Why “Loser’s” Compromise?

[Update: I forgot to include the link back to Lifeguard's original comment; fixed now.]

Well, I’m back, sort of, and from the looks of things you guys didn’t miss me too much. I don’t suppose I’ll ever catch up on the comments backlog, but I’m sure you will let me know if there are any important points I’ve missed in my quick skim.

Meanwhile, I did notice this interesting comment (stuck in the moderation queue) from a commenter by the handle of “Lifeguard.”

I guess what I’m struggling with here is what the exact difference is between the Loser’s Compromise and simply acknowledging the very real possibility that despite the certainty of your beliefs you may be mistaken about which conclusion is the most justified, the best of the bunch, to say nothing of absolutely proven to be true?

That’s an excellent question, and I’m happy to have the opportunity to explain this further.

There’s a big difference between the Loser’s Compromise and the reasonable practice of acknowledging a certain margin for error in one’s conclusions. In the latter, the goal is to keep one’s mind open in order to be receptive to receiving new information that might change one’s conclusions. The goal of the Loser’s Compromise, by contrast, is to deprive us of the ability to benefit from new information, or even already existing information. The loss of this ability to distinguish between truth and falsehood is precisely what makes it a “Loser’s” Compromise—we’re trying to lose a faculty we could otherwise use to learn that our beliefs are already false.

The Loser’s Compromise is, in effect, the exact opposite of admitting that there’s a real possibility we could be wrong. If we are wrong, the only way we’ll ever find out is by noticing that the evidence is inconsistent with the conclusions we wish to believe. The whole point of the Loser’s Compromise, however, is to make the evidence sound equally consistent with all conclusions, thus causing us to lose the ability to identify incorrect conclusions.

The feature that makes the Loser’s Compromise stand out as a rationalization, and that betrays the compromiser’s motives, is when we try to use the Loser’s Compromise to claim that we have a justification for our beliefs, despite the fact no such justification exists. If the evidence fails to favor one conclusion over the others, then they are all equally UNjustified, not equally justified. That’s an important distinction, because when the evidence is uniformly ambiguous, the only position that can be legitimately justified is agnosticism, not belief.

Now, there may indeed be circumstances in which the available evidence is insufficient to distinguish between different possible conclusions. I would not use the term “Loser’s Compromise” in such situations, provided that we were openly agnostic about our conclusions and that we were actively seeking more evidence and information with the goal of ultimately discovering which answers were right and which were wrong. The term “Loser’s Compromise” only applies to the specific case of trying to make the existing evidence sound inconclusive, via arguments intended to deny or distort the facts, in order to avoid acknowledging a clear inconsistency between the available facts and a particular conclusion.

So yes, I endorse the practice of acknowledging the possibility that one’s conclusions might be incorrect, and that new information might invalidate previously-held beliefs. I myself could be wrong about heliocentrism, or about God, though at this point I’d say the odds would appear to be about equal in either case. Acknowledging the possibility of error is a good thing, but it’s the exact opposite of what the Loser’s Compromise attempts to accomplish—lip service to human fallibility notwithstanding.

 
1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
Loading...Loading...
Posted in Comment Rescue, Loser's Compromise, Realism, Unapologetics. 27 Comments »

27 Responses to “Why “Loser’s” Compromise?”

  1. ssjessiechan Says:

    I missed you a lot, actually! I really look forward to reading your posts, so when you’re gone I just check a lot and think, “gee, I wish there was something new here to read.” Your perspective is entirely new and unique to me, so I get a good deal of pleasure out of each post. ^^

  2. John Morales Says:

    DD, a criticism regarding applicability and another regarding your logic,

    If the evidence fails to favor one conclusion over the others, then they are all equally UNjustified, not equally justified. That’s an important distinction, because when the evidence is uniformly ambiguous, the only position that can be legitimately justified is agnosticism, not belief.

    I disagree, inasmuch as I consider justification categorically different to preference. As it seems apposite, I quote something I wrote to cl at his blog in the context of challenging his similar contention that, in real life, there can be “multiple criteria upon which competing hypotheses may be preferred – I mentioned parsimony, explanatory power and predictive power, as examples.”
    As with his, yours is a claim regarding the possible existence of multiple hypotheses consistent with a given dataset about which there may not be a justifiable discriminant.”

    Let me give an actual example – in real life, for example, I am agnostic as to whether there exists telos yet, by the principle of parsimony, in functional terms I can be said to “believe” there is no telos.
    Even if the evidence is ambiguous, I have chosen to not believe in its existence, and I further believe justifiably so.

    Regarding “… they are all equally UNjustified, not equally justified” – you do realise that both in logic and in natural language, if two things are equal in one aspect, they must be also equal in its reciprocal?
    (e.g. when two people are equally old, they’re also necessarily equally young).


    NB at which point they’d be classed as theories.

  3. John Morales Says:

    Hm, my quoting went a tad awry, but I hope it’s clear nonetheless. I also want to clarify I refer to my epistemology and pragmatic provisional belief respectively when I speak of both being agnostic yet believing.

  4. pboyfloyd Says:

    John says, “(e.g. when two people are equally old, they’re also necessarily equally young).”

    I think that when we are talking about really old people, when we say, “They’re both 90 years YOUNG!”, we’re joking(a la, “It’s funny because it’s true!”)

    That goes for the justified/unjustified thing too.

    The pragmatic solution for myth/”possible esoteric reality” is to, since we’ve already discounted hundreds of gods AS myths, is to go ahead and discount these last hangers-on too, isn’t it?

    Seems to me that if we’re justified in not believing in Thor, Osiris, Hera et al, then we can claim some justification for lack of belief in Yahweh.

    cl could easily be trying to justify Santa.

    Just ‘saying’ that things are equally justified or unjustified is fine unless, it seems to me anyway, we are talking about hypotheses which ARE unjustifications of each other.

    Duncan is saying that B’LIEF is unjustified because it’s a myth but cl isnt trying to say that reality is just a myth, rather cl is saying that believing Christianity is a myth is unjustified.

    But turning tables doesn’t make Christianity any more unmythlike.

  5. ThatOtherGuy Says:

    @John

    I think the snag you’re hitting with “equally unjustified vs. equally justified” is implications. To be perfectly correct, yes, two things that are equally unjustified are also equally justified. However, both describe different directions in magnitude. When one says two things are equally justified, there is an implication that both are in the black, as it were. Saying they are equally unjustified is semantically important, because it implies that both are in the red.

    That’s basically what I think DD’s getting at. You’re correct in saying that both terms describe the same thing, but the connotations of each are very different.

  6. R.C Moore Says:


    Regarding “… they are all equally UNjustified, not equally justified” – you do realise that both in logic and in natural language, if two things are equal in one aspect, they must be also equal in its reciprocal?
    (e.g. when two people are equally old, they’re also necessarily equally young).

    Axiom 1: I always lie. (A paradox, but a valid construct)

    Axiom 2: (the reciprocal) I always tell the truth.

    Set A is a set of all constructs equivalent to Axiom 1.
    Set B is a set of all constructs equivalent to Axiom 2.

    Set B is not the equivalent of Set A, there is in fact no mapping function from A to B, as A is a paradoxical set.

    I prefer Bayesian logic, where one does not get caught up in semantics and logical traps.

  7. cl Says:

    The goal of the Loser’s Compromise, by contrast, is to deprive us of the ability to benefit from new information, or even already existing information.

    I submit that such has never once been my goal, despite all your annoying implications to the contrary.

    There’s a big difference between the Loser’s Compromise and the reasonable practice of acknowledging a certain margin for error in one’s conclusions.

    Now we’re getting somewhere..

    The whole point of the Loser’s Compromise, however, is to make the evidence sound equally consistent with all conclusions, thus causing us to lose the ability to identify incorrect conclusions.

    I also submit that despite your annoying protestations to the contrary, making your GH and MH entail identical predictions has never once been my goal.

    If the evidence fails to favor one conclusion over the others, then they are all equally UNjustified, not equally justified. That’s an important distinction, because when the evidence is uniformly ambiguous, the only position that can be legitimately justified is agnosticism, not belief.

    I define agnosticism as lack of knowledge, and I strongly disagree with you here. The very core of rationalism is to accept ideas that are supported by evidence, and this should not change just because more than one idea is currently supported by evidence. When two or more hypotheses are equally consistent with all of the available data – although provisional belief in either would be rationally justified – truth claims remain unjustifiable until further evidence favors one hypothesis over another. Truth claims are entirely different philosophical beasts than rationally justified beliefs.

    I would not use the term “Loser’s Compromise” in such situations, provided that we were openly agnostic about our conclusions and that we were actively seeking more evidence and information with the goal of ultimately discovering which answers were right and which were wrong.

    I do not know whether or not God exists, but I believe God exists, and I am always actively seeking more evidence and information with the goal of ultimately discovering which answers were right and which were wrong.

    Can I get just a little respect now, seeing how I’ve demonstrated that your “Loser’s Compromise” relates zero to my arguments?

  8. nal Says:

    John Morales:

    Regarding “… they are all equally UNjustified, not equally justified” – you do realise that both in logic and in natural language, if two things are equal in one aspect, they must be also equal in its reciprocal?

    True, but, I think there’s a point-of-view difference. If two views have zero justification, the claim they are equally justified, while technically correct, puts the emphasis on the wrong side of the reciprocal.

  9. pboyfloyd Says:

    Anyone want to talk about whether cl ought to get more respect or disrespect?

    (ribbit.. ribbit..),(… … …),
    (tweet… tweet…)

  10. R.C Moore Says:


    I define agnosticism as lack of knowledge…

    Why, are you unable to access a dictionary?

    From wikipedia:


    Agnosticism (Greek: ?- a-, without + ?????? gn?sis, knowledge; after Gnosticism) is the philosophical view that the truth value of certain claims — particularly metaphysical claims regarding theology, afterlife or the existence of deities, spiritual-beings, or even ultimate reality — is unknown or, depending on the form of agnosticism, inherently impossible to prove or disprove. It is often put forth as a middle ground between theism and atheism,[1] though it is not a religious declaration in itself, and the terms are not mutually exclusive, since agnosticism refers to knowledge, while atheism and theism refer to belief.[2]

    Oh, I see, it is because the actual definition makes


    The very core of rationalism is to accept ideas that are supported by evidence, and this should not change just because more than one idea is currently supported by evidence. When two or more hypotheses are equally consistent with all of the available data – although provisional belief in either would be rationally justified – truth claims remain unjustifiable until further evidence favors one hypothesis over another. Truth claims are entirely different philosophical beasts than rationally justified beliefs.

    this nonsense. A philosophical construct has not bearing on the empirically based results of rationalism, so conflating agnosticism with a lack of empirical knowledge lets you pretend to play with rationalists without getting your hands dirty with all the facts.

    Clever in a way, but only if one doesn’t think to much about it. All evidence in favor of a particular supernatural religious claim is quickly and soundly overwhelmed by Bayesian probabilities that accumulate the knowledge obtained by objective measurement.

    Trying to invert the process and declaring God a unitary probability space and arguing is must be chipped away at by empiricism is invalid in any context.

  11. cl Says:

    R.C.

    In your glee to get a gotcha, you echoed my claim – apparently without even realizing it.

    Trying to invert the process and declaring God a unitary probability space and arguing is must be chipped away at by empiricism is invalid in any context.

    Uh…?

  12. cl Says:

    DD,

    BTW, where is Lifeguard’s comment? I’d like to read it.

  13. R.C Moore Says:


    In your glee to get a gotcha, you echoed my claim – apparently without even realizing it.

    I re-read everything, and was about to come to the same conclusion, when I realized you don’t know the difference between hypotheses and conclusions.

    So I am not going to agree for now, because it is all too muddled. I may agree with full clarification. That means I am not going to grant you being correct on an irrelevancy so you can inflate into a full endorsement.

    We have been down that road before.

  14. Hunt Says:

    The LC either is, or reminds me of, the fact that it’s always easier to denigrate another person’s position than it is to buttress your own argument, especially if you happen to be on the more solid foundation. You see this in common practice when creationists attempt to poison the well of evolutionary theory and conclusions. It’s always easier to cast doubt on a strong argument than it is to continue a sound theory into near certainty. You might consider this a corollary to the law of diminishing returns. When the opposition stops building their case and devotes themselves wholeheartedly to obstructing their opponent, you know that they consider themselves in the losing camp, whether they actually are or not. It’s all a mind game, really. By observing your opponent’s behavior, you can tell how well they think you’re doing — and you can tell exactly what it is that they fear. By and large, Christians hate evolution, thus we may conclude that they consider the truth of it a deadly threat. It directly undermines their teleological conceptions of reality. When you know you’re on the weaker side, what is your strategy? Cast aspersions and doubt into the opposition camp. This is exactly what we see.

  15. Hunt Says:

    I think R.C. Moore is on the right track by bringing up Bayesian inferences, but I haven’t fiured how it applies the GH and MH. Intuitively I have a hunch that a lot of this will boil down to the subjective/objective controversy in the Bayesian approach. How does one assign prior probability to hypotheses? I note that both sides have used this controversy to their advantage. William Lane Craig has a favorite Bayesian technique when he states that the Resurrection should seem quite ordinary so long as we accept the hypothesis of God as given. Usually this stops his mathematically unversed opponent dead in his or her tracks. It’s an absurd argument, of course, because you can make any extraordinary event ordinary by assuming the requisite prior probabilities. I can, for instance, make the possibility that I’m going to have to do battle with Darth Vader tomorrow quite probable so long as I assume with probability 1.0 that I live in the world of Star Wars and that I’m a Jedi warrior.

  16. R.C Moore Says:


    William Lane Craig has a favorite Bayesian technique when he states that the Resurrection should seem quite ordinary so long as we accept the hypothesis of God as given

    Excellent example to what I was trying to say. You said it much better.

    I can only apply a Bayesian inference to the MH, as I am unable to formulate the GH as a query.


    How does one assign prior probability to hypotheses?

    In order to preempt undue criticism, we really should get more consistent in our language (I am the worst offender). The hypothesis is the question, the search target. We arrive, through Bayesian inference at a probability the hypothesis is true (or false, makes no difference, subtract 1).

    Because we are using Bayesian logic, the probability is not statistical (Gaussian)– it is not the result of flipping a coin. It is only more or less probable than the conclusion of other hypotheses, but more importantly, over time, the probability begins to converge or stabilize.

  17. R.C Moore Says:


    You might consider this a corollary to the law of diminishing returns. When the opposition stops building their case and devotes themselves wholeheartedly to obstructing their opponent, you know that they consider themselves in the losing camp, whether they actually are or not. It’s all a mind game, really. By observing your opponent’s behavior, you can tell how well they think you’re doing — and you can tell exactly what it is that they fear.

    You put this very well. Consider it stolen for future use (apologies)

  18. John Morales Says:

    I appreciate the responses to my criticism; I’d’ve responded sooner but it was a long day at work :(

    ThatOtherGuy, I grant your point, but I was addressing their equality in terms of justification, not their rank position on some scale of justifiability.

    R.C Moore, the issue of justification is critical to the issue, as you say.
    Both DD and cl were referring to hypotheses being equally justified, but the only criterion mentioned was consistency with observation; I am saying there are multiple criteria that are applicable, and that I think once additional criteria are included in the considerations, I doubt there could not be one preferred (least unjustified) hypothesis.

    pboyfloyd, as per my above, were we to apply the criterion of pragmatism to the hypotheses and one thereby become preferable (more justifiable) on that basis, there would no longer be equality so the consideration wouldn’t arise.

    R.C Moore, I agree Bayesian methods would be appropriate for such determinations, given that participants can agree on intersubjective quantifications for various propositions.
    (… and what are the odds of that? :)

    Regarding your logic exercise, since your axioms 1 and 2 are contradictory, it is no surprise they do not map reciprocally as one must map onto the null set. Furthermore, they relate to the attribute (justifiability) not to the entities (the hypotheses), and so are not an apposite analogue.

    Your axiomatic conjunction is of the form (where L represents lies always, J represents justification)
    ∀ILI ∧ ∀I~LI
    where to reflect my contention regarding hypotheses, it should be of the form
    ∀x,y(Jx=Jy)

  19. John Morales Says:

    [meta]
    Whoops, just reviewed the thread and noticed I unconsciously assumed readers would be familiar with my atheism; as a result, I expressed myself very poorly.

    When I wrote “I speak of both being agnostic yet believing”, the belief is that no gods exist.

  20. John Morales Says:

    cl,

    I define agnosticism as lack of knowledge

    The word for a lack of knowledge is ignorance. Your argument from agnosticism thus becomes an argument from ignorance.

  21. John Morales Says:

    correction:
    (where L represents lies, J represents justification)
    ∀ILI ∧ ∀I~LI

  22. cl Says:

    [ This comment has been moved to the discussion forums.]

  23. R.C Moore Says:

    John Morales —


    I agree Bayesian methods would be appropriate for such determinations, given that participants can agree on intersubjective quantifications for various propositions.
    >/i>

    Ahh, but the point of Bayesian inference is that no one has to, they just vote. Think commodity markets or your spam filter.


    Regarding your logic exercise, since your axioms 1 and 2 are contradictory, it is no surprise they do not map reciprocally as one must map onto the null set. Furthermore, they relate to the attribute (justifiability) not to the entities (the hypotheses), and so are not an apposite analogue.

    Your axiomatic conjunction is of the form (where L represents lies always, J represents justification)
    ?ILI ? ?I~LI
    where to reflect my contention regarding hypotheses, it should be of the form
    ?x,y(Jx=Jy)

    Yes, just trying to be pointlessly clever, and to reinforce why I prefer a Bayesian approach to these problems.

    You have upped the standards, using formal notation. I will try to meet them.

  24. Hunt Says:

    cl,

    I realize you haven’t explicitly aimed this at me, but I feel the need to combat the negative stigma DD assigned to me all out of his misunderstanding my argument.

    You’re right, it was meant as a general statement and can apply to anyone arguing any point. I haven’t been around long enough to completely get your arguments, but I’m happy there are opposing viewpoints. Discussion is pretty dull otherwise.

  25. John Morales Says:

    [meta]
    NB – cl is not being in any way censored, his entire comment is now in the forum and I personally have responded there.

    (The comment was also responding to RC and Hunt, as well as to me.)

  26. Lifeguard Says:

    DD,

    Thanks for answering my question. I suspected that was the direction you were headed in, but it didn’t seem that clear to me from the original LC post.

    In any event, great blog and great threads… even if a little high above my head.

  27. cl Says:

    [Moved to the forums.]