Santa Everywhere Fallacy
Santa Everywhere FallacySantaEverywhereFallacy is based on the situation, real or perceived, where person A asks person B to do the equivalent of "prove Santa Clause does not exist". By some accounts, this is an unrealistic expectation, let alone impossible on a large scale.
"SantaEverywhereFallacy" is not a formal fallacy, but related to issues involving obligation and BurdenOfProof in debates (the ReversalOfBurdenIssue), whereas 'fallacy' has to do with unsound and misleading reasoning. The name is related to the primary analogy driving the discussion, which considers the (modern) existence of the legendary and mythological 'Santa Claus'.
Of course, if person B had claimed some equivalent of "Santa Clause does not exist", then BurdenOfProof requires that person B should be able to, upon request, provide or point person A towards convincing evidence for that claim. If person B would find doing so an 'unrealistic expectation', then person B should not have made the claim in the first place. (Alternatively, person B can recant the statement equivalent to "Santa Clause does not exist" and usefully issue a new claim equivalent to "We have no reason to believe that Santa Clause has an influence on the processes being described here".)
Discussion Ensues
If I am skeptical that Santa Claus exists, I would say, "You have no evidence that Santa exists". Some seem to think that I am then obligated to show that Santa does not exist. However, this is not the case. The existence of Santa is like a Null or Nil value in a programming language or database. Santa's status does not change from that Null until more evidence is brought to light.
To prove that Santa does not exist (on Earth), one would have to put a camera in every spot and keep them rolling at the same time. Even then, the believer could claim that Santa is dressed in regular street clothes.
If you are skeptical that Santa Claus exists, you should say, "I am skeptical that Santa exists." If you are skeptical that the evidence I possess for Santa's existence is flimsy, you should say: "I find your evidence for Santa's existence uncompelling." Both of these are statements about yourself (thus possessing very low burden), and both of them put burden on the other guy if he cares to change your mind. But if you say, "You have no evidence that Santa exists", you're obligated to support that statement. I.e. you need to support "You have no evidence that Santa exists". A complete, proper job would require showing that: (a) that which I believe to be evidence isn't actually evidence, and (b) that I also don't have any evidence that I don't believe to be evidence.
It is better stated as, "You have PRESENTED no evidence that Santa exists". The proper response of the Santa party would then be to list evidence they allegedly are presenting or have presented.
Incorrect. The accused party MAY present what they believe to be evidence if they wish (especially if they agree with you), OR they may demand that you prove the claim you just made. "You have presented no evidence" IS a stated claim just like any other. And if the accused party believes they HAVE presented evidence, demanding proof is the "proper response". After all, what you just stated also directly implies: "nothing you have presented thus far qualifies as evidence". That counts as an 'extraordinary claim' and requires 'extraordinary evidence' from YOU. Simply parroting themselves, restating that which they have presented, is NOT a proper response to your claim - doing so implies that either (1) the accused party believes that repeating something makes it truer, or (2) that they believe you a simpleton incapable of understanding what they've already stated. If you want them to reorganize their argument to something more comprehensible, it's better to ask nicely than to make accusations.
Whether the presented evidence is any good is perhaps a different issue than this topic. This is about a deflection of burden. The person in the above example would not be obligated to show that Santa does not exist; for that was not their claim. If they are burdened to show that the other person has not presented any evidence, I am not sure how one could practically do such even if it was their burden (for the sake of argument). Non-existence generally is impossible to prove without some kind of scope limit placed on it.
Correct. The issue here is of burden of proof. And the person who claims "You have PRESENTED no evidence" IS burdened to show that the other person has not presented any evidence. If said person CANNOT do so, then they SHOULD NOT have made the claim. That is true of ALL claims: in reasonable debate, people shouldn't open their mouths and make ANY claim that they aren't willing AND able to support with what they honestly believe to be sound reasoning. That includes claims that the other person hasn't presented any evidence. That includes claims that the evidence the other person hasn't presented any 'valid' evidence. That includes claims of non-existence. If you can't prove it, don't claim it.
Now, I'm fairly certain that your statement that "Some seem to think that [you] are obligated to show that Santa does exist" was directed at me. I consider that idiocy on your part. You possess a burden of proof to support the statements you make... and only the statements you make. Allowing, of course, for error: if the issue is poor communication or wording, you ought, rather, to clarify and correct. I don't believe I've ever stated anything that implies you need to support more than your own statements, and if I did, it certainly wasn't intentional..
Agreed. Thanks for improving the wording. --AnonymousDonor
Q: But what if we need a working assumption about Santa now (analogous to picking a programming paradigm for an existing project)? What is the logical way to determining the default or current working value? It seems clear that the existence of Santa stays "null". If one side wants to bump the working assumption to another value, they need provide evidence.
A: In classical logic, when there is data to which you cannot associate a truth value, you simply don't reduce it. You carry 'Exists(Santa)' along through the logic. Sometimes you'll be able to reduce something like 'Exists(Santa) || ~Exists(Santa)' to 'True', or reduce 'Exists(Santa) && ~Exists(Santa)' to 'False'. You can also reach perfectly reasonable conclusions, like: 'Exists(Santa) -> Something' (if Santa exists then something is true), '~Exists(Santa) -> Something Else'.
This is not the same as reducing 'Exists(Santa)' to 'NULL'. If you perform such a reduction, then you cannot perform some of the reductions above because you have lost information. I.e. 'Exists(Santa) || ~Exists(Santa)' reduces to 'NULL || NULL' which then reduces to 'NULL', even though the actual value can be determined to be 'True' by applying the excluded middle. Similarly, you won't be able to reach useful conclusions such as "If Santa Exists then Something is True".
And if you need to make a decision when there are unknowns, tracking the original source of the unknown data still allows one to make useful decisions when hedging one's bets. I.e. if there are 12 unknowns that are each either true or false, then there are 2^12 = 4096 combinations that could be examined, which might be rather expensive to explore. By examining the logic, you might be able to determine that the 'Exists(Santa)' unknown is only critical to just a few combinations that you might be betting against for other reasons, and thus efforts to determine whether Santa exists can be delayed for longer than other unknowns.
Anyhow, there are different logics that deal with unknowns in different ways. Like scientific theories, logics are proven by application in the real world in controlled environments, and are disproved by counter-examples. (I.e. there are an infinite set of possible logics, but only a small subset of those are of any value in this universe.) Examples of proven logics: Intuitionist logic effectively wraps every proposition with 'Provable()'. I.e. 'Exists(Santa)' is unknown, but 'Provable(Exists(Santa))' is false, as is 'Provable(~Exists(Santa))'. There are rules to determine what is provable relative to which other things are provable ('provable' in intuitionist logic isn't absolute; rather, it is relative to a set of data and assumptions). Bayesian logic, OTOH, works with unknowns by applying a confidence to each proposition and applying heuristics to determine the confidence of derived propositions. Bayesian logic trades out repeatability and rigor for speed and ability to work with unknowns. Fuzzy logic is (fuzzily) similar to Bayesian logic, but, where Bayesian logic still asserts expressions are 'probably true' or 'probably false' (truth with varying confidences), Fuzzy logic uses 'partially true' or 'partially false' (varying truth with absolute confidence). Fuzzy logic has not proven valuable in many domains, but it does work well with 'vague' human language and prototype-based classifications (i.e. 'Chair(X)' gets turned into a 'chairness' percentage, such as 'X is 87% chair').
So there isn't "just one" answer about what "logic" does (though one can hedge bets and stick with the classical logic answer). There is one answer for each logic. You can create a logic where 'Exists(Santa)' reduces to 'NULL'... but, if you do, then you'll need to prove that your invented logic is applicable in the real world.
Okay, but sometimes in the real world we cannot carry around the unreduced form. If your manager says, "We need to pick a programming paradigm tomorrow. This is the mandate given to our department." The reduction must be made by the deadline. We have to decide if X is better than Y, and so forth. -- top
When you need to make a decision with an unknown, you ultimately make a gamble. Making a good bet with a time limit is the province of heuristics and decision strategies. Given a data set, logic supports decisions by leveraging data to produce higher-level information, reducing the number of unknowns, and associating probabilities or with unknowns based on predictive models (logic is generally flexible enough to support multiple models simultaneously, though contradictory models may result in an inconsistent logic if one doesn't somehow associate imperfect 'confidence' with models and their axioms). If one keeps the logic traces, logic also helps when determining which data would be most beneficially pursued in further information gathering.
Unknowns introduce risk, and the cost of such a risk is generally estimated in terms of (probability of event * estimated cost of event) under a given decision. While a heuristic could further reduce unknowns to big question marks and leave it at that, most approaches associate various unknowns with costs under each decision (e.g. the cost of a particular requirements change might be different under a choice of one paradigm than another even if the probability of that change is the same, so risk for a given unknown can vary based on the decision). Beyond risk costs, there are also known costs of decisions.
One applies "policy" to choose from among the possible decisions. I.e. based on policy, one can select the decision with the lower estimated probabilistic cost (i.e. possibly favoring the slow-but-sure approach), the decision with the lowest concrete cost (even if it might be riskier), the decision with the lowest (cost+risk), etc. While one may separate the application of a policy from the process of evaluating the decisions, when given a deadline the two processes are usually interleaved as an optimization (such that only the most promising decisions under a given policy are further examined and evaluated). Similarly, one can interleave policy and decision into the logic itself, e.g. by lazy logical analysis of propositions. Further, it all interleaves back into information gathering: if there is a critical unknown that can be measured, discovering it might simplify logic analysis and reduce risk of dependent decisions.
All this amounts to: yes, in the real world you must make decisions with deadlines... but, you would probably be better off keeping unknowns in unreduced form. Doing so helps with decision heuristics and aides you in gathering the right information to further reduce the risks of your decision.
And you'll probably be happy to hear: working with unknown paradigms introduces risk as well as training costs. There are some teeth behind MindOverhaulEconomics. Also, knowing more paradigms and keeping oneself educated gives one more options, but more options also has a cost when it comes to decision making. (If all you have is relational, then every round problem will be forced into a rectangular table... but at least the decision was cheap and easy.)
PageAnchor: Non-Existence
RE: Non-existence generally is impossible to prove without some kind of scope limit placed on it.
Regarding the above: non-existence IS provable under certain conditions. Particularly, you can prove non-existence if existence should provide obvious evidence that is not... evident. E.g. one can reasonably say that "there are no level 4 hurricanes sweeping through Colorado today" by watching the weather channel - if there were level 4 hurricanes in Colorado, we'd hear about them.
But you can't prove that Santa doesn't exist, and you can't prove that the invisible Great Spaghetti Monster doesn't exist. Rather than trying to prove they don't exist when someone claims they do, you should simply recall that their claim created a burden of evidence. You can just say: "I've seen nothing that compels me to believe these exist. Until you substantiate your claim of the Great Spaghetti Monster's existence with real evidence, it deserves no credibility.".
(Regarding the hurricane in Colorado): That's an example of the mentioned "scope limit".
True. I did mention "through Colorado". But "scope limit" doesn't always help, either - e.g. a claim that 'Santa doesn't exist on Earth' is considerably easier to work with than 'Santa doesn't exist', but you're unlikely to get much further with it. The real issue isn't scope limit, but rather the degree to which implied evidence-of-existence should be obvious (i.e. within one's ability to observe it). Simple logic: if A->B, then ~B->~A. If B should be obvious, but isn't observed when sought, then ~B is easy to prove. The 'scope limit' is just an aide to this more general rule: shrinking the scope of a search generally increases the probability that you would find the evidence you seek. But Santa, who supposedly can enter a few hundred-million homes in one night without being seen... you can't even prove "Santa isn't standing in my kitchen, looking for milk and cookies, right now". You'd never know he's there. Sometimes a scope limit doesn't help..
In the case of the level-4 hurricane in CO, one could counter with, "if there was a L4 hurricane in CO, then it would be detectable with existing equipment."
All this is nice, but would somebody please add a summary of what the SantaEverywhereFallacy actually is at the top?
Sure thing.
An slightly related page is ReversalOfBurdenIssue, which discusses the following pattern:
Person A: X is objectively better than Y
Person B: I am skeptical. I have not seen you present any evidence that X is better than Y.
Person A: In that case YOU prove that Y is better than or equal to X.
It is allegedly a misuse of burden rules for person A to turn the burden back on B. It is being compared to:
Person A: Santa exists.
Person B: I am skeptical that Santa exists. I have not seen you present any evidence that Santa exists.
Person A: In that case, you prove that Santa doesn't exist.
The pattern in actually discussed in SantaEverywhereFallacy is closer to:
Person A: Santa exists because foobars are brumbles.
Person T: You haven't PRESENTED any evidence that Santa exists.
Person A: Oh? I don't believe you. Prove that I haven't presented any evidence. (Remember: Person A believes his argument that foobars are brumbles provides sufficient evidence. He can, quite reasonably, be upset at Person T's arrogant dismissal of it.)
Person T: (all indignant) Don't try to turn burden of proof on me! I don't need to prove that Santa doesn't exist! I declare SantaEverywhereFallacy! I'll even go make a new page for it, right now!
Person A: ...
I'm not aware of doing such. If I have, please point it out.
People HAVE pointed it out. You insisted on remaining blissfully unaware. I'm sure opportunities will arise to point it out again in the future. I'm confident you'll do the same as you always did. But at least in the future, I can point you to the page: 'SantaEverywhereFallacy'.
Most of the cases they point out are instances of where I refused to accept anecdotal evidence. I don't consider that objective evidence. I invite you to select a strong example of where I have blatantly done the above. If I am as sinful as you paint me, then that should not be hard. (Please limit it to technology benefit issues, not definitional ones. I'm trying to diet from definition fights.)
As you stated so clearly above, top, this is a burden of proof issue. Your words: "whether the presented evidence is any good is perhaps a different issue than this topic." Even if you refuse to accept examples and anecdotal evidence, you suffer burden-of-proof every time you dismiss it as not being evidence, 'real' evidence, 'valid' evidence, or even not being 'objective' evidence. Further, the issue here is you doing it AT ALL; even if you consider yourself in the right for some other reason (e.g. because anecdotal evidence was used) that doesn't "make it okay" to pretend that a reversal of burden has occurred. And while I could point you towards quite a few past arguments, I think I'll just wait until you do it again. I already have a prediction for what your response will be: deny, deny, no matter what, deny. No need to pollute a bunch of old pages with it. If you think my unwillingness to dig up a bunch of pages that are resting in peace just to start catfights is sufficient proof of your innocence, then smile and grin and continue living your life with your strange fallacy.
And as a slight digression: you're incorrect to assume that examples and "anecdotal evidence" are not "objective" evidence. Anecdotal evidence is, by nature, not stochastic (won't tell you about general truths over trials), but it isn't necessarily "subjective", either. It's okay to dismiss statements of feelings as subjective, but anything more than that becomes quite controversial.
Anecdotes are generally non-dissectible and non-verifiable. I will indeed consider them weak evidence, but very weak. See EvidenceTotemPole.
A convenient abstraction
http://donaldr.noyes.com/ThinkingOutLoud/SantaClausAsAConvenientAbstraction.htm
Truth is rarely a matter of convenience (if that were so, my Bachelor pad would be 'clean enough for company' merely because I desired it so). And LaynesLaw debates where you simply redefine "Santa" until you can see the guy "Everywhere" are true MentalMasturbation. I think your proposal here might be better located in 'ObjectivityIsAnIllusion' or 'VaguesDependingOnVagues?'.
Abstractions can be subjective or relative, and abstractions can be absolute. But once they are utilized in common language, you don't possess freedom to change definitions arbitrarily, and, whether you like it or not, they inextricably become tied into "truth" and "logic". Even given the subjective and relative nature of the statement, my apartment isn't 'clean enough for company.' Similarly, 'Santa' already has a meaning. Even if you find it convenient to see "Santa Everywhere", you'd simply be telling falsehoods when describing a car accident to the police: "Santa was driving Santa and struck Santa when the Santa was red." The invention of new words and abstractions can be performed for 'convenience' - operating definitions make a fine example. Useful abstractions are more likely to make it into common language, and convenience is a fine test of value when you have the power to shape language.
Nonetheless, most other speakers of your language will gaze at you in confusion. You see kindness and generosity and call it 'Santa'. You see kindness and generosity in many places and call it 'SantaEverywhere'. Others hear 'Santa' and think jolly old man with white beard and red suit. Others hear 'SantaEverywhere' and think of jolly old men on every street corner. Words in a given language reference abstractions that aren't so easy to strongarm to the meaning you desire. Of course, word phrases can possess a meaning largely independent of their constituent parts. I can see how 'SantaEverywhere' as a word phrase could adopt the meaning you describe, at least if it captures the people's attention or imagination.
It is evident that it Has.
Related: ReversalOfBurdenIssue | UsefulLie | AllAbstractionsLie | 
EditText of this page
(last edited January 10, 2009)
or FindPage with title or text search