in. But, again, I think I'm short on the killer competitive instinct many of you have, and more of a 'the rapture of pursuit is the prize' kinda nerd...

Oh, I wouldn’t call it a ‘killer competitive instinct’ by any means; I’m just saying that it can be horribly misleading when we get an incorrect answer to a question.

Zju wrote: ↑Wed Dec 16, 2020 1:17 pm If I were to make amendments to the latest rules, it would be that each participant would have one phylogenetical-OR-geographical yes/no question to ask every 48 hours and one is-it-this-specific-language question to ask every 24 hours. The time slots would be offset relative to the start of the current round, not per player; so if one really wanted, they could ask four such questions within minutes. Questions about linguistic topology, though, such as 'does the language have an X feature' would still not have any time limit.

I don’t think this is a good idea, for the simple reason that it’s horribly complicated. (I’m not even sure I fully understand what you’re saying!) In general, I prefer simpler rules for a game like this.

I thought keeping track of a single time limit is simpler, but whatever.

Zju wrote: ↑Thu Dec 17, 2020 1:50 am I thought keeping track of a single time limit is simpler, but whatever.

Oh, was that what you were saying? If so, I must have misread your earlier post — I interpreted you as suggesting that we have one time limit for phylogenetic questions and a separate one for specific language guesses, which seemed a bit overcomplicated to me.

------

Anyway, after thinking about what everyone else has said, here’s the rules I’m now leaning towards:

• People are given 4 guesses, to be spent on either phylogenetic affiliation or language. (Other guesses are unrestricted. Not sure what we should do with questions like ‘spoken in Polynesia?’, which are nearly synonymous with phylogenetic questions; my preference would be to count those as part of the 4 guesses.)
• Guesses must be made no less than 24 hours apart from each other, to ensure that no one person can monopolise the game.
• Each round has a time limit of four days. If after four days no-one has successfully guessed the language yet, the game goes to whoever guessed closest.

Does that sound reasonable?

bradrn wrote: ↑Thu Dec 17, 2020 4:18 am Anyway, after thinking about what everyone else has said, here’s the rules I’m now leaning towards:

• People are given 4 guesses, to be spent on either phylogenetic affiliation or language. (Other guesses are unrestricted. Not sure what we should do with questions like ‘spoken in Polynesia?’, which are nearly synonymous with phylogenetic questions; my preference would be to count those as part of the 4 guesses.)
• Guesses must be made no less than 24 hours apart from each other, to ensure that no one person can monopolise the game.
• Each round has a time limit of four days. If after four days no-one has successfully guessed the language yet, the game goes to whoever guessed closest.

Does that sound reasonable?

Almost. IMO phylogenetic guesses should be limited, in order to discourage tree-tracing. Note that in my last post I made a distinction between "asking for a language family" (where you'd get an answer that everyone could build on) and "guessing a language family" (where I assumed that no answer would be given before the end of the round, but which would count as a winning guess if nobody guessed closer). Of course this distinction probably wasn't obvious, and it probably wouldn't be easy to always tell apart either...

So I think I'll try running the next round with a slightly more restrictive version of your proposal: Everyone gets three guesses, one of which can be spent on phylogenetic affiliation. Let's just see how it works out.

cedh wrote: ↑Thu Dec 17, 2020 6:46 am

bradrn wrote: ↑Thu Dec 17, 2020 4:18 am Anyway, after thinking about what everyone else has said, here’s the rules I’m now leaning towards:

• People are given 4 guesses, to be spent on either phylogenetic affiliation or language. (Other guesses are unrestricted. Not sure what we should do with questions like ‘spoken in Polynesia?’, which are nearly synonymous with phylogenetic questions; my preference would be to count those as part of the 4 guesses.)
• Guesses must be made no less than 24 hours apart from each other, to ensure that no one person can monopolise the game.
• Each round has a time limit of four days. If after four days no-one has successfully guessed the language yet, the game goes to whoever guessed closest.

Does that sound reasonable?

Almost. IMO phylogenetic guesses should be limited, in order to discourage tree-tracing.

I did try to limit phylogenetic guesses: you only get four of them. (Less, if you want to ask about a specific language.)

Note that in my last post I made a distinction between "asking for a language family" (where you'd get an answer that everyone could build on) and "guessing a language family" (where I assumed that no answer would be given before the end of the round, but which would count as a winning guess if nobody guessed closer). Of course this distinction probably wasn't obvious, and it probably wouldn't be easy to always tell apart either...

Personally, I don’t see the difference: ‘Is this X?’ and ‘I think this is X’ are practically identical statements. I think we should avoid making such fine distinctions for what is, after all, just a fun game.

So I think I'll try running the next round with a slightly more restrictive version of your proposal: Everyone gets three guesses, one of which can be spent on phylogenetic affiliation. Let's just see how it works out.

Sounds like a good plan to me. Do we have a time limit between repeated guesses?

(But if we’re still trying out different rules, maybe we shouldn’t make a new thread just yet? Perhaps let’s wait until we’ve all agreed on the rules.)

OK. Here's a test run for the new version of the game: Name That Language! v2.0

The new rules are as follows:

• In each round, one person presents a paragraph of text in a natural language. The source text should be used as-is whenever possible (i.e. no re-transcriptions for the sake of obfuscation); however, small adjustments or straightforward transliterations into Latin or IPA are OK if the script/orthography itself is a dead giveaway.

• All readers are invited to try identifying this language. To this end, you may ask only yes-or-no questions. Whoever is the first to name the language correctly becomes the next poster of a sample text.

• Everyone gets a maximum of three guesses at the identity of the language. These guesses must be made no less than 24 hours apart from each other, to ensure that no one person can monopolise the game. No googling please!

• One guess per person may optionally be spent on phylogenetic affiliation instead, i.e. asking whether this language belongs to a specific language family. These guesses count to your maximum budget of three guesses, and they may target only top-level families or major subgroups of large families, so that the target group typically still contains at least 10-20 languages. Certain geographical questions like ‘Is it spoken in Polynesia?’, which are nearly synonymous with phylogenetic questions, are counted as pyhlogenetic questions.

• The number and frequency of non-identifying questions is not limited. A few examples: “Does this language have more than 1000 speakers?” “Does it have any famous native speaker?” “Is the scholarly literature on this language mainly in Russian?” “Is this language spoken in the vicinity of two different language isolates?” “Does it have adjectives?” “Do verbs conjugate differently depending on whether the direct object is definite or not?” The possibilities are endless...

• Each round of the game has a time limit of four days. If the language hasn't been identified by then, the game goes to whoever guessed closest (usually based on language families and subfamilies, but in exceptional circumstances it could also be based on geographical distance).

And here's the next language (to be identified before Dec 21):

Ngebuje ibukaypacap ibararak osodop taperadup pe ibibodo. Oibuꞌun ip jebibododon. Wararaꞌacat muyꞌun am ip oibuꞌun. Jebukaypacaptup oꞌtupmuꞌawero ip tupmupik am soatup. Oꞌtupmupik ip soat etabutpe. Teibong cincan osodop tuptei.

(I got very confused for a moment when I tried to reply and phpBB told me ‘this topic does not exist’, before I realised that you had created a new thread and then merged it into the old one…)

Anyway, the phonological inventory is pretty limited, with lots of palatals and stops, and no fricatives… South American, then?

Is it spoken in an area with rainforest?

Does the nominal plural suffix have a labial?

In your opinion, cedh, does this language have an unusual alignment system?

See, this is awesome, these four weird & interesting questions pouring right out I'm loving the renovation work on this thread!

Is the base of the number system something other than 10?
Does the language have a native word for kangaroo?

Very good questions!

bradrn wrote: ↑Thu Dec 17, 2020 7:52 amAnyway, the phonological inventory is pretty limited, with lots of palatals and stops, and no fricatives… South American, then?

Yes. (Although the language does have fricatives, even in the sample.)

Creyeditor wrote: ↑Thu Dec 17, 2020 8:32 am Is it spoken in an area with rainforest?

Yes.

Vilike wrote: ↑Thu Dec 17, 2020 9:44 am Does the nominal plural suffix have a labial?

No.

Moose-tache wrote: ↑Thu Dec 17, 2020 11:17 am In your opinion, cedh, does this language have an unusual alignment system?

The way you formulated this question, the answer depends on my own judgement of what constitutes "an unusual alignment system". From what I know about this language, I'm quite sure that there are some details in the alignment system that I would indeed consider unusual, but I haven't come across such details yet (but I haven't studied the language all too thoroughly). And the overall alignment system is not really unusual in my opinion, so you get a tentative No.

Zju wrote: ↑Thu Dec 17, 2020 1:46 pm Is the base of the number system something other than 10?

Yes.

Zju wrote: ↑Thu Dec 17, 2020 1:46 pm Does the language have a native word for kangaroo?

No. (Actually I don't know, but it would be a big surprise.)


And since we're on the next page already, here's the sample again:

cedh wrote: ↑Thu Dec 17, 2020 7:43 am

Ngebuje ibukaypacap ibararak osodop taperadup pe ibibodo. Oibuꞌun ip jebibododon. Wararaꞌacat muyꞌun am ip oibuꞌun. Jebukaypacaptup oꞌtupmuꞌawero ip tupmupik am soatup. Oꞌtupmupik ip soat etabutpe. Teibong cincan osodop tuptei.