“When you have eliminated the impossible, whatever remains, however improbable, must be the truth” - Sherlock Holmes.
Wrong! - You can never know all other options, and you can never know they are impossible.
In 1999, Sally Clark was convicted of murdering her two sons and was jailed for life. She served three years before being released on appeal in 2003. The expert testimony of Pediatrician Professor Sir Roy Meadow was instrumental in this conviction, as it was in the conviction of Angela Cannings, Donna Anthony and Trupti Patel for murdering their two, two and three children respectively. These three women were also eventually freed on appeal.
The children died due to Sudden Infant Death Syndrome (SIDS), known as cot death in the UK. But Meadow testified for the prosecution that the odds of a child dying from SIDS were 1 in 8,543, and so the odds of two children from the same family dying from SIDS were 8,843 squared – or 1 in 72 million. The jury bought the idea that these odds were so astronomical the mothers must have murdered their children.
Anyone with a basic knowledge of logical fallacies would recognize this as a false dilemma – a limited number of options (usually two) are given, while in reality there are more options. Let’s suppose the odds of two children dying from SIDS by pure chance really are 1 in 72 million. Meadow and the jury made the mistake of assuming there were only two choices – chance or murder. They ruled out chance so it must be murder, right? Wrong, because there are at least two other options, for example: a genetic pre-disposition; or something in their home. There are probably others we don’t know about because you can never know all other options. It wasn’t chance that these children died, but that didn’t mean it was murder.
Critical thinkers would also note that no positive evidence was offered for murder, just negative evidence against SIDS. If you want to prove something true you must actually prove it true, not just claim that all alternatives are impossible, because you can never know all other options are impossible. This reversal of the burden of proof is a fallacy known as argument from ignorance - unless something (murder, in this case) has not been proven false, it must be true.
It’s mind-blowing that a jury didn’t see these logical errors, unbelievable that the defense attorney didn’t make the case for them and disgraceful that a prosecution could even be brought with such flimsy evidence. It gets worse - Meadow testified in the Patel trial after the law lords had already called his same evidence in the case of Cannings and Clark, “grossly misleading”. How did this happen? Meadow, with his long list of qualifications, was considered an authority on the subject, and prosecutors and jurors were in awe of him. And this neatly illustrates a third fallacy, argument from authority - even experts can be wrong; what counts are the data not the qualifications of the person presenting them.
Meadow is currently facing a General Medical Council hearing into charges of serious professional misconduct. Last week it heard he gave "misleading and flawed" evidence in Clark’s trial. He could be struck off if found guilty.
A key tool of critical thinking is to know and understand logical fallacies: don’t rely on them yourself and recognize them in opposing arguments. It seems to me the application of critical thinking and recognition of fallacious reasoning in these cases could have prevented these innocent women from going to jail.
As for Sherlock Holmes, despite his reputation he really wasn’t much of a critical thinker. But that shouldn’t be too surprising: his creator believed in fairies at the bottom of the garden.
Thanks for taking on that silly Holmes quote! Nonsense dressed up as profundity. When you have eliminated the impossible, you are left with all the possibilities to consider.
Posted by: VKW | June 28, 2005 at 12:33 PM
If you want to prove something true you must actually prove it true, not just claim that all alternatives are impossible, because you can never know all other options are impossible.
This argument is not critical thinking, it's patent foolishness.
It doesn't apply in mathematics, nor in logic, and it doesn't apply in the courts either.
Where a prosecution lacks sufficient direct evidence, they can bolster their case by demonstrating that an accused is very unlikely to be innocent. Agreed, this is not a popular approach. It makes for a weak case as it's the long way around and because the defense has merely to establish that there is a reasonable doubt, by emphasizing other possibilities, as in the case you cite above.
But it remains a valid approach, because the standard winning the case is intentionally discretionary. Conviction does not require absolute proof.
Posted by: Blue State | June 28, 2005 at 07:02 PM
The fact that these women were actually innocent, proves you are wrong.
Posted by: Skeptico | June 28, 2005 at 08:04 PM
Logical flaws or not, the statistics are bogus anyway. It may be true that the odds of two SIDS deaths in a given family are 1 in 72 million, but that does not translate to "astronomical" odds for it happening.
Imagine that you flip a coin 8 times. The chances of you getting all heads is 1 in 256, which seems rather low. But if you have 100 people flipping coins, the odds that at least one of them gets all heads is something like 64% (hope I got the numbers right; but you get the idea). Those odds don't seem low at all.
In other words, you have millions of families trying this "experiment" when they have children. The chance that it will happen to someone is probably not that unbelievable. The 1 in 72 million figure would only apply if you chose, a priori, which family to watch.
Also, Skeptico is right to point out that a calculation based purely on odds is assuming a flat distribution. Since almost nothing is known about SIDS/cot death, the possibility of genetic of environmental factors is too significant to ignore.
Having said that, I do think there is a role for this kind of argument in some cases. I'm having trouble remembering the details, but I recall reading about a court case where two expert witnesses (I believe it was Henry (Sr) and William James) were called upon to verify the authenticity of a signature. The prosecution theorized that the signature was a forgery, added to a new document after the death of the signatory by the defendant. The James' found a signature in the correspondence of the signatory that matched the one on the new document pretty much exactly (right down to its position on the page!). They tried to quantify the odds of such a close match occuring by examining the variance in other signatures, and came up with some ridiculously small probability. Though there analysis was pretty good, the unsophisticated jury didn't buy it.
Posted by: Eric Wallace | June 28, 2005 at 10:17 PM
I may be incorrect in stating this, but say for argument's sake the chances of one child passing from SIDS is truly 1 in 8,543. Then it occurs. Does it lessen the chances of the second child passing from the same syndrome? In the same vein, if you flip a coin and get heads, does that lessen your chances of getting heads on the second flip? Please point out any flaw in my logic and forgive me, it's been awhile since statistics. But would not the chances of 2 children passing from SIDS be 2 in 17,806 and not 1 in 72 million?
Posted by: Rockstar | June 29, 2005 at 06:12 AM
Eric:
You are correct on the bogus statistics argument – if there were 72 million families, the odds are one would have 2 deaths by chance alone.
Rockstar:
If one child had already died, the chance of having a second die would still be 1 in 8,543. But the chance of a family having two deaths by chance would be 8,543 squared.
Posted by: Skeptico | June 29, 2005 at 09:23 AM
The fact that these women were actually innocent, proves you are wrong.
Skeptico - your comment is another example of uncritical thinking. The fact that the women were found innocent is irrelevant and does nothing to prove an assertion that discrediting alternatives is an invalid method for building a case.
Posted by: Blue State | June 29, 2005 at 11:01 PM
Were I one of these innocent women, I certainly would not want to rely on flawed methodology. For instance:
... rarity by itself shouldn't necessarily be evidence of anything. When one is dealt a bridge hand of thirteen cards, the probability of being dealt that particular hand is less than one in 600 billion. Still, it would be absurd for someone to be dealt a hand, examine it carefully, calculate that the probability of getting it is less than one in 600 billion, and then conclude that he must not have been dealt that very hand because it is so very improbable. --John Allen Paulos
Posted by: Rockstar | June 30, 2005 at 06:02 AM
Blue State:
Re:The fact that these women were actually innocent, proves you are wrong.
Skeptico - your comment is another example of uncritical thinking. The fact that the women were found innocent is irrelevant and does nothing to prove an assertion that discrediting alternatives is an invalid method for building a case.
Actually, yes it does. Alternatives were discredited and yet the women were innocent, so the method is clearly flawed. I know that this type of evidence may often be used in courts of law, but I am not talking about courts of law – I am talking about how science is done. (And by extension, critical thinking. Science and critical thinking have a higher standard than a court of law – they are different beasts altogether.)
Please read the following and try to understand this fallacy,
You are relying on a logical fallacy if your conclusions are not arrived at from the premises you present. If you just discredit alternatives, and provide no evidence for your theory, you are relying on the Argument From Ignorance fallacy:
This is the exact same arguments creationists use. They say the eye, or the flagellum or whatever couldn’t have evolved by natural selection, and so a designer must have been involved. No evidence of a designer is shown – they think it is enough merely to show it is “irreducibly complex” so it couldn’t have arisen through evolution. Flawed reasoning that Richard Dawkins called “argument from personal incredulity” – a subset of argument from ignorance.
Another example: crop circle enthusiasts say some crop circles must be made by aliens, or some other unknown force, because humans couldn’t possibly have made all the really complex circles. Flawed reasoning again – they have to demonstrate what this other thing is that is making the crop circles: it is not good enough just to say that it’s too difficult for a human to have done that so it must be_______ (and there you fill in your preferred alternative). They need positive evidence for what they think is causing it.
Here is another definition from wikipedia
Please try to understand the error here: it’s one of the most common fundamental flaws in reasoning people employ.
Posted by: Skeptico | June 30, 2005 at 10:08 AM
Skeptico, you don't seem to understand that the argument from ignorance fallacy is not the same as disproving all the alternatives. In maths, logic, science, and critical thinking, disproof of all the alternatives is the equivalent of proof of the assertion. It mathematically constitutes an absolute proof.
The trick is in covering all the alternatives. It's a lot harder to prove the assertion that all crows are black by confirming that all non-black things are non-crows. But it's logically correct to do so.
The women were initially found guilty on the assertion that there is a 1 in 73 million chance of two SIDS deaths in the same family. But that's a grossly misleading statistic, and the problem is that the argument based on it falsely excludes other possibilities, not that it excludes them.
That's a key distinction.
The misleading evidence should have been challenged in court the first time. According to the AAP, deaths erroneously attributed to SIDS that are actually infanticide or abuse are estimated at 1%-5% (although the number of SIDS diagnoses has declined in concert with a rise in the number of abuse cases in recent years). But from that statistic it could be argued that the chance of two infant deaths occuring in the same family and not being murder is 90% (95%x95%). That may be a little high, as the statistic comes from independent cases, not cases from the same families. But it's quite a long way from 1 in 73 million figure, and certainly should be enough to induce more than reasonable doubt about the prosecution case.
Also, it's false to assert the women in the case were innocent, in a logical sense. That's unknown. They did not prove their innocence on appeal. They showed that the case against them was flawed, so now they are presumed innocent, by the standard legal convention.
Posted by: Blue State | June 30, 2005 at 11:46 PM
I’m sorry, you are just wrong. I think maybe you are hung-up of a legal definition – what is allowed in a court of law. But I am not talking about what is allowed in a court of law; I am talking about logic and fallacious reasoning. Disproof of all the alternatives is not the equivalent of proof of the assertion. Maybe it is in math, maybe it is allowed in a court of law, but not elsewhere, because outside of math, you can not know all the alternatives.
From my link, argument from ignorance is:
…a special case of a false dilemma, since it assumes that all propositions must either be known to be true or known to be false.
Think about that. It obviously first has to assume all propositions must be known, yes? As I said, outside of math, you can’t possibly know all alternatives, which is why it’s flawed logic.
Argument from ignorance says, “I don’t know what happened therefore it must be ________" (and here you enter your own preferred pre-conceived explanation). But it must be obvious this is wrong. If you don’t know what happened then you just don’t know what happened – period. If you have “disproved all the alternatives”, all you have done is “disproved all the alternatives you can think of.” It’s only evidence of your own lack of imagination, nothing else. Look at the arguments creationists use – that’s exactly the form of their argument: we have disproved evolution (they think), so there must be a designer. They have no evidence for the designer, just evidence (they think) against all the alternatives.
Re: According to the AAP, deaths erroneously attributed to SIDS that are actually infanticide or abuse are estimated at 1%-5% (snip). But from that statistic it could be argued that the chance of two infant deaths occuring in the same family and not being murder is 90% (95%x95%).
That’s absurd. So if there were ten deaths in one family the chance that they were not murder would be 60%. I think you are ignoring the probability of a child dying in the family in the first place.
Posted by: Skeptico | July 01, 2005 at 02:53 PM
You seem to think I'm taking issue with the logical fallacy of argument from ignorance. I'm not. I'm sure that fallacy is used rhetorically every day by people who argue for what they want to believe, like creationism, or whatever.
What I'm saying us that it's not what we have here. You asserted that you can't build an argument by disproving all the alternatives, and I took issue with that. It works in mathematical logic, and, as it relates to this example, it will stand in court if the unknowable alternatives you failed to imagine are unlikely to be true.
For example, we don't know it wasn't space aliens that killed the children either, but basing a legal defense on that extremely remote yet open possibility is, I'm sure you'll agree, a shortcut to the electric chair.
Although we'd like to believe otherwise, this case was not overturned because a junior skeptic with his list of logical fallacies wrote a letter to the judge. It was overturned because the evidence excluding other possibilities was "grossly misleading". That's a different thing entirely.
Skeptico, your tag line mentions critical thinking, so I assume you take that idea seriously. Producing a post a day it may be hard to keep the assertions tight and logical and free from snipes like myself, but with practice you'll get better at it. What you needed to do here was say "This case reminded me of a logical fallacy I see commonly in creationist arguments..."
Posted by: Blue State | July 01, 2005 at 08:53 PM
And BTW, on the AAP statistics - the point is not that the statistic is the right number, it is that a probability of 90% is a long way from 1/73 million, and should cast doubt on the latter.
(That said, given the assumptions, the calculation is correct.)
Posted by: Blue State | July 01, 2005 at 09:19 PM
Blue State:
You’re reading a critical thinking blog, so I assume you take that idea seriously. Producing a comment a day, it may be hard to keep your assertions tight and logical and free from being refuted, but with practice you'll get better at it. Although you'd like to believe otherwise, my argument was not overturned because a junior skeptic wannabe with a flawed understanding of logical fallacies wrote a comment in my blog. What you needed to do here was say, "This is a logical fallacy but it might work in a court of law and mathematics”.
I’ll try one more time to explain for the slow student. For some reason, despite repeating this about 4 times now, you don’t seem to get that A FLAW IN LOGIC IS NOTHING TO DO WITH WHAT WOULD BE ALLOWED IN A COURT OF LAW. It will work in math. It will perhaps be allowed in a court of law. But in the world of science and critical thinking it fallacious reasoning. What may be allowed in a court of law and what is critical thinking are two different things.
As for the statistics – your figures show that if 13 children died in a row, the odds of murder would be only50%. That is absurd. There is a glaring error in your figures.
Now, to see if you have followed this time, some questions.
1) Argument from ignorance is “…a special case of a false dilemma, since it assumes that all propositions must either be known to be true or known to be false”. Please explain how this could be so without knowing all the options. Please explain how you can disprove all the alternatives unless you know what they all are.
2) Alternatively, please explain how you can know that you know all the alternatives. (Note: you will fail this question if your answer is to do with mathematical proofs.)
3) Please explain how a creationist’s argument:
…is different in principle from:
4) If thirteen infants died in a row in the same family, would you say the probability of the deaths being murder was only 50/50 (95% ^ 13)
I think if you think carefully about these questions and try to answer them honestly, you will learn something about critical thinking. Please don’t be afraid to ask if you don't understand something. But please try to get in right this time – if you keep getting poor grades in this subject I may have to hold you back from posting in the more advanced threads in the future.
Posted by: Skeptico | July 02, 2005 at 10:29 AM
Skeptico, you need to slow down and read what I've written.
1) & 2) Once again, I'm not taking issue with the existence of the fallacy of argument from ignorance. I said above, I'm sure it's used every day.
I'm saying that you were wrong when you wrote that you can't prove an assertion by disproving the alternatives. You've already yielded that, grudgingly, in your responses, in the fields of mathematics and law.
I'm going to explain how disproving the alternatives is also used in another field, a very topical one.
As I wrote originally, the problem is covering all the alternatives; it's rarely used to prove assertions, as it makes for a weak case legally, and more direct approaches to proving an assertion are usually more fruitful.
It is precisely this impracticality of covering alternatives, by the way, that is known as the "skeptics burden". Skeptics hold that they can never know, never really know with 100% certainty, to the millionth decimal place, that some outrageous assertion is untrue. Skeptics are condemned forever to politely accept the possibility that the fruitcake's assertion is true, and politely ask the fruitcake to show evidence or proof for the claim. But a condemnation it is, because a skeptic is obliged to refrain from just calling the fruitcake a fruitcake.
For example, you cannot prove that I'm not in fact a psychic spoon-bending space-alien; you are obliged to admit that the possibility, however remote, exists, and to ask me for proof. You cannot prove that titanium necklaces don't work; you must politely ask for proof that they do from the maker. You cannot prove that acupuncture does not work, nor astrology, nor mind-reading, nor ESP, nor nothing.
This is why Randi had the idea to put $1 million on the line. He can't prove these things don't work either. So he made a polite put-up or shut-up to those who claim to have supernatural gifts. He does not just say that the claimants are liars or fools or both. He cannot, he is a skeptic.
Carl Sagan's "Demon Haunted World" is a classic skeptical work. You can read the entire thing and he won't just flat-out assert that demons don't exist. If he made the claim, he would then need to prove it, and would find it difficult to nail down all the possibilities, and if he forgot one, some fruitcake would claim it as proof that demons exist. So if he can't prove it, how can he make his case?
He asks questions, over and over again of those who would believe. He makes a very effective case. He makes it indirectly, not proving directly that demons don't exist, but calling into question evidence that they do, relentlessly, and opening the possibilities that they don't. An interesting way to proceed, don't you think?
I'm saying that casting doubt on possibilities and disproving alternatives is a major tool of skepticism itself.
3) The argument as you state it is the same principle. But the argument made by the prosecution was different It was not that the deaths could not have arisen by chance, but that two SIDS deaths could not have arisen by chance.
This is different because it excludes a bunch of possibilities by the nature of SIDS. SIDS is by definition an unknown cause. No one really knows why it happens. If a child dies from an identifiable cause (accidental suffocation, choking, drowning, falls, etc), it's not a SIDS death.
But even so, for prosecution purposes, the argument as made is also incomplete. The prosecution can't rest on the assertion that it couldn't have arisen by chance to it must be murder. They also have to prove who did it (ie. not the babysitter, not the husband, not someone else), how it was done and that the medical record doesn't contradict this, that there was opportunity and motive - proceeding thereby to exclude other possibilities beyond a reasonable doubt, as I mentioned.
The apparent flaw in the case was that the children's deaths both being SIDS was much more probable than represented. That was the key evidence that was grossly misleading.
That is not the same thing as the case suffering from the fallacy of argument from ignorance. It's very different.
4) The statistic is drawn from the article reporting AAP findings that you cited. In that report it said that 1%-5% of infant deaths diagnosed as SIDS were in fact infanticide. So using the higher incidence, the chance of one SIDS death actually being infanticide is 95%, and the chance of two is 90.25%. That's simple maths. As I said, there are assumptions in that figure that make it too high; the stat is from independent cases, not two in one family. It also depends on a constant accuracy in diagnosing SIDS instead of infanticide.
But, I'll say this for the third time, the purpose is that the probability figure of 90% is a strong contrast from 1/73 million, and it should cast doubt on that figure being accurate, and should thereby cast doubt on the prosecution's case.
Posted by: Blue State | July 03, 2005 at 08:24 AM
On point 4, I inverted the causes. It should read: "So using the higher incidence, the chance of one SIDS death actually being SIDS is 95%, and the chance of two is 90.25%."
Posted by: Blue State | July 03, 2005 at 08:29 AM
Blue State:
You need to slow down and answer the questions. They were:
1) Argument from ignorance is “…a special case of a false dilemma, since it assumes that all propositions must either be known to be true or known to be false”. Please explain how this could be so without knowing all the options. Please explain how you can disprove all the alternatives unless you know what they all are.
2) Alternatively, please explain how you can know that you know all the alternatives. (Note: you will fail this question if your answer is to do with mathematical proofs.)
3 ) Please explain how a creationist’s argument:
…is different in principle from:
4) If thirteen infants died in a row in the same family, would you say the probability of the deaths being murder was only 50/50 (95% ^ 13)
Your answers in bold:
I'm saying that you were wrong when you wrote that you can't prove an assertion by disproving the alternatives. You've already yielded that, grudgingly, in your responses, in the fields of mathematics and law.
I did not concede this was so in a court of law. I said it is often allowed in a court of law, but it is still a fallacy.
And I did warn you not to mention math proofs. Starting off with negative points here.
As I wrote originally, the problem is covering all the alternatives; it's rarely used to prove assertions, as it makes for a weak case legally, and more direct approaches to proving an assertion are usually more fruitful.
It’s not that it “makes for a weak case”. It is that makes for a logically fallacious case. So far you are agreeing with me.
It is precisely this impracticality of covering alternatives, by the way, that is known as the "skeptics burden". Skeptics hold that they can never know, never really know with 100% certainty, to the millionth decimal place, that some outrageous assertion is untrue. Skeptics are condemned forever to politely accept the possibility that the fruitcake's assertion is true, and politely ask the fruitcake to show evidence or proof for the claim. But a condemnation it is, because a skeptic is obliged to refrain from just calling the fruitcake a fruitcake.
Sorry, but you failed to even address the questions.
You appear to be confused. You are talking about the impossibility of proving a universal negative. That is not the same as saying that disproving all possible options proves the one option left.
You cannot prove that titanium necklaces don't work; you must politely ask for proof that they do from the maker.
The burden of proof is with the maker of these things. I don’t have to prove they don’t work; he has to demonstrate they do. If he can’t, I am entitled to assume they don’t work. I haven’t done that by “disproving all the alternatives”.
I’m sorry but you have demonstrated such a misunderstanding of this that I must fail you on these questions. Actually an F-, since you still don’t understand that math proofs are something different.
3) The argument as you state it is the same principle. But the argument made by the prosecution was different It was not that the deaths could not have arisen by chance, but that two SIDS deaths could not have arisen by chance.
But the only difference between one and two deaths is in the magnitude of the probabilities. The principle is exactly the same.
This is different because it excludes a bunch of possibilities by the nature of SIDS. SIDS is by definition an unknown cause. No one really knows why it happens. If a child dies from an identifiable cause (accidental suffocation, choking, drowning, falls, etc), it's not a SIDS death.
Wrong – that is not a difference in principle. If natural selection is not possible then by definition there is an unknown cause. Exactly the same.
But even so, for prosecution purposes, the argument as made is also incomplete. The prosecution can't rest on the assertion that it couldn't have arisen by chance to it must be murder. They also have to prove who did it (ie. not the babysitter, not the husband, not someone else), how it was done and that the medical record doesn't contradict this, that there was opportunity and motive - proceeding thereby to exclude other possibilities beyond a reasonable doubt, as I mentioned.
Exactly the case with the creationists. If natural selection is ruled out, that does not mean that god did it by default: they also have to show evidence that god actually did do it. They don’t do this – they just assume it is the only alternative. This is argument from ignorance, just as it would be in the SIDS example. This is exactly the same. You have just agreed with me.
I’m afraid I have to fail you on this question too since you apparently don’t see the significance in what you just wrote – they are exactly the same.
The apparent flaw in the case was that the children's deaths both being SIDS was much more probable than represented. That was the key evidence that was grossly misleading.
That is not the same thing as the case suffering from the fallacy of argument from ignorance. It's very different.
That was an additional error in the SIDS court case. I am aware of that, I didn’t comment on it because I was demonstrating the logical fallacies only. It doesn’t mean there wasn’t also a logical fallacy.
4) The statistic is drawn from the article reporting AAP findings that you cited. In that report it said that 1%-5% of infant deaths diagnosed as SIDS were in fact infanticide. So using the higher incidence, the chance of one SIDS death actually being infanticide is 95%, and the chance of two is 90.25%. That's simple maths. As I said, there are assumptions in that figure that make it too high; the stat is from independent cases, not two in one family. It also depends on a constant accuracy in diagnosing SIDS instead of infanticide.
But, I'll say this for the third time, the purpose is that the probability figure of 90% is a strong contrast from 1/73 million, and it should cast doubt on that figure being accurate, and should thereby cast doubt on the prosecution's case.
I am aware of that. But you didn’t answer my question so I am going to have to fail you on this too.
Four Fs. Must do better.
To recap, the error was in not proving murder, but instead just saying we have no knowledge of how the children died. They were drawing conclusions from a lack of knowledge, but it should be obvious that you can not draw any conclusions from a lack of knowledge. That literally is argument from ignorance.
Posted by: Skeptico | July 03, 2005 at 02:22 PM
Hey Skeptico,
It's long been a principle of science that theories are subject to review and improvement, or even being changed by new discoveries. We saw how Newtonian physics replaced the crude understandings before that, and how it was in turn improved by the correction factors introduced by relativity. A scientific theory is never held to be the absolute truth; it's simply a postulate that happens to predictably explain observable phenomena, and it stands or falls on it's ability to do so.
Creationists seized upon this, rhetorically, to say evolution is a "theory", not a fact. And they have another theory - so why not some equal time for their theory?
But they don't want to prove their theory. Instead, they want to attack evolution. And they do it by arguing about fringe cases, for which the evidence may yet be to accumulate. If you can't prove this possibility, if there is a hole somewhere in the fossil record or other evidence, then you can't prove evolution, they argue. They fulminate that it's logically fallacious to exclude these possibilities - and that scientists lack the imagination to think of others. Therefore, they argue, arguments for evolution are based on a logical fallacy.
It's a specious argument of course. The evidence for evolution continues to grow, and creationism hangs on to increasingly remote fringes. But it will never completely go away, because we can't prove it absolutely, we can't close down all the possibilities. We can't know all the possibilities about any scientific theory.
We can simply continue to build the case until those that hold it not to be true are holding on to fruitcake positions. That's how all science works. Theories on gravity are only "theories" as well, but that doesn't mean we are going to start floating around just because we can't exclude possibilities we haven't imagined yet.
So I'm saying that foreclosing on alternative possibilities is not only applicable in maths, law, and skepticism, it's also the basis of how scientific knowledge is accumulated.
Do I get another F?
Posted by: Blue State | July 03, 2005 at 05:20 PM
It's revealing, come to think of it, of the weakness of any argument that argues fallaciousness on the basis of not excluding possibilities. That's where creationists have been pushed. Creationists used to argue that God created man, and all the types of animals and plants, like it says in the Bible. Then they began to back away from Biblical literalism and instead argue that there must have been a "creator". Now they have backed away from that, and ask only that scientists admit the possibility that "intelligent design" must have had a role in the creation of species.
All they want is for evolutionists to grant that tiny logical possibility.
Posted by: Blue State | July 03, 2005 at 05:32 PM
Yes you do. F minus actually, since you have introduced another fallacy, the false analogy. There is strong evidence for evolution. Overwhelming evidence, actually. There was no evidence for murder in the cases of these women, just negative evidence against chance. Do you really not see the difference?
Scientific theories are based on strong evidence for the theory, not evidence against the alternatives. Saying there is evidence against all alternative possibilities is absolutely with 100% certainly NOT the basis of how scientific knowledge is accumulated. This is 100% fallacious reasoning.
Posted by: Skeptico | July 03, 2005 at 05:40 PM
Sure I see the difference. But the case for murder is not based on negative evidence against chance. It was based on negative evidence against many possible other causes (i.e. all non-SIDS causes), plus additional evidence that it was in fact the mother, etc, as I explained above. See the difference?
The correct contrast is between the assertion:
"the flagellum could not have arisen through natural selection, therefore god did it"
and
the children’s deaths could not have arisen through any of the thousands of known diagnosable causes, and it was 1/73 million against being two cases of SIDS (that's the false bit), therefore there is a case that it was murder (which we will go on to provide evidence for).
That argument will not rule out absolutely all imaginable possibilities. But it intends to push all the alternatives into the realm of improbability, down their with the other non-excluded "demons did it" possibilities. It's a roundabout method, but it's not based on logical fallacy. Do you see how that case stands, despite not excluding possibilities?
It's the same with the theory of evolution. Creationists troll arguments like "how could something really complex like (example) evolve?". They advance the argument that the possibility that intelligent design cannot be excluded. They'd like to proceed from there into the argument from ignorance to say that therefore it was God. But first they need scientists to yield the logical possibility. Should they yield that intelligent design in specific fringe areas is a non-excluded possibility?
And as I said, scientific theories stand or fall on their ability to predict observable phenomena. Their inability to exclude fringe possibilities weakens them somewhat, depending on how far out the fringes are. And in that effort, any ability to disprove alternatives strengthens them. That's certainly been the case with the gradual strengthening and acceptance of the theory of evolution over the past 130-odd years. I'm sure there are other examples of where we have strengthened scientific knowledge through evidence against the alternatives.
Evidence against the alternatives is weaker than evidence for the assertion, because alternatives are theoretically infinite. But it still leads in the direction of the truth. What's so hard to understand about that?
Posted by: Blue State | July 03, 2005 at 06:45 PM
Blue State:
Are you having me on or are you really this dense? You are having me on, right? I hope so for your sake.
Sure I see the difference. But the case for murder is not based on negative evidence against chance. It was based on negative evidence against many possible other causes (i.e. all non-SIDS causes), plus additional evidence that it was in fact the mother, etc, as I explained above. See the difference?
WRONG!
The case for murder WAS based on negative evidence against chance. Chance is 1 in 73 million – highly unlikely therefore murder. JESUS CHRIST what is wrong with you?
The correct contrast is between the assertion:
"the flagellum could not have arisen through natural selection, therefore god did it"
and
the children’s deaths could not have arisen through any of the thousands of known diagnosable causes, and it was 1/73 million against being two cases of SIDS (that's the false bit), therefore there is a case that it was murder (which we will go on to provide evidence for).
That is just sophistry. They are the same.
It's the same with the theory of evolution. Creationists troll arguments like "how could something really complex like (example) evolve?". They advance the argument that the possibility that intelligent design cannot be excluded. They'd like to proceed from there into the argument from ignorance to say that therefore it was God. But first they need scientists to yield the logical possibility.
No they don’t. They have to provide evidence FOR god did it. There is no need for “scientists to yield the logical possibility”. Of course there is that possibility. That isn’t the point. The point is
THERE IS NO EVIDENCE FOR GOD DID IT, JUST NEGATIVE EVIDENCE AGAINST EVOLUTION (they think). And that is a fallacy.
Seriously, does anyone else see what I am getting at here or is it just this one idiot?
Posted by: Skeptico | July 03, 2005 at 07:31 PM
I don’t know what else to say to you. You seem willfully to misunderstand this subject and there is really nothing much more I can do if you are determined not to understand what I am telling you. I don’t know why you have this mental block on this subject but I am now out of patience – I have no more time to continue with this stupid argument of yours. Saying you have disproved all the alternatives is argument from ignorance. All you are saying is you have dismissed all the alternatives you can think of. It’s pretty clear. You are just wrong. Give it up.
Posted by: Skeptico | July 03, 2005 at 07:41 PM
Hey, you don't dislike all junior skeptics do ya? :) I'm probably more of a sophomore...
Posted by: Rockstar | July 05, 2005 at 06:13 AM
(I am new here, but definitely not new to critical thinking or logic. I disagree with creationists and atheists and maintain an agnostic stance about most things)
I am with BlueState on this argument. All (s)he is saying is that "if all alternatives can be identified"... (practically impossible in real life, but IF) Eg: If the dice didn't show 1,2,3,4 or 6, then it has to be 5
Skeptico says: "Scientific theories are based on strong evidence for the theory, not evidence against the alternatives."
This is incorrect. Scientific theories are "chosen"... The one theory that is consistent with all experiments and requires least number of assumptions. There ARE alternatives to the current scientific theory of gravity, which cannot be disproved. Scientists CHOOSE to exclude them using Occam's Razor There is no Proof for Occam's Razor being right, but without it science will not exist.
You also stated that Since you cannot prove that ghosts do not exist, they must exist. This brings up the corollary which you seem to be sticking to, "Since you cannot prove that ghosts exist, they must not exist." I hope you realize that this is also a logical fallacy.
Posted by: dont_know | July 06, 2005 at 08:29 AM
don't know:
I am with BlueState on this argument. All (s)he is saying is that "if all alternatives can be identified"... (practically impossible in real life, but IF) Eg: If the dice didn't show 1,2,3,4 or 6, then it has to be 5
That is a math proof. As I wrote to Blue State until I was sick of repeating myself, math proofs are an exception to this logical fallacy because you can often know all the alternatives. Please don’t go on about math proofs – we have covered this already.
Skeptico says: "Scientific theories are based on strong evidence for the theory, not evidence against the alternatives."
This is incorrect. Scientific theories are "chosen"... The one theory that is consistent with all experiments and requires least number of assumptions. There ARE alternatives to the current scientific theory of gravity, which cannot be disproved. Scientists CHOOSE to exclude them using Occam's Razor There is no Proof for Occam's Razor being right, but without it science will not exist.
I never said scientific theories are proven. I said there is evidence for them. There is evidence for the theory of gravity. There is evidence for the theory of evolution. There is evidence for the theory of quantum mechanics. Etc etc. These theories are not based on “we have dis-proven all the alternatives so _____________ must be right”.
Regarding Occam’s Razor - you are confusing two things. I am saying “disproving all the alternatives so _____ is right” is a fallacy. Occam’s razor is not about disproving alternatives – actually O.R. specifically recognizes you cannot disprove a particular alternative but lets you rule it out anyway. O.R. just says don’t add unnecessary entities to the explanation. In other words, suppose we have a theory backed by evidence – you don’t need to add: "and a fairy sits on top” to the theory, although you could and the theory would work just as well. O.R. just says don’t add this unnecessary entity (it’s unfalsifiable).
You also stated that Since you cannot prove that ghosts do not exist, they must exist. This brings up the corollary which you seem to be sticking to, "Since you cannot prove that ghosts exist, they must not exist." I hope you realize that this is also a logical fallacy.
They are both logical fallacies – the same one. The same one you are saying is not a fallacy.
You cannot prove something by disproving all the alternatives – all you are doing is disproving the alternatives you can think of. This is the creationist’s argument:
“the flagellum could not have arisen through natural selection, therefore god did it”
Even if the flagellum could not have arisen through natural selection, that doesn’t mean god did it, it would just mean “we don’t know what happened”. You can not know all the alternatives - that is a fallacy.
Posted by: Skeptico | July 06, 2005 at 10:11 AM
I guess we all agree on everything but for some very subtle semantics.
I am saying “disproving all the alternatives so _____ is right” is a fallacy
True for all practical (real life) applications because the phrase "all the alternatives" can never be used accurately in real life.
Posted by: dont_know | July 07, 2005 at 07:26 AM
Couple of points, if I may.
Technically, the women who won their appeals against conviction have not been found "innocent" -- the House of Lords decision was that their convictions were unsafe. In the English criminal justice system, the verdicts are "guilty" and "not guilty", and the prosecution has the burden of proving its case beyond a reasonable doubt. Thus a "not guilty" verdict does not equate to innocence, but rather that the prosecution's case was not strong enough to convict, and "guilty" does not mean that the jury are 100% sure that the defendant comitted the crime. Just to add a little more confusion, the Scottish system has a third verdict, "not proven".
Second point, Meadow has been struck off.
Posted by: Nigel Pond | July 19, 2005 at 10:00 AM
I think I side with Blue State. I don't see why it must be impossible to eliminate all possibilities but one. The obvious example here is that in a court case, a defendant can be found either guilty or not guilty. And dont_know mentioned the dice example ("If the dice didn't show 1,2,3,4 or 6, then it has to be 5").
To say that situations where there are a finite number of possibilities are "mathematical" and are "an exception to this logical fallacy because you can often know all the alternatives," or that "outside of math, you can not know all the alternatives" seems to me to be a cop out. There are many situations where we can see some unknown as one of a finite number of possible options. Must all dilemmas be false dilemmas?
It seems that, in your disagreement with Blue State, you (and others) are saying there is always an infinity of (unknown or unknowable) possibilities when dealing with real-world situations. For example, in deciding whether Skeptico or Blue State is right on this issue, there are a finite number of possibilities:
1. Skeptico is right, Blue State is wrong
2. Blue State is right, Skeptico is wrong
3. Both Skeptico and Blue State are wrong
4. Both Skeptico and Blue State are right
Clearly, it is not the case that there is an infinity of options here. Your premise (that you can not know all the alternatives) is not true. Or would you say that this is also "mathematical"?
Posted by: | August 09, 2005 at 01:08 AM
That last post was from me, mefisto. It was my first post here, and it seems I wasn't logged in when I wrote it.
Posted by: mefisto | August 09, 2005 at 01:17 AM
I’m not saying there are always an infinite number of options. I am saying you cannot know what all the options are.
It’s really quite simple – if you want to prove an option true you have to prove it true, it is not good enough to say all the other options cannot be true so ______ (your preferred option) must be true. All you would be demonstrating is your own lack of imagination. Please read this link:
I didn’t just dream this up you know.
Posted by: Skeptico | August 09, 2005 at 09:01 AM
Whether you dreamt it up or not has no bearing, does it? Are you attempting an appeal to authority?
If we assume that the page you linked to is correct (ie, that it is authoritative) then you must concede that there are situations where all the options can be known. That page lists three types of reasoning from ignorance that resemble the "appeal to ignorance" fallacy (they're under the heading "Exposure"), yet are valid. It's not that they are simply "exceptions" but rather that the reasoning is not really fallacious.
You say that excluding all possible options but one is not good enough, because it only shows a lack of imagination. A "lack of imagination" here means that all possible options have not actually been excluded. In other words, the condition that all the other options must be excluded has not been satisfied. It does not mean that the reasoning (exclude all but one option) is fallacious.
Posted by: mefisto | August 10, 2005 at 02:32 AM
Yes it does mean the reasoning is fallacious. It is fallacious because (one more time) you cannot know all the other options. That’s why (one more time) you need evidence for the thing you are trying to prove, not evidence against what you think are all the other options. It is really very simple. The “guilty verdict” court case is a perfect example of this fallacy in action. I’m sorry you don’t understand this, but it is correct.
Posted by: Skeptico | August 10, 2005 at 08:45 AM
OK, answer this: Is it possible to know ANY options?
I'm sure you agree options are knowable. So what makes it impossible to know all options? If all options were knowable, would you agree that excluding all possible options but one is logically valid?
You don't seem to be making a distinction between the validity of the reasoning and the truth/falsity of the premises. You're saying the reasoning is faulty because the premise is untrue.
You can say "I'm sorry, but it's just correct" as much as you like. You're not showing how or why it is correct.
Posted by: mefisto | August 11, 2005 at 05:07 AM
mefisto:
Of course it is possible to know some options. But it is impossible to be sure you know all the options. Unless the options are strictly limited by the question (as in the case of math problems), you can not know all the options. That is why you need evidence for a proposition, not what you think is evidence against all the other options. It is fallacious reasoning because you can never know that you have included all the other options. I don’t know how much clearer I can make it – it’s very arrogant to think you know everything, and it is a logical fallacy. Perhaps examples would help.
Example 1
If you read the original post, I gave an example that shows the fallacious nature of the argument. The evidence given in court was just evidence against chance. Since chance had been ruled out, the only option left to consider was murder. But there were other options. Do you think you know them all? Here are some:
• Genetic predisposition
• Mould in the house
• Radon gas in the house
• Poor food storage and preparation methods
• Poor treatment of the children due to ignorance
• Or perhaps they lived close to a major transport hub?
Did I include all options? If I rule out all of these is the Mother guilty of murder? Are there any other options – please list them all for me.
Example 2
Creationists say the eye could not have evolved through small beneficial changes, and so it must have been designed.
They have no evidence for the designer, just evidence against evolution. But how do they know it could not have arisen through evolution? Can you list all the possible ways the eye could evolve? Here are some clues. Here are some more. Do you think you know all the ways? List them all please.
Example 3
Some people believe that crop circles are caused by aliens. But they have no evidence for this; just evidence of very complex circles they claim couldn’t have been produced by humans. But this is just evidence of their own lack of imagination – how do they know humans could not have produced them?
This is not evidence for aliens. Other solutions could be:
• Ball lightening
• Other atmospheric conditions
• Wind vortexes
• Large groups of well organized and skilled human hoaxers.
By “ruling out” humans they have not shown aliens were involved. To do that they must show positive evidence for aliens, not negative evidence for humans. Unless you can list them all – please do so for me.
Unless you are sure you can list ALL the other options above, you need evidence for the thing you are trying to prove, not evidence against what you think are all the other options.
Posted by: Skeptico | August 11, 2005 at 12:34 PM
Since you look like you need a hug or something, Skeptico, I know exactly the point you're making.
Sherlock's quote might be useful in a crunch, but a trial is not a crunch.
The prosecution didn't prove its case. It just claimed the defense was improbable. That, however, doesn't mean anything: The burden of proof is on the prosecution, since they are the claimant. If it weren't for the idiocy of jurors, the defense could have slept through the whole thing and won by default.
Posted by: BronzeDog | September 15, 2005 at 03:38 PM
Scenario: I win the lottery. The odds of it happening are one in 400 bajillion. It's very unlikely I won by luck, therefore I must have rigged the lottery, and there's no need to show security camera footage, signs of tampering, or fingerprints on the balls. I am thus convicted of fraud.
Posted by: BronzeDog | September 15, 2005 at 05:21 PM
Ok Skeptico,
It's time to admit you were wrong.
Otherwise this discussion is going to bring the whole idea of skepticism and critical thinking into disrepute.
Yes the expert, the defence lawyer, the jury, the judge, and the law were all incompetent in their own various ways and the women should all have been found not guilty in their original trials.
And yes, their original convictions were unfair and that's a bad thing about which we should all feel angry.
BUT
Our emotional reaction to that particular set of examples should not be used to try to add weight to a basically fallacious argument.
In fact the statement that you claim to be disproving is logically unassailable and your argument against it is totally bogus.
Let's look again at what Sherlock Holmes is alleged to have said: “When you have eliminated the impossible, whatever remains, however improbable, must be the truth”
You challenge this by identifying an example where *none* of the conditions of Holmes' alleged thesis apply.
1 What was "eliminated" in the trials was not "the impossible" but just the merely improbable. (That the estimate of improbability was obtained by multiplying suspect probabilities of manifestly non-independent events is only a small part of what was wrong with that analysis, but that issue is of course irrelevant to the point under discussion(as is also the fact that exactly the same argument could have been used by the defense to prove the "impossibility" of the mothers being guilty!))
2 "Whatever remains" after eliminating SIDS (if in fact that had been done) would have been "anything but SIDS" and certainly not just "murder". It is not Mr. Holmes' fault that everyone involved lacked the imagination to consider the possibility that the children had been, for example, bitten by poisonous amazonian spiders (brought into the house in the trousercuffs of their anthropologist uncle whose footprints in the hallway had the unmistakable colour of ...)
Of course it is often difficult to come up with an exhaustive list of possibilities (except by using the obvious ploy of negation - though I admit that that would be using logic, which is close to mathematics, which I gather makes you uncomfortable).
But just because it is often difficult to meet Holmes' criteria in a useful way does not mean that it is never possible.
For example, if a body is found at the bottom of a swimming pool, it is easy to identify an exhaustive set of possibilities regarding the relative times of death and uncovered facial immersion, and the elimination of all but one might usefully establish the actual case.
Science and jurisprudence abound with such legitimate uses of the Holmes' criterion, and the claim that one (or many) examples of its misuse make it invalid is definitely not an example of good "critical thinking".
Alan
P.S. The value of your posting might be saved by changing "Wrong! - You can never know all other options, and you can never know they are impossible." to something more appropriate such as "True, but it is often hard to identify all the other options and to show that they are completely impossible."
Posted by: | October 30, 2005 at 01:31 PM