Variations on the Milgram Experiment, and Trolls. (Thought Experiments)

[psiberzerker]

This is going to be TL: DR, and nerdy.  First, let me describe the basic Psyche 101 test procedure.

What you'll need:

Volunteers.  A test environment, where they can be isolated, a small room will do the trick.  A lab coat, a clip-board, and a "Buzz Box."  (Or just copies of the Milgram test data, and a basic background in Statistical analysis, since the tests have been run several times, by several researchers over decades since the initial study was published in 1974.)

The Buzzbox can be as simple, or elaborate as you want.  In the interest of eliminating variables (Occam's Razor) let's start with a simple switch, a dial, wires leading to the wall, and a nice gauge marked Voltage.

The "Test Subject" could be an actor, a recording, or even the Tester yourself.  If you have a Soundproof Booth, then you can just conceal a microphone, and wire it to a speaker in the wall.  Otherwise, you can record the Responses, including sounds of Pain, and play them back as the Volunteer goes through the test.

NO ONE GETS SHOCKED.  This is important, in the test, the "Subject" at no point receives a real shock, whether an Actor, the Tester, or a recording.  The only shock is delivered to the Gauge, to make the needle jump, the wires aren't even wired up to the power supply.  They're dummy wires, at no point is anyone actually causing Physical pain, but the Volunteers are stimulated with Sympathetic pain, by listening to the screams.  Believing them to be a response, though they are actually the Stimulus to see whether they hesitate to "Shock" the one asked the questions again.

Procedure:

Tell the Volunteer that you're testing Negative Feedback, and how it affects performance on Oral Exams.  There's another volunteer in the next room, and you're going to ask them a series of questions.  Either you ask the questions, or you can have the quiz printed out for the Volunteer to ask them, with the answers provided.  Every time the "Test Subject" get's something wrong, you are to push a button to deliver a shock.

Stimulus, Response.  That's what Milgram was testing, though his conclusions were mapped out in the Abstract, before testing was even begun.  He claimed to be testing Obedience to Authority, on the assumption that the Lab Coat made him (Or the Tester) an Authority, for the purposes of the test.  

I suggest another possible Stimulus at play here, which I'm going to call Mental Momentum, and Emotional Inertia.  I probably don't have to tell you how the experiment turned out, the vast majority of the Volunteers kept delivering shocks, long after the screams became loud, and the "Mistakes" (Part of the setup) became so frequent that they were "Shocking" the "Test Subject" every answer.

Stimulus Response:

It's quite possible, I'd even call it plausible that one of the most measurable reasons why the real test subjects (The volunteers) kept pushing the button is because of all the times they pressed the button before.  "For science," it's also possible that there was some Sadism at play, and deep down we're all sadists at heart.  Another possibility is we're all predisposed to Obedience to Authority, as Milgram positied in the Abstract, but that's what he set out to prove in the first place.  He set up this experiment to test how random volunteers follow Orders.

However, he used Volunteers.  So, I have to wonder what kind of person shows up, voluntarily to test something like this?  Someone who's interested in Science?  Someone who's mentally, and emotionally predisposed to Obedience, or Rebellion?

I think we can eliminate Rebels from the list of people who're likely to volunteer for scientific experiments.

[psiberzerker]

Variation 1:  The Rigged Casino.

This is the one test I helped with myself.  It was represented as a study into the Science of Gambling, and Gamer Theory for Control.  (If the volunteers know what they're testing, it skews their behavior, which you're studying.)  This was an exercise of Occam's Razor:  Removing variables which may create noise in the data.  Basically, the Simplest Experiment yields the cleanest data.  The results hypothetically have less noise from things like Sadism, opinions on Science, Authority, and our roles in a world of Technology aren't what we're testing.  

Stimulus

The mock Casino is set up with Slot machines, with a lever you can pull, 3 dials which spin to line up, and some sort of reward that pays out when you line up 3 of a kind.  (We used Alcohol.  Free shots of your choice to purposefully add a variable, and an Incentive.  For consistency, the Shots were all 40% Alcohol by volume, and the test subjects were also given the choice of Mixers, or nonalcoholic chasers, if they so desire.  This is also a variable in real Casinos, where free drinks are often provided for the guests, to keep them in the seats, pulling handles.)

Risk/Punishment, and Reward.

The slot machines are, of course rigged.  One of the local Casinos (Incidentally, the study was conducted by the University of New Mexico, Los Alamos.  The machines came from Jemez Pueblo, and I was onboard in the Setup of the study.  I didn't participate.)  Instead of having an actor, or a recording that the volunteer was instructed to Shock, each machine was set up to deliver a safe, irritant level current to the Handle, every 20 pulls.  Not at random, the dials were real, the random variable was the payoff, but the Shocks were at the same current, and exactly every 20 pulls of the handle.  

The volunteers continued to shock themselves.  There were empty machines, at all time, and multiple volunteers in the mock Casino, but there was never a time when all of the machines were taken.  The volunteers were allowed to stop at any time.  Switch machines, and many did after getting up to report that the machine shocked them.  Some left, some didn't, but regardless, none of them was able to Game the system by picking up on the pattern.  (It was noted that if you realized that the shocks came every 20 pulls, you could pull 19 times, and then switch machines.  Nobody did.)  Not one abandoned the experiment on the first shock.  They were told "That's why the machines were recalled," and they were given the option of quitting, after they found out that ALL of the machines have this flaw.  Many didn't even bother to ask, nor give up until they earned their first drink.

The interesting result was the people that kept pulling the handle after the Payoff.  Just like in a real casino, there are compulsive gamblers who will continue pulling the lever, after they win the payoff, because at some point they're there to pull the handle.  Regardless of how many time they played to Bust, ran out of money, and were forced to quit (Since there was no money involved, this variable was removed) they still returned, and went through whatever rituals they believed to bring them "Luck" until they hit the jackpot.  Some of the truly habitual gamblers will win a jackpot, and keep playing, until they run out of money, and are forced to quit.

At some point, you lose count.  How many times before I won the jackpot?  How many times since the last shock?  It isn't so much that we are reduced to rote action, switch pullers we could design a machine to replace.  It would be simple enough to create a machine that pulls switches, without the pain of continuing to go home broke, and tell it'self "Better luck next time," but we weren't testing Luck.

For lack of a better term, we're testing the Rote Response.  The longer someone participated, no matter how drunk (They were administered Breathylizer tests before each shot/mixed drink.  We also told them that the cutoff was 7% BABV, when they'd be too drunk for useable data.  That's 1% below the limit to drive in the state of NM at the time, and we also had shuttle service provided, if needed.)

I'm going to stop here, because this is just another example, and the Thought Experiment is even more complex.  So, in the interest of Brevity...

[psiberzerker]

"Here come the test results: 'You are a horrible person.'  That's what it says: 'A horrible person.' We weren't even testing for that."

~GLaDOS

http://www.youtube.com/watch?v=ohDB5gbtaEQ

Trolling

I'm a troll, so let me speak as a troll, and not a Volunteer.  Nor a test subject, nor the Experimentor, nor an actor pretending to be shocked when someone makes a mistake.  It takes one to know one.

Stimulus Reward

At first, it's just a simple comment, maybe a reply.  Oh look, someone mispelled 'Incorect.'  Maybe they would like to know that they were wrong.

Pull the switch.  Oh, they didn't like that, there must be something wrong with them.

Pull the switch.  Oh, someone is backing them up, it must be a clique of people who don't care about proper diction as much as I do.

Pull the switch, twice.  One for each of them.

At what point does the troll realize that he, or she is trolling?  Is it after they start seeking our grammatical errors in the comments?  Whatever the threshold for it to begin the be considered "Trolling" is, at what point does the Troll realize that we are looking through every comment for Speling Erors, and ignoring the content of the message he's spellchecking for?

At what point do re realize that the vast majority of people, on the internet have access to spelcheck, and auto-corect?  Now, let's do another experiment to test that, shall we?  My favorite Stimulus to detect Spelltrolls is to intentionally mispel words, from a fairly small set:

{Spelling, and it's variants.  {Spell, Misspell...}

Correct, and variants.  {Incorrect...}

Error...}

What does this prove?  Well, first let me state explicitly why I chose those words, in particular:  The premise is that if someone is so intent on finding erors to prove you're wrong that they will overlook the obvious nature, and reply with "You spelled misspell wrong" then you know that they are there simply to proofread the comments.  

They will ignore everything else, you're saying, as long as there's 1 flaw in your reply.  Even if it's an unrelated flaw,

Pull the switch:

At some point, a troll will meet a troll, online.  I shouldn't have to prove this hypothesis, you're in 1408, you've seen 2 troll beat each other's heads against each other, each insisting the other one's head is a wall.  In the original Milgram Experiment, there was also a dial, and a gauge, in addition to the switch.

Pull the switch, each time the guy in the Lab Coat, we'll say Dr. Milgram reaches over to turn up the dial.

Pull the switch, and the needle on the Ammeter goes higher.  The actor screams louder, and so forth.

1408 removes the Actor.  Here, everyone is a volunteer, with a switch.

[psiberzerker]

Variations (Trolls)

This is a multi-axis categorization, to better understand the Kind of troll you're butting heads with.  Which stimulus they respond to, and how to get the response you want out of them.  So, starting at 101, the first 2 axes are Stimulus (Tactics) and Response (Motive.)  What do you do, to get what you want out of them, and conversely, what they do hints at what they want from you.  Beyond a Reply, that's just binary to the act of trolling.  "The only winning move is not to play."  

~WOPR

{B.A.S.I.C.

The Spelcheck troll is probably the easiest, based on the Turing Test:  If a 'bot can interact with a user, without telling the difference, then it might as well be as Intelligent as the User.  (Not AI, UI.)  Your B.A.S.I.C. speltroll can be replaced by the same proofreading programs that we all have on our word processors, auto-corect, and the like.

Dictiontroll:  This is a more advanced (1.1) strategy, where the automatic proofreader lines don't pop up when they hit "Quote" to point out the speling erors for them.  This would require a more sophisticated 'bot to pass the Turning Test, basically Grammarly, or Dragon.  (So, the C++ equivalent, on the hierarchy of Programming Languages.)

Binary:  This is your False Dichotomy, either I'm right, or you're wrong.  Heads I win, tails you lose, with no concept of a draw, where you both lose.

FORTRAN:  This is the troll that Compiles a set of Statistics.  A limited set, where the Confirmation Bias is preserved, throwing out any outlying datum which would skew the conclusion they're after, in the Abstract.  This is Occam's Guillotine: "Deciding the conclusion, then looking for the Evidence to Prove it is like cutting off your own head, so you don't have to shave."  The opposite of Occam's Razor, rather than removing steps to refine the Data, they remove the Experiment completely, and refine the Data (Statistics) to the set that correlates with their assumptions.  Syntax Error: The numbers don't add up to the "Truth" as I see it.  (Alternate speling is 4tran, to allude to the message board where it was first discovered.)

Redditorial:  Your basic Strawmaster, "So what you're saying is;"  That's how they declare their approach to your argument.  They don't even bother to read the entire comment, they skim for the Keywords, and edit out the ones that they don't deem important.  That's how you get "Exonerated" from "Not Exonerate," or the full predicate "Neither exonerate, nor provide sufficient Probable Cause to Press Charges."  All of these are motivated by the same thing:  Proving you wrong, to prove themselves right (Or Write in the case of the Autocorect troll) escalating to more advanced tactics, to strategies, and ever more complex "Truths."}

Within this Set, it is assumed that they have decided the Conclusion, whatever it is, and are only arguing with you to steer the conversation around to their "Truth."  Everything else in the debate is discarded, out of hand, as log as it comes back around to their assumption, which they will restate, until you give up, or agree with them.  There is no other option.

;

Triggers

This is the other (Why) axis, where they pull the switch often enough that the act of pulling the switch is all they need.  Quite often, they announce themselves in the Abstract, like a hot chick in a sex story describing herself bra first.  "36D-24-36."  That's all that matters, and they will use any excuse, or none, to spam that button.  They only have a hammer, so every argument looks like a nail, even when it's actually a Dead Horse.

Once is an instance, twice coincidence, 3 times a pattern.  So, if they spam that button 3 times in a row, you have reason to expect that's how they score points.  Every troll keeps score, but one of the tricks is not telling the other trolls what counts in your rulebook.  If you can figure out what their Hot Button is, then you can try to disable it, so they can't just give themselves free points.  if they immediately post this:

or something like it, then you know they score points by how many people they can "Trigger."

Sets:

{"No U!"  Doesn't matter what you're proposing, the simplest way to avoid an actual argument is to accuse the other, of whatever.  It often belies a subconscious flaw, they want to avoid, so if they don't want to sound like a racist, they will preempt that by calling you a racist, reverse racist, or accuse you of equating everything with racism.  (Sexism, or whatever your argument is about.)

Ad Hominem:  It doesn't even matter what the keyword is, it could be as simple as calling you a "Troll."  I'm not a troll, you're a: {Liberal, Pinko Commie, Nazi, Sociopath, ...} whatever.  The inverse accusation is "You always call everyone {X} when they don't agree with you."  This is separate from Godwin's Law, which is a Sociological (Forest) trend, not a Psychological (Tree) trait.  However, if you use the terms {Nazi, Fascist, Holocaust/denier, Concentration Camp...} then expect to get dinged for a violation of Godwin's Law by someone that thinks that's an automatic WIN.  

If they call you a "Snowflake" 3 times in a row, then you know that's all they've got.  You're a snowflake, everyone's a special little snowflake (Except me, of course.)  You are an unique individual, just like everyone else.

"Eat shit, go fuck yourself, shitbag.  Douche, go back to where you came from, I hate you, you make me sick, you're just a..."  That's when you know they either don't have an argument, or regressed to the point that they can't articulate it any more eloquently than a 6th grader picking on younger kids in the Elementary Schoolyard.

Lord of Chaos:  AKA "Edgelord..."  This is your standard troll, who's just here to read the comments...  They stimulate the forum by throwing out something controversial, then sitting back, and keeping score on the raging argument they created.  You can generally tell, because they only come back to contribute something meaningful (More so than standard buttons like memes, slogans, or whatever) when the argument begins to die down.  So, look for them, if you want to chase them around the forum, in places where there's Partisan subjects like {Politics, Religion, or Sex.}  Basically anything you shouldn't talk about at the dinner table, because it's going to be a fight?  Yeah, if it says {Race} in the title, {Abortion, Guns, Evolution...}  This is why certain topics, or proclamations are considered "Edgy."  For instance, here you have {Pedophiles, Rape, Incest, Necrophilia, and Bestiality.} So, we attract a lot of Edgelords, 4 teh chaos!  Lulz.

Sealioning:  This is getting more advanced, and for the sake of Clarity, I use it all the time.  (I'm not a Sealion, I play the Role of Sealion, when that appears the best way to approach a given argument.  A Sealion either defaults to this as their first strategy going in, or doesn't know any other way to argue, so uses it even when it's counter-productive.)  It's complicated, rational, and appears to be Civil, when it's not.  Typically involves Gaslighting  "You're insane, which makes me the voice of reason here."  Also backhanded compliments, and "Sideswipes" (Forehanded insults.)  A sealion gets our points when the other guy gets frustrated, and less intelligible, making us seem more civil, and coherent in contrast.  The best counter for Sealioning seems to be feigning a breakdown, until they score a point, then returning the question they're avoiding, but you want to hold onto it.  I don't get win debates, I have no intention of "Winning," because then the discussion is over, and I have to find someone else to talk to.  Because eloquence, and rationales are how they keep score, Sealions fighting tend to be the most long winded, and detailed as we put the ANAL in analysis, but I still score points when the other guys gives up trying to argue, and devolves into one of the more basic trolls (Above.)  It just doesn't make up for the interesting discussion being over.  (That's still a Loss in my scorebook, so a pyrric victory at best.)

...}

[psiberzerker]

Pathos

Not to be mistaken for any_1(31) in particular, but this isn't a kind of troll.  This is a kind of Purina Troll Chow:

The trait of being Pitiable, or Wretched.  Specifically a Sadistic troll, who just wants to cause pain and sufferering Shadenfreude, or Edgelords who thrive on chaos, because a Victim brings the Sadists in to Feed.

This is the perfect couple, a Sadist, and the mythical Masochist.  Likewise a self professed "Dominant," and the long sought after submissive that exist only to satisfy His (Almost inevitably) every whim, however when He doesn't get it, He may claim it instead.

"You just want to Play the victim."  Look into my eyes, you want to be My victim...  That's your Dominant, seeking Pathos.  When they actually run into someone who just wants to play the Victim, that's when they hit the Jackpot or they just convince themselves that she only wants to play the Victim.  "She was asking for it, look what she was wearing!  She's a slut, and you can't rape a slut, they enjoy it too much."
 The stereotypical Edgelord Incell assumes that this is the default "Female" state.  There's 2 kinds of people, victims, and Me.  Not that they'd stoop to playing the victim, of course.  That's just a False Allegation.

As a literary term Pathos denoted the Mask of Tragedy.  You know the classic Thesbian masks, Comedy, and Tragedy?  Pathos appeals to Emotion, Athos is a joke, Ethos (Or Aramis) appeals to Truth, or the Authority of a Higher Power, and Porthos is just a glutton for punishment.  

Not his punishment, of course.  That's what he needs Pathos for, to feed his appetites.  Nobody truly exists for Pathos, and survives very long.  There's too many Predators out there, "A fool and his money were lucky to get together in the first place."

~H. Anderson

D'artagnan just wants a fight, and doesn't care about your cause.

#Resist.

to someone, that would make Joan Milady de Winter, in his mind.

[psiberzerker]

Rules Lawyering, or Memberation.

This is an attempt at a Win button, for instance Godwin's Law.  At some point, Mike Godwin himself was called on calling detainees "Concentration Camps," so the "Law" was expanded to the odds of someone being cited for Godwin's Law reached 1:1.  (Godwin being cited his Own Law.)

This is where the scorecards are tallied up, but remember that each Troll scored their own points, and determines themselves the Winner.  "I win, you lose."  That's what happens when the players keep score, and make up the rules when they go along, at some point Clive Anderson, or Dara Ó Briain has to hit the buzzer so someone else gets to take a turn (And is Mocked from both sides for being Bald.)

However, this doesn't settle anything, honestly.  It changes the subject, to basically "Your a troll," no "U R!" but that's the point where the Debate ends, everyone loses.  

As with the other Tactics, there's warning signs to look out for, only this set consists of "The Rules" as interpreteded by the one who smelt it.  (Meaning, he's playing That Card, and pretending to be the Dealer, so he doesn't have to Ante up, to claim the Pot.)

{Godwin's Law:  As referenced earlier, it doesn't mean that the opponent automatically loses by default, when they utter the N word.  Nazi=/=Nigger.  Pardon my French, but that's the false equivalency here:  Using the latter "N word" makes you a Racist, and reveals that you're argument is basically racist.  Discussing the Third Reich as an analogy of the worst case scenario of totalitaran government ruled by racists, to justify crimes against humanity, and trying to take over the world just signifies that the debate has gone on long enough to notice some parallels to the most famous example in human history.  It doesn't count as a point, to either side, it's closer to the point where most of the Pawns are sacrificed, and cleared from the board.  "Checkmate, libtard."

Occam's Razor:  "The simplest answer is most likely the right one" is a logical fallacy.  I already used this one in the OP, which is what it actually applies to:  A thought experiment, not a debate between 2 pedants over who's got the best vocabulary, and diction.  Basically, the more complex the example, the more variables are involved in the equation, and the more likely that some variable or datam will be discarded as insignificant due to Human Eror.  That's really one of the central points on this multi-axis correlation:  It's not that simple, it's never that simple, and you'd have to be pretty damned simple to believe that something as complex as inter-national online communications on the sociopolitical ramifications of the escalation of animosity can be summed up by a meme, an image macro, or a #Hashtag.

;}

#AllLivesMatter well, actually, #NotAllMen.  Both of these were coined to counter an overly simplistic argument by adding the word All, or Not All.  On the implication that the speaker, whether a Feminazi, or a Reverse Racist just forgot to say All Men, or implied that Black Lives matter more, instead of Less.  Which is what #BLM stated, explicitely from the start.  Black Lives are being treated as expendable, the death of unarmed black men aren't treated as a tragedy, and any black voice that speaks out about it is treated as a reverse racist.  A half black/Pollak kneeling respectively, in silence was taken as an afront to America, because he dared to respect the African side over the American side.

Just for 2 examples, an over-simplification is an oversimplification.  Ever since #Manslpaining was coined, trolls have been trying to coin ...splaining to compare any intelligent discussion to a straight white guy talking over a man of color, woman, gay man, or transwoman (Just to isolate the different prejudices without compounding it by intersectionality, but if you want all the axes together, a mixed race polyamourous intersex atheist.)

The simple answer, to every debate is I am superior ∴ you lose.  Or the inverse, you're inferior ∴ therefore I win.  Both trolls keep score, and the net result is 0.

1:  Start

;

1986: Execute Clock.

1999:  If Clock=line+1 then GOTO 1986.

Rinse, repeat.  That's why the same argument happens every day, on every board, Twitter Feed, Facebook wall, News Channel, Editorial, middle-school playground (After school) and senatorial debate.  

Because that's the central hypothesis:  I'm right, you're wrong, and the one who "Wins" is the one that's stubborn enough to keep insisting on that, the loudest, long enough for the other guy to give up.  

#NoOneWinsOnTheInternet.  No matter who keeps score, you want the Occam's Tool{Set} in a nutshell?  The WORD OF GOD(win) condition?

^That's it.  If anyone ever truly won, then we'd have to stop fighting, and find something productive to do with our time.

{*Occam's Crucifix:  "You're a monster.  Sacrilege!  Mertyl, get the torches, I already got mah Pitchfork right here..."  Dogma.

Occam's Pitchfork is "So what you're saying is:\;"

Occam's Torch is "A witch, burn her!"

Occam's Noose is "You gonna burn that cross?"

Occam's Teacup: "The simplest explanation is Aliens!" is just a rewording of what Godwin's Law mutated into.  "The easiest win condition is because Nazis."

Punchline:  ^  "Because it rests in a saucer."}

2000: /ENDIF.

[psiberzerker]

Millgram's Dungeon

This setup is merely a thought experiment, where you give each Volunteer 2 buttons.  Tell them 1 is wired to electrodes, and the other is wired to a Sybian, then tell the actress "We're about to begin."

Inside her chamber, just put 2 lights, so she can see let's say Red for Pleasure, and Blue for Pain.  The volunteer can hold down either, or both buttons, as long as they like, and the intensity is set.  (On, or off.)

Now, to the Volunteer, it doesn't really matter as long as she's a good enough actress to convince him.  The default state is that both buttons are off.  He also has a microphone, but no pre-scripted questions, no right or wrong answers, he can ask her anything he wants.

Just get [PAIN] "You like that?"  Out of the way.  Every one tries it at least once.  That's the con/troll, then what?

You don't even have to Observe this experiment, because it's a Thought Experiment.  There is no Cat.  Just for the sake of argument, put yourself in the place of the Volunteer, and consider what questions you would ask her.  Once the guy in the labcoat leaves you alone, what stimulus you want to give her, and what feedback you want from her?  There's no way of knowing whether she's lying or not, or even if the buttons do what you were told they do.  What if they're backwards, the Blue one causes her pleasure, and the Pink one shocks her.  How would you find that out?  What questions would you ask to get that Truth out of her?

There is no wrong answer, because these are the questions you have to ask yourself.  What am I here for?  What do I want out of her?  This is a blank canvas, you have Carte Blanche, just like you have here, in 1408.

You can't win, and since she's an actress, the buttons don't do anything but tell her how to act, the only things you can learn from this are about yourself.  

Here's the problem, the dilemma in this paradox:  Anything you bring with you besides the buttons in front of you, and the questions you want to ask, skews your interpretations of her answers.

If you go in with the hypothesis that she must be a troll, a feminist, or a Nazi, or a White Supremacist (When you can't see her face, let alone the color of it.) A fat guy sitting at his computer catfishing men on the internet by claiming to be a bisexual woman.  What color are her pubes, does she have any, and if so, do they match her scalp?  Her eyebrows, if any, is she wearing any makeup?  What color are her eyes?  Is she a criminal, a virgin, a hedonist, or a prude?  Too fat, too skinny, too old, too young, too bossy, or high maintenance?  A slut, or too shy to speak up for herself, and make eye contact.  A masochist, or maybe she just like to play the victim.

How would you know?  You've got pleasure, and pain control, right here at your fingertips, you just have to think of the right question to ask.  If you start with what you assume, coming in, and only believe what confirms your personal "Truth," you lose.  You learn nothing in here about the only person that really matters:  Yourself.  You're doomed to repeat the same argument with yourself (And your impression of other people you only yell at in plaintext) that you came to this forum with.

Ask, don't assume.

[psiberzerker]

Q&A

Again, there is no Conclusion, the entire Thought Experiment is an Abstract.  A Koan:  "Do you understand Zen?"

The first question you need to answer is "What kind of troll am I?" then have the self awareness to realize what shuts you down.

If you're not a troll, the first thing you answer is "What kind of troll are they?"  (Don't ask them, they will lie.  "I'm not a troll, you are!")

Trolling is B.A.S.I.C. binary.  Stimulus Response.

A discussion is Question, and Answer.  Listen and respond, both ways.  If you don't read their answer, you're just a troll.  It isn't a dialog, it's 2 monologues.

Now, how do you stop a Sealion?