Linguistic Fraud

[Sir Thopas]

Tossups on the Winter War should not have a place name ending in "-niemi" in the lead-in (I'm looking at you, CO 2007 tossup). Tossups on Hungary should not have "Ferenc" on the second line (I'm looking at you, NAQT tossup from a couple of years ago). Tossups on Jean Sibelius should not have Finnish-sounding place names in the first line.

Clearly, the first of these is much less egregious than the other two. Nevertheless, as Andrew said, it disappoints me whenever writers, however good otherwise, don't pay attention to avoiding this.. I want to win based on having more knowledge on the topic at hand, not because I figure out that the moderator is struggling to pronounce a digraph that appears in Albanian.

[Important Bird Area]

Let's talk about "linguistic fraud."

Guy's complaint about linguistic fraud in high school sets, and Jeff's response that NAQT doesn't believe that enough high schoolers have the knowledge to convert tossups by figuring out linguistics, is indicative of this behavior. Does NAQT really believe that language transparency (which it admits is over the heads of most high schoolers) draws in casual teams? Or is linguistic transparency a tenet of good quizbowl that casual players don't even realize exists and thus cannot color their experience, and NAQT is simply refusing to abide by community standards because it's too much trouble to address the criticisms of a few "elite" players?"

I'm going to go ahead and define what I understand to be linguistic fraud/language transparency. Please post and let me know if I miss something.

Linguistic fraud happens when a player recognizes the language affinity in the proper names found in a question, and uses that to sort the answer space to arrive at the correct response faster than a player who has deeper "real" knowledge (for now, let's use the Potter Stewart test for what "real" knowledge is) of the item in question.

So, linguistic fraud works like this:

1. A player hears a clue such as "... this castle was besieged by Takeda Katsuyori...".

2. Player A has deeper knowledge of 16th-century Japanese history, and would likely buzz with "Nagashino" on a later clue like "... innovative firearms tactics ensured a Tokugawa victory at what 1575 battle?"

3. Instead, Player B combines the recognition that "Takeda Katsuyori sounds Japanese!" with "there was a famous Japanese battle at a place called Nagashino", buzzes in, and answers correctly.

The two threads on this have been merged. --the mgmt

[Important Bird Area]

Guy and I were posting simultaneously, I see.

OK, here's part two.

Thus, the prevalence of linguistic fraud depends on the following:

-how easy is it to recognize the language in question?

-how large is the answer space of whatever the indicated pronoun is?

Thus, I would submit that linguistic fraud is less likely at tournaments like Gaddis and Chicago Open, where there are multiple conceivable answers for just about everything.

Classic examples would include the Kalevala and the Treaty of Waitangi, both because Maori and Finnish are distinctive-sounding and because the answer space of "documents from New Zealand history" and "collections of Finnish traditional epic" is nearly uniquely identifying.

My initial response to Guy was not meant to dismiss the complaint of linguistic fraud entirely. It was meant to suggest that it is less likely to occur at lower levels of difficulty. I suspected that very few average high school players would, say, recognize "Lemminkainen" as Finnish but simultaneously not know that "Lemminkainen is a character from the Kalevala."

That being said: it is entirely possible to find examples of this fraud at any difficulty level. Please feel free to cite such questions from any cleared NAQT set (or anywhere else, for that matter).

[Sir Thopas]

My initial response to Guy was not meant to dismiss the complaint of linguistic fraud entirely. It was meant to suggest that it is less likely to occur at lower levels of difficulty. I suspected that very few average high school players would, say, recognize "Lemminkainen" as Finnish but simultaneously not know that "Lemminkainen is a character from the Kalevala."

I mean, why would Lemminkainen be the first name mentioned in a Kalevala tossup? The whole point is that it's happened pretty frequently (especially in NAQT, I've found) that I've buzzed on knowledge I don't have otherwise. In fact, if one were to game the system, linguistic fraud might be a pretty good way to do it.

Also, I'm not quite sure what you're getting at when you say it doesn't happen at lower levels of difficulty. It's precisely because the answer space is so small that it's a problem in lower levels. Sure, most high school players may not be able to do it, but most high school players can't name the Kalevala either. I fail to see why that's a reason not to pay attention to it.

[Brian Ulrich]

This issue has actually been discussed among NAQT editors before. The current internal conventional wisdom is that something like a Japanese-sounding name functions much like a clue, and should be placed wherever it fits pyramidally. Knowledge of what sounds Finnish is rare enough that, in my questions, I'm not afraid to put it early. Japanese is a different story, particularly when combined with evidence of CE-ness, which effectively narrows it to Aso Taro or Akihito. English is distinctive, but usually unhelpful due to a huge number of possible answers.

[Matt Weiner]

Thus, I would submit that linguistic fraud is less likely at tournaments like Gaddis and Chicago Open, where there are multiple conceivable answers for just about everything.

Classic examples would include the Kalevala and the Treaty of Waitangi, both because Maori and Finnish are distinctive-sounding and because the answer space of "documents from New Zealand history" and "collections of Finnish traditional epic" is nearly uniquely identifying.

This seems true.

My initial response to Guy was not meant to dismiss the complaint of linguistic fraud entirely. It was meant to suggest that it is less likely to occur at lower levels of difficulty. I suspected that very few average high school players would, say, recognize "Lemminkainen" as Finnish but simultaneously not know that "Lemminkainen is a character from the Kalevala."

And this is invalidated by the above. In high school, the Kalevala is the only Finnish literature that could come up (and it's high school nationals level, for the most part). If you're buzzing on Finnish names at ACF Nationals, maybe you're negging yourself out of a tossup on Frans Sillanpää. You are not doing this at a high school tournament unless the editor has completely lost his mind.

In other words: Linguistic fraud becomes exponentially more of a concern as the difficulty level goes down.

[Important Bird Area]

why would Lemminkainen be the first name mentioned in a Kalevala tossup? The whole point is that it's happened pretty frequently (especially in NAQT, I've found) that I've buzzed on knowledge I don't have otherwise.

I wasn't suggesting that it ever should be. I don't see any reason to think that linguistic fraud is any less likely in a middle clue than a leadin. (Although I suppose by that point it could be considered less fraudulent...)

The whole point is that it's happened pretty frequently (especially in NAQT, I've found) that I've buzzed on knowledge I don't have otherwise.

Examples welcome, of course. Here's the Hungary example you mentioned above (from IS #64):

"This country became autonomous through an 1867 agreement between Franz Joseph I and Ferenc Deac. After World War I, Admiral Miklos Horthy signed the Treaty of Trianon and overthrew the Soviet Republic begun by Bela Kun. Janos Kadar led the (*) "goulash communism" system after Soviet troops deposed Imre Nagy in 1956. For 10 points--name this country whose capital is Budapest."

I think this might be less linguistically transparent and more just transparent transparent, in the sense that the first sentence gives away "it was part of the Austro-Hungarian Empire" while the second sentence provides more substantive Hungarian history content.

[Sir Thopas]

I buzzed on Ferenc. The question may have been bad in other ways, but it was certainly linguistic fraudable.

For what it's worth, another type of linguistic fraud leads to me thinking that Bela Kun shouldn't be in power. People are much more likely to know that Bela Bartok is from Hungary and buzz on first names than actually know of Bela Kun.

I'm going to find fraudable tossups in the first non A-set I see, which is the 2005 HSNCT. Expect a post in a while.

[Important Bird Area]

In other words: Linguistic fraud becomes exponentially more of a concern as the difficulty level goes down.

I think this is true for certain levels of difficulty: from Gaddis-level down to say, high school nationals. Below that, I think the prevalence of linguistic fraud at least levels off and probably decreases, simply because very inexperienced players are less likely to be buzzing off of some fragment of an uncommon language. In short: I agree with Brian that very few 10th-graders will recognize Finnish but not have heard of the Kalevala.

[Important Bird Area]

I'm going to find fraudable tossups in the first non A-set I see, which is the 2005 HSNCT. Expect a post in a while.

Great, looking forward to it.

For what it's worth, another type of linguistic fraud leads to me thinking that Bela Kun shouldn't be in power. People are much more likely to know that Bela Bartok is from Hungary and buzz on first names than actually know of Bela Kun.

This is one of those "quizbowl arms race" phenomena that happen between players and editors. Roughly looks like this:

-Once upon a time, players complained about vague leadins.
-Editors produced leadins containing clear pronoun referents and lots of proper names.
-Players used some of those names to linguistically fraud questions.
-Editors resorted to using common languages like English to limit linguistic fraud.
-Players complained that the leadins contained too little non-western content.

(analogous process for, say, those minor works by Schopenhauer that we were discussing earlier this fall)

I think a stasis that is good quizbowl can certainly be produced from this effect. Whether that stasis meets an individual player's standards for optimal fraud-avoidance (or whatever) may be another matter.

Note that I'm not trying to excuse any particular example of linguistic fraud (which may indeed be terrible quizbowl).

[Matt Weiner]

-Players complained that the leadins contained too little non-western content.

Who is complaining about this, exactly?

[Important Bird Area]

-Players complained that the leadins contained too little non-western content.

Who is complaining about this, exactly?

This probably doesn't have a direct connection to linguistic fraud (the argument being that increasing non-western requirements in the distribution will widen the answer space, making fraud less likely).

But "too little non-western content" is certainly on my radar as one usual criticism of NAQT. (see examples in these threads) In particular, I had in mind the Waitangi tossup from last year's SCT, which prompted some discussion of linguistic fraud offboard.

[Matt Weiner]

Yes, I will put myself on record as being someone who wants to see more non-western content in NAQT in general. But I'm wondering who is saying "I'm fine with your general level of non-western content, but the fact that the content isn't specifically in the leadins of tossups is unacceptable!" I find it implausible that even one such person exists.

[Important Bird Area]

But I'm wondering who is saying "I'm fine with your general level of non-western content, but the fact that the content isn't specifically in the leadins of tossups is unacceptable!" I find it implausible that even one such person exists.

As do I; I was postulating the likely reaction from someone who (like you) believes in increasing the non-western distribution, were a hypothetical editor to overzealously remove distinctively non-western names from leadins.

[at your pleasure]

That being said: it is entirely possible to find examples of this fraud at any difficulty level. Please feel free to cite such questions from any cleared NAQT set (or anywhere else, for that matter).

Nobles joined this in part because of the Doctrine of Lapse, such as the Rani of Jhansi’s desire for her son to rule, and the death of Mangal Pandey also helped trigger it. Its final stronghold was in Oudh, and the leader of forces at Kanpur in June had been denied his father's title by Lord Dalhousie. That man, Nana Sahib, caused the death of Commander Wheeler in an incident known as the Cawnpore Massacre. Believed to have started due to the presence of animal fat in new weapons, FTP, name this rebellion by the namesake Indian soldiers against the British.

It sounds Indian and vaugly historical/military. Given the fact that there are few Indian military events other than the sepoy rebellion that come up on a reguar basis at normal HS events , it is quite fraudable. On the other hand, on higher-level questions, many miltary events/movements in Indian history could be reasonably asked. This is why a larger non-western canon-and canon expansion in general-makes for less transparency;the possible answer space for any given set of nonspecific criteria is larger.

[Sir Thopas]

I gotta say, there's not much here. Most of these questions are fraudable or poorly written in other ways long before anyone with linguistics can gain an advantage. Nonetheless, there are a couple.

"In Utgard it..." (3.25) How many objects from Norse myth are askable? The -gard ending gives it away.
"It was opposed by King Waldemar IV..." (8.26) While this is a pretty misplaced lead-in in and of itself, what polities could there be opposed by someone from Northern Europe besides the Hanseatic League? Schmalkaldic League isn't askable at this level.
"The earliest ones formed in 1980 under Colonel Enrique Bermudez...." (10.16) OK, I guess this could be Contras or Sandanistas, but it's still pretty poor.
"To its extreme north are the Vuntut and Ivvavik national parks..." (12.3) Like, this is the Yukon territory, and it's immediately either that or a hose for Nunavut. I probably would have buzzed with the latter, but Inuit names are very distinguished.
"While working as an apothecary's apprentice in the village of Grimstad, he wrote a play..." (13.17) This is not Bjornstjerne Bjornson.
"The works of Arnaut Daniel and Jaufre Rudel are still in existence..." (17.7) If you can figure out what the hell's going on here, your knowledge of Middle French/Provencal names will get you 15 points for "troubadour".
"The protagonist fails to buy them from Nozdryev..." (18.2) This is fraudable in other ways, but the Russian name is an immediate buzz for Dead Souls for anyone with a modicum of knowledge about the book. Actually, yeah, just chalk this up to regular fraud.
"Faults along the west side of this mountain range have led to a chain of volcanoes including Mount Amiata and the Alban Hills...." (18.6) I mean, I hear Amiata is an Italian name and Alba Longa is a famous city in Italy. These are indeed the Apennines.

And that's all I can find. Admittedly, this wasn't as many as I remember, so you may think I'm just blustering. Nevertheless, all of these were pretty avoidable. I also think it's worth saying that I'd forgotten how generally fraudable this set was. When there's so many other ways to fraud questions, linguistics isn't always the most frequent method. Make fo this what you will, I guess. I still don't like linguistic fraudability and wish it would be avoided much like all fraudability. NAQT doesn't seem to care much about either much of the time, honestly.

[Important Bird Area]

And that's all I can find. Admittedly, this wasn't as many as I remember...

Some of those are indeed susceptible to linguistic fraud (eg, the title items purchased from someone Russian-sounding). Some of them are other kinds of fraud (like Waldemar IV, stock-clue enemy of the Hanseatic League- that's probably why this got banished to tossup 26 in the first place). Others may not be fraudulent- if you can distinguish Provencal from the other Romance languages from two personal names, I would be quite happy to award you 15 points for a troubadour tossup at this level.

[The Laughing Man]

"While working as an apothecary's apprentice in the village of Grimstad, he wrote a play..." (13.17) This is not Bjornstjerne Bjornson.

This lead in has a more fundamental flaw than its fraudability; namely, the fact that the only possible buzz on the lead in is a fraudulent one. For the player with knowledge of Norwegian village names, clearly it allows a fraudulent buzz,but I contend that even the player who gets it because he knows where Ibsen grew up has frauded the question. A person who has memorized where Ibsen grew up should not get a question over someone who has read 6 Ibsen plays and familiarized himself with character names from several more.
In addition to rewarding fraud, this lead in wastes a bunch of characters which could have been saved by saying "In one of this author's plays...." This would allow for actual description of the works, which is notably lacking in many IS and even national championship literature questions. I suspect that biographical literature lead ins are not so much an institutional problem of NAQT as a result its dearth of good writers; however, this is definitely something that should be strongly discouraged in the question writing guides(or whatever it is) that NAQT distributes to its question writers. I have seen several egregious biographical literature questions in IS sets this year and it needs to be addressed.

[Ike]

I finally got a book on Occitan and I can surely distinguish between the two languages, so I guess that would mean 15 for me. But...
Seriously, Jaufre is an older French name, I mean many of the troubadours have French names that are an older form of the language. Seriously, you don't come across people named Jaufre nowadays.
Combine the fact that it doesn't sound modern, so its probably a couple hundred years old + French + an artiste of some sort, and you get where Guy is headed.

For what its worth, I think its a slippery slope to say

I would be quite happy to award you 15 points for a troubadour tossup at this level.

Linguistic fraud is a bad thing, and I find it disturbing that instead of actually rewarding deep knowledge on the subject of literature, we are rewarding knowledge on languages. (And not even really on this tossup)

Edit: Grammar and Redundancy

[Sir Thopas]

Yeah, Ike fleshed out what I meant to say but didn't. Arnaud is an old-timey name from France. Whether or not it's Provencal, I don't specifically know. It's not necessary to distinguish between that and Middle French to linguistic fraud that question.

[BuzzerZen]

I contend that even the player who gets it because he knows where Ibsen grew up has frauded the question. A person who has memorized where Ibsen grew up should not get a question over someone who has read 6 Ibsen plays and familiarized himself with character names from several more.

Well, let's be careful about terminology here. It's not fraudulent to buzz off a clue giving an author's birthplace and name the author. You're using explicit information in the question; there's no metagaming, guessing, or lateral thinking involved (if, in fact, you know Ibsen was born in Grimstad). What is going on here is the question is rewarding knowledge that it ought not to reward. For my own part, I think the notion of "fraud" in quiz bowl is a sketchy one anyway, and seems to just be a gut appraisal of how many levels of meta-knowledge are appropriate for someone to be buzzing off of. But this is a topic, of course, for yet another thread.

[Matt Weiner]

Fake clues that you get from memorizing a list of nonsense like author birthplaces are the core definition of what "fraud" is.

[BuzzerZen]

Fake clues that you get from memorizing a list of nonsense like author birthplaces are the core definition of what "fraud" is.

In that case, I'm pretty sure people use "fraud" to mean a variety of suspicious practices, ranging from this one (buzzing on purposefully included but disfavored clues) to things more along the lines of meta-gaming, which is what the "linguistic fraud" under discussion seems to fall under. I think we have a slight poverty of vocabulary, here, because clearly writing tossups with disfavored clues is always bad, whereas meta-gaming probably isn't always bad since I think most top players do a lot of it.

[Terrible Shorts Depot]

While I agree that linguistic fraud is a huge problem and one that I desire to be free from, I don't know how one would get rid of it. All efforts to hide the nationality of the answer, e.g., not mentioning any people in a tossup on Mishima, seem very similar to those stupid tossups on Marie Curie where she is referred to with genderless and vague pronouns, e.g., "this scientist" or "they". If you are writing the aforementioned tossup on Mishima, is it possible to keep distinctly Japanese character names out of it until somewhat late in the tossup without a myraid of uses of "main character" or other such words? Or, perhaps, can we assume that the canon of Japanese authors or female scientists is large enough at whatever level to not make "she" coupled with some science words a dead giveaway? What about subjects with smaller canons?

EDIT: Cleared some stuff up and added some questions.

[Sir Thopas]

Yeah, there are definitely enough Japanese authors to make that viable. I've found that linguistic frauding is usually avoidable—I've been taking pains to keep it out of Prison Bowl, for example, hopefully successfully. I'd like to think that I'm acutely aware of it, and that helps me write, but we'll see.

[Maxwell Sniffingwell]

Fake clues that you get from memorizing a list of nonsense like author birthplaces are the core definition of what "fraud" is.

I was under the impression that "fraud" is answers that you get without memorizing any clues at all - like buzzing with Carmen because the answer looks like it's going to be an opera set in Spain.

If you memorize a specific piece of useless information and it gets you points because that specific piece of information is mentioned, that's bad clue choice, not fraud... I guess what I'm saying is that fraud must by definition (in my opinion) involve lateral thinking.

[Maxwell Sniffingwell]

19. A premonition of the title character’s death is found in her “Card Song”, while another major moment of the opera is the duet “Seguidilla”.The male lead’s fate is sealed when he draws his sword on Captain Zuniga, leading him to flee to the mountains and join the smuggling gang led by Lillas Pastia. Shortly after singing his “Flower Song” he abandons his sweetheart Micaela, but his beloved soon cuckolds him for the bullfighter Escamillo, who sings the famous “Toreador Song”. FTP, what is this opera about a gypsy woman, the masterpiece of Georges Bizet?

By the end of the first sentence, you've got a opera with a Spanish female title character... buzz. To me, that's fraud in its purest form.

[Sen. Estes Kefauver (D-TN)]

Who cares what we call it? It's all bad.

[setht]

I gotta say, there's not much here. Most of these questions are fraudable or poorly written in other ways long before anyone with linguistics can gain an advantage. Nonetheless, there are a couple.

"In Utgard it..." (3.25) How many objects from Norse myth are askable? The -gard ending gives it away.

If that's the opening sentence of the question, I think there are several plausible answers after those three words, at an IS-level tournament. Perhaps the next few words rule out most or all of the possibilities I can think of and do in fact make the question susceptible to linguistic fraud, but buzzing off of those three words thinking, "time to guess a thing that appears in Norse myth because I heard Utgard" seems like a terrible idea to me. It's not even clear to me that the question has to be exclusively Norse myth, since possible answers include things like fire, thought, old age, and the sea.

-Seth

[AlphaQuizBowler]

I think this may be a little much. Yeah, there are blatant examples of fraud (such as the Carmen TU), but how many people are sorting this out while the TU is being read? My point is, it's easy to go back and say, yeah, those were French-sounding names, I should have had "troubador", but does anyone think that in a game? And, Guy, I have to say your evidence is very tenuous in my opinion. Most of them were straight-up fraud, and I think you went a little overboard on this:

"Faults along the west side of this mountain range have led to a chain of volcanoes including Mount Amiata and the Alban Hills...." (18.6) I mean, I hear Amiata is an Italian name and Alba Longa is a famous city in Italy. These are indeed the Apennines.

First of all, if you know Alba Longa is a famous city in Italy (and can relate that to Alban Hills), you have real knowledge of Italian geo that should be rewarded. Besides, don't pretend that the Apennines are the only mountains in Italy that are askable. You would be risking it trying to fraud this.

All in all, I can agree that maybe this happens sometimes, but I think you overstate the prevalence of linguistic fraud.

[Skepticism and Animal Feed]

One way to avoid Linguistic fraud is to use alternate names. A lot of really Slavic-sounding things have alternative German names, and vice-versa.

Hungary is particularly problematic because its first names are distinctive. A Slavic-sounding name can be from a bunch of different countries, but only one country is in the business of exporting guys named Ferenc. The way I usually solve this is by translating the name; I would say Francis instead of Ferenc, Emmeric instead of Imre, Nicholas instead of Miklos, etc. I think this is justified when saying the ethnic form of the name would constitute a giveaway to Guy or anyone else with linguistics knowledge.

Also, generous pronunciation guides can solve the "struggling over an Albanian digraph" problem. Many words that look extremely foreign actually sound very bland when they are properly said.

[Sir Thopas]

Also, generous pronunciation guides can solve the "struggling over an Albanian digraph" problem. Many words that look extremely foreign actually sound very bland when they are properly said.

This is true. Still something to watch out for, though.

[grashid]

One way to avoid Linguistic fraud is to use alternate names. A lot of really Slavic-sounding things have alternative German names, and vice-versa.

Hungary is particularly problematic because its first names are distinctive. A Slavic-sounding name can be from a bunch of different countries, but only one country is in the business of exporting guys named Ferenc. The way I usually solve this is by translating the name; I would say Francis instead of Ferenc, Emmeric instead of Imre, Nicholas instead of Miklos, etc. I think this is justified when saying the ethnic form of the name would constitute a giveaway to Guy or anyone else with linguistics knowledge.

Also, generous pronunciation guides can solve the "struggling over an Albanian digraph" problem. Many words that look extremely foreign actually sound very bland when they are properly said.

Wouldn't this put someone who has learned it as Miklos or Imre, but doesn't know the translation, at a disadvantage?

[setht]

One way to avoid Linguistic fraud is to use alternate names. A lot of really Slavic-sounding things have alternative German names, and vice-versa.

Hungary is particularly problematic because its first names are distinctive. A Slavic-sounding name can be from a bunch of different countries, but only one country is in the business of exporting guys named Ferenc. The way I usually solve this is by translating the name; I would say Francis instead of Ferenc, Emmeric instead of Imre, Nicholas instead of Miklos, etc. I think this is justified when saying the ethnic form of the name would constitute a giveaway to Guy or anyone else with linguistics knowledge.

I think this would be problematic the other way, confusing people that do have knowledge of a topic by using some variant of a name that no one uses in normal conversation/books/the web (that is, a variant that no one uses to refer to the person the question is referring to). If you want to sneak a reference to Imre Nagy or Imre Kertesz past players, I'd say that putting in "Emmeric" or "Emmerich" in place of "Imre" is a bad idea, almost as bad as disguising dates that would otherwise give things away too fast (e.g. 1066 CE) by announcing them in hexadecimal. If a clue is going to give things away to too many people at some point in a question due to its factual content or linguistic associations or whatever, just don't use it at that point in the question.

-Seth

[STPickrell]

I consider the presence of words in the 'target language' to be a clue in and of themselves, to be placed appropriately the same as any other clue. Some of these workarounds strike me as 'cures worse than the disease,' such as using non-standard names (i.e. not Posen/Posnan or Danzig/Gdansk) in the stead of 'instant language recognition' clues.

Is it 'fraud' to get a TU off of a stock lead-in that's been relegated to mid-level or pre-giveaway status?

Is it 'fraud' to get a TU off another type of clue, such as knowing a subcommander at a battle or knowing minor characters from a work of literature?

Why is it 'fraud' to get a TU off of recognizing a word in a language and then making a good guess, using your knowledge of the canon?

Likewise there are compound clues, clues which can only exist after two separate clues (which are independent in and of themselves.) The Hans Christian Andersen TU quoted in another thread contains such a clue -- the International Children's Reading Day and the poem 'I am a Scandinavian' are separate clues that together form a compound clue 'Scandinavian children's lit author.' Compound clues are often non-specific but can do a great deal to narrow the potential field of answers to 1-3 possible answers.

I think some of the linguistic fraud investigators are concerned about players who can (1) recognize the language and (2) be filtering through ever-more difficult answers (e.g. Ahtisaari, Silanpaa, etc.) while a slightly-less talented player buzzes with the stock answer for that language (Mannerheim, Sibelius, Kalevala, etc.) I'm not sure how prevalent this really is.

[Sir Thopas]

Is it 'fraud' to get a TU off of a stock lead-in that's been relegated to mid-level or pre-giveaway status?

Is it 'fraud' to get a TU off another type of clue, such as knowing a subcommander at a battle or knowing minor characters from a work of literature?

Why is it 'fraud' to get a TU off of recognizing a word in a language and then making a good guess, using your knowledge of the canon?

Dude, all you've shown me is that you have no clue what fraud is. The first two are not fraud because they require actual knowledge of the clue. A stock clue too early is bad quizbowl, but it's not fraud. The last one, on the other hand, is the very definition of fraud. It doesn't use any knowledge of the subject at hand, just knowledge of the canon. If you don't get that this is, quite literally, fraud in its purest sense, I can't tell you anything further.

Likewise there are compound clues, clues which can only exist after two separate clues (which are independent in and of themselves.) The Hans Christian Andersen TU quoted in another thread contains such a clue -- the International Children's Reading Day and the poem 'I am a Scandinavian' are separate clues that together form a compound clue 'Scandinavian children's lit author.' Compound clues are often non-specific but can do a great deal to narrow the potential field of answers to 1-3 possible answers.

I'm not sure what you're saying here. Are you defending it? Because the compound clues you're talking about are bad because they lead to fraud. I've never heard of "I am a Scandinavian", but I'm sure as hell going to buzz with HC Andersen because I've been told that it's a Scandinavian children's lit author. This is not a good thing. This is fraud, because it doesn't actually require me to know anything about Andersen's oeuvre.

I think some of the linguistic fraud investigators are concerned about players who can (1) recognize the language and (2) be filtering through ever-more difficult answers (e.g. Ahtisaari, Silanpaa, etc.) while a slightly-less talented player buzzes with the stock answer for that language (Mannerheim, Sibelius, Kalevala, etc.) I'm not sure how prevalent this really is.

You're thinking wrong. Linguistic fraud goes with knowledge of the canon. I'm concerned about someone using a narrow canon and knowledge of a language to beat someone who has more legitimate knowledge. Anyone who thinks that freaking Silanpaa is going to come up in a high school packet is being absurd. The whole problem is that someone with legitimate knowledge—for example, someone who has read plot summaries of the Kalevala and knows that Vainamoinen defeats a brash young man in a singing contest and causes her sister to kill herself—could lose to someone who knows nothing about the Kalevala but is buzzing off Joukahainen. Do you see why this is bad?

[Mechanical Beasts]

Why is it 'fraud' to get a TU off of recognizing a word in a language and then making a good guess, using your knowledge of the canon?

Here's a very simple example. I did not know what the Meiji Restoration was until senior year. I knew that it was a thing that happened in Japan. I thought it was spelled Magi; I kid you not. I heard a Japanese name, the question sounded like a non-war event, I buzzed.

Is this risky at high levels of quizbowl? Surely. Is it risky at high levels of high school quizbowl? Maybe. But the canon of tossupable Japanese history is particularly small in high school.

I should never get points for knowing that the "Magi" Restoration happened in Japan before someone with real knowledge about it precisely because "FTP, name this event that's not a battle or war [implicit: that is in the high school canon; sorry, lovers of the Keian Uprising] that happened in Japan" is a giveaway-level clue if you understand the implicit part.

Linguistic fraud allows you to answer a tossup off canon knowledge alone earlier than you should. If you don't see that that's bad, I can't help you.

[Skepticism and Animal Feed]

I think this would be problematic the other way, confusing people that do have knowledge of a topic by using some variant of a name that no one uses in normal conversation/books/the web (that is, a variant that no one uses to refer to the person the question is referring to). If you want to sneak a reference to Imre Nagy or Imre Kertesz past players, I'd say that putting in "Emmeric" or "Emmerich" in place of "Imre" is a bad idea, almost as bad as disguising dates that would otherwise give things away too fast (e.g. 1066 CE) by announcing them in hexadecimal. If a clue is going to give things away to too many people at some point in a question due to its factual content or linguistic associations or whatever, just don't use it at that point in the question.

The thing is, plenty of sources do refer to these people by translated names. Virtually any famous historical figure from Hungary from before 1918 is, in my experience, known interchangably by the English, German, and Hungarian forms of his name in historical sources, history books, etc. Many of them even referred to themselves this way: Miklos Horthy signed his autobiography as "Nicholas von Horthy". And I'm not the only one to do this in quizbowl: virtually any reference to Ferenc Gyulai in the canon is as "Franz Gyulai".

If the alternate forms are out there, I say use them strategically.

[Skepticism and Animal Feed]

One way to avoid Linguistic fraud is to use alternate names. A lot of really Slavic-sounding things have alternative German names, and vice-versa.

Hungary is particularly problematic because its first names are distinctive. A Slavic-sounding name can be from a bunch of different countries, but only one country is in the business of exporting guys named Ferenc. The way I usually solve this is by translating the name; I would say Francis instead of Ferenc, Emmeric instead of Imre, Nicholas instead of Miklos, etc. I think this is justified when saying the ethnic form of the name would constitute a giveaway to Guy or anyone else with linguistics knowledge.

Also, generous pronunciation guides can solve the "struggling over an Albanian digraph" problem. Many words that look extremely foreign actually sound very bland when they are properly said.

Wouldn't this put someone who has learned it as Miklos or Imre, but doesn't know the translation, at a disadvantage?

Doesn't referring to Aachen put somebody who learned it as Aix-la-Chapelle at a disadvantage? Doesn't referring to it as the Battle of Tannenberg put somebody who learned it as Grunwald at a disadvantage? These things happen in quizbowl, and people just have to deal with it, or learn more things

[setht]

some variant of a name that no one uses in normal conversation/books/the web (that is, a variant that no one uses to refer to the person the question is referring to).

The thing is, plenty of sources do refer to these people by translated names...If the alternate forms are out there, I say use them strategically.

Sure, I think we agree on this. Now, when there aren't plenty of (reasonable) sources that refer to a person by a translated name, would you still advocate using alternate forms strategically? I would not.

-Seth

[No Rules Westbrook]

Well, Seth, when you say that the answer could be "fire" or "thought"...yeah, maybe it could, if I were confident that the question was written by a good writer. But, a lot of fraud in general comes from the fact that you're at a tournament where you're already pretty distrustful of the questions.

If I go to a high-quality event written/edited by known good writers, I come in with a certain trust that employing fraud isn't going to pay off. That trust quickly evaporates when you hear questions that are fraudable (in any number of ways), contian unimportant trivial clues, pyramidality problems, and the like. Good players very quickly adjust their strategy from "let's play honestly and buzz when I know something" to "let's try to game the questions." Repeated instances of linguistic fraudability are a sure way to get people to lose trust in a set, and then you're left with...frustration! Now where have I heard that word before?

Also: On the Bruce thing, don't start translating and using alternate titles for things unless they appear with some frequency in sources. That just equals confusion...and more frustration.

[setht]

Well, Seth, when you say that the answer could be "fire" or "thought"...yeah, maybe it could, if I were confident that the question was written by a good writer.

I'm not sure why you feel that a tossup on "fire" or "thought" could only be written by a good writer, but even if I discard those possible answers I'm still not sure whether the tossup is on Mjollnir or the Midgard Serpent (Guy or someone else who saw that question: what was the answer, anyway?). A clue that lets a player narrow things down to one of two (or three, or four, or any small number) of answers within three words without having deep knowledge is a bad clue, even if it's not super-buzzable because there are still multiple options. Whether it's a misplaced linguistic clue or a misplaced whatever clue doesn't matter; it's bad.

-Seth

[STPickrell]

Linguistic fraud allows you to answer a tossup off canon knowledge alone earlier than you should. If you don't see that that's bad, I can't help you.

OK. What you're saying is that a Meiji Restoration question shouldn't contain actual Japanese names until maybe mid-way through. I agree with you.

My point is that 'Japanese-sounding names' are a clue in and of themselves, to be put in their proper place. Or from my first post: "I consider the presence of words in the 'target language' to be a clue in and of themselves, to be placed appropriately the same as any other clue."

So Japanese-sounding names showing up midway through a Meiji Restoration tossup are OK in my book.

Something like 'Toba-Fushimi' in the first line is a little too giveaway-ish. OTOH, I think 'Republic of Ezo' or 'Boshin war' don't scream 'Japanese' in the same way, and ending the 'four divisions of society system' certainly requires you to know a little bit about Japan.

Guy --

My point is that fraudable clues are OK, in their proper place..

You don't want 'I am a Scandinavian' and 'Children's lit' in the power part of a tossup. But having the latter of the two appear as a mid-level or pre-giveaway clue (depending on the level of play)? Not so bad. See above regarding language clues.

A less-talented player would still need to recognize the Finnish-sounding names as Finnish to get the tossup. But, to a very experienced group of players, these clues are pre-giveaway level clues in the leadin or post-leadin.

[Birdofredum Sawin]

Well, Seth, when you say that the answer could be "fire" or "thought"...yeah, maybe it could, if I were confident that the question was written by a good writer. But, a lot of fraud in general comes from the fact that you're at a tournament where you're already pretty distrustful of the questions.

If I go to a high-quality event written/edited by known good writers, I come in with a certain trust that employing fraud isn't going to pay off. That trust quickly evaporates when you hear questions that are fraudable (in any number of ways), contian unimportant trivial clues, pyramidality problems, and the like. Good players very quickly adjust their strategy from "let's play honestly and buzz when I know something" to "let's try to game the questions." Repeated instances of linguistic fraudability are a sure way to get people to lose trust in a set, and then you're left with...frustration! Now where have I heard that word before?

There's something to what Ryan says here, but I think "fraud" is much more widespread and difficult to combat than parts of this thread suggest. As a case in point, consider that a lot of my criticism of Gaddis was on grounds similar to those identified in this thread. (E.g. the Sykes-Picot tossup, which was "fraudable" for a Gaddis audience in much the way that the Carmen tossup critiqued earlier in this thread would be fraudable for a lower-level audience; or the questions which might have been fraudable for anyone who knew the personal predilections of the question writer.)

I'd suggest the following:

1. Fraud is a possible problem at any event, no matter how "high-quality" and no matter how reputable the writers.

2. Fraud is actually a more severe problem at higher-level events, because in general it is only good players who are capable of making the intellectual shortcuts which we're referring to as "fraud" during a tossup. Mediocre, or "casual," players are much less likely to be good enough at the game to make those leaps. I'm not saying fraudable questions are a good thing. Rather, I'm suggesting that a fraudable question in an A-set is much less likely to be "frauded" by the average A-set player than a comparably fraudable question at ACF nationals is likely to be frauded by the average ACF nationals player.

Andrew

[Whiter Hydra]

This issue has actually been discussed among NAQT editors before. The current internal conventional wisdom is that something like a Japanese-sounding name functions much like a clue, and should be placed wherever it fits pyramidally. Knowledge of what sounds Finnish is rare enough that, in my questions, I'm not afraid to put it early. Japanese is a different story, particularly when combined with evidence of CE-ness, which effectively narrows it to Aso Taro or Akihito. English is distinctive, but usually unhelpful due to a huge number of possible answers.

Actually, for IS-79, there was a tossup in the final round that was very easily fraudable (and I think those who have seen it would agree). I don't want to get into more detail on a public forum, but it caused a bit of discussion when we saw it in the War Room.

[Brian Ulrich]

This issue has actually been discussed among NAQT editors before. The current internal conventional wisdom is that something like a Japanese-sounding name functions much like a clue, and should be placed wherever it fits pyramidally. Knowledge of what sounds Finnish is rare enough that, in my questions, I'm not afraid to put it early. Japanese is a different story, particularly when combined with evidence of CE-ness, which effectively narrows it to Aso Taro or Akihito. English is distinctive, but usually unhelpful due to a huge number of possible answers.

Actually, for IS-79, there was a tossup in the final round that was very easily fraudable (and I think those who have seen it would agree). I don't want to get into more detail on a public forum, but it caused a bit of discussion when we saw it in the War Room.

I'm not, of course, saying that there is a common policy. What I outlined was the sense I get of how most editors approach this issue after the inconclusive internal discussions that did go on.

[Captain Sinico]

I find myself agreeing to a certain extent with Shawn here, but I think he's understating the problem. For example, I wouldn't find it acceptable to reveal "this is a Scandinavian children's author" until the giveaway: Indeed, that probably should be the giveaway. I think Andrew's absolutely right in what he's saying, though I'd like to note (or perhaps counter?) that his second point doesn't, in my mind, lift from NAQT the responsibility of trying to minimize fraudability (of whatever kind: I definitely agree with Charlie that arguing about the definition of that term is fruitless,) even in low-level sets because, while it's perhaps less bad there, it's still unquestionably bad.
I also want to express the view that knowledge and understanding of a language is itself an important form of knowledge that is inevitably tested to some extent in quizbowl (albeit indirectly, excepting odd formats.) Therefore, to my way of thinking, it's far from illegitimate in se for a player to use knowledge of the linguistic origin of a clue to deduce things about the answer: in fact, I'd argue that this inevitably happens. The problem comes in when, as has been noted, the origin knowledge is widely held and admits strong guesses about the answer, though I can't define how widely held or how strong. Thus, I'd assert that the linguistically fraudable question and the good question are different only in extent rather than type per se and that the extent difference is imprecise and, to some extent, subjective. I want to stress again that I don't think this absolves a writer or editor from attempting to minimize linguistic fraud (subject to other constraints, for example that outlined by the debate between Seth and Bruce), only that doing so is, rather like most things in question writing and editing, an art rather than a precise science.

MaS

[Mechanical Beasts]

2. Fraud is actually a more severe problem at higher-level events, because in general it is only good players who are capable of making the intellectual shortcuts which we're referring to as "fraud" during a tossup. Mediocre, or "casual," players are much less likely to be good enough at the game to make those leaps. I'm not saying fraudable questions are a good thing. Rather, I'm suggesting that a fraudable question in an A-set is much less likely to be "frauded" by the average A-set player than a comparably fraudable question at ACF nationals is likely to be frauded by the average ACF nationals player.

It's certainly an arms race between the canon's size and the player's 'canon sense'; the average ACF nationals player has far more canon sense than the average A set player to, heck, ten? more?), so, ceteris paribus, they're better at frauding. But they're also more hamstrung by the larger canon. This assertion is true if a player's canon sense grows faster than the canon, which is possible; I don't know if we have any good measure to tell. What we can pretty safely say is that the standard deviation of canon sense is higher among A-set players than among ACF Nationals players. I say that this is actually what creates a greater perception of fraud at low difficulty levels.

[at your pleasure]

Doesn't referring to Aachen put somebody who learned it as Aix-la-Chapelle at a disadvantage? Doesn't referring to it as the Battle of Tannenberg put somebody who learned it as Grunwald at a disadvantage? These things happen in quizbowl, and people just have to deal with it, or learn more things

The difference between this and anglicizing Miklos is that one reguarly sees Achen used to refer to the city in question.

Faults along the west side of this mountain range have led to a chain of volcanoes including Mount Amiata and the Alban Hills...." (18.6) I mean, I hear Amiata is an Italian name and Alba Longa is a famous city in Italy. These are indeed the Apennines.

Assuming that you had learned Alba Longa in a history class somewhwere, I think this is more "knowning a geo question from a non-geo source", which is going to happen regardless of whether or not the question is transparent. If there's a problem there, I would think it would be "Alban Hills" coming too early, which is debatable.

[Stained Diviner]

The Apennines example was not really a good example of linguistic fraud. You could make an argument that Amiata sounds Spanish or Japanese when read aloud, a mountain anywhere in the world could have an Italian name, and if somebody gets the question because they know Alban Hills is in Italy, then that is knowledge rather than fraud. This is a minor problem with an otherwise good post, though.