Affirming the consequent and doing science and history 3 Jan 20123 Jan 2012 Here’s a conundrum for the simple minded: One of the classic fallacies is the fallacy of affirming the consequent: If P then Q, Q, therefore P It’s an obvious logical fallacy because there might be many reasons for Q. And yet, all science rests on doing just that. Suppose I wish to test whether Quargles are caused by Pruntles. I set up the conditions for Quargles to be produced by Pruntles, and find that Quargles result. I have therefore tested the hypothesis If Pruntles then Quargles. No, says the logician (i.e., Popper), you haven’t. But remove affirming the consequent from science, and science would grind to a full stop. Science done Popper’s way would end science altogether (and this is shown by the fact that those scientists who claim to be Popperians have to do all kinds of acrobatics to show that testing hypotheses is actually a kind of falsification of non-hypotheses). Experimental replication is still affirmation of the consequent. Now historical hypotheses are of the form: If P then Q[Event E is a kind of P; Outcome O is a kind of Q]E occurredO resultedE caused O (If P then Q is true) The test is that the rule If P then Q is consistent with E and O for many E’s and many O’s. But each one is a case of affirming the consequent. Therefore all history is unscientific (I have read people saying exactly this). But I say we do know about the past, and we do it by affirming the consequent, just as we use affirmation of the consequent for most science. What’s up with that? Epistemology Philosophy Science
Freedom Phobosophy 14 Aug 201714 Aug 2017 As everyone knows, philosophy comes from the two Greek words philo and sophos, and means, roughly, the love of wisdom, although as everyone also knows, Socrates declared his wisdom was his knowledge that he knew nothing. In recent years (by which I mean increasingly since the 1970s), there has been… Read More
Epistemology 50 words for snow 6: constructing phenomena 11 Dec 20171 Mar 2019 Series Conceptual confusion The economics of cultural categories What are phenomena? What counts as sociocultural? Species Constructing phenomena Explanations and phenomena There is a naive empiricist view held by nobody on close inspection, that phenomena merely present themselves to the observer, and call for explanation. At least since Kant, such… Read More
Epistemology More Whewell 26 Jul 2009 When we have thus obtained, with reference to any such subject as those we have here spoken of, these two portions of science, a systematic description of the facts, and a rigorous analysis of the causes, — the Phenomenology and the Ætiology of the subject, — we are prepared for… Read More
The cigar is in the mail. But it is a way of raising some common confusions about fallacies and science.
Yes, good one! Being a bit illiterate I had to look it up, but especially with Pierce’s emphasis on economy of explanation it does seem to fit the bill.
It’s an obvious logical fallacy because there might be many reasons for Q. And yet, all science rests on doing just that. No, that is not at all how science works. Suppose I wish to test whether Quargles are caused by Pruntles. I set up the conditions for Quargles to be produced by Pruntles, and find that Quargles result. I have therefore tested the hypothesis If Pruntles then Quargles. That confuses an operation on abstract propositions with a physical action testing causation. Moreover, the scientist will vary the conditions of his experiments an attempt to ensure that there was no extraneous cause involved. The result might not be 100% certain, but it is not at all the same as confirming the consequent. And, I should add, there’s usually a lot of other supportive knowledge, else the scientist would not have been trying that experiment anyway. But remove affirming the consequent from science, and science would grind to a full stop. No. But removing experiment testing would indeed be a problem. Science done Popper’s way would end science altogether This is probably true.
I think your logic is too terse to apply to scientific testing. You’d need more than just P and Q, say H for hypothesis. IMHO, scientific testing then goes: Deduction (or is it abduction?): If H was true then doing P should result in Q. Test: P is done, Q results. Conclusion: H cannot be refuted. It is tentatively affirmed.
Ha. For a start your paragraph beginning “It’s an obvious logical fallacy…..” doesn’t in any way touch upon ‘affirming the consequent’. Unless, that is, you meant the italic phrase to read ‘If Quargles then Pruntles”. In any case, your wording is sloppy. Your sentence “I wish to test whether Quargles are caused by Pruntles” is not logically the same as “If P then Q” or “Q therefore P”. Your sentence wording needs to include words like ‘always’ and ‘only’. Also, if you “set up the conditions for Quargles to be produced by Pruntles,” then by definition, Quargles will result. I doubt it shows anything. I don’t think that any science in which I was ever involved, involved this fallacy, and certainly no scientist that I have ever met would agree with your set of statements. Science just doesn’t work that way. (I have heard the odd student try it, but they soon get corrected. On the other hand I have heard the odd medical person try to use it.) I suspect that removing ‘affirming the consequent’ would probably not make a ripple in the pond of scientific endeavour. Andrew
Oh dear! I think the respected Wilkins is a little sleepy today after new year whisky and risks accidentally confirming the hypothesis that philosophy of science are of interest to philosophers but not to scientists according to his own brand of logic… or illogic… because Andrew Hobbs is correct: for someone proclaiming a logical fallacy you immediately jumped feet first into your own by implying that: if P then Q implies: if Q then P Oh dear, oh dear! As to the Popperish scientific test of: if P then Q it is not saying: if P and Q then (hypothesis is confirmed) rather it is rather a test of: if P and not Q then (conjecture is refuted) and the Popper trick is to try to find experiments that would be expected to fail if the conjecture/hypothesis is bad. But surely we’re beyond this simple-minded first order logic and its childish fallacies these days? Doesn’t that Reverend Bayes wave his magic wand here? The Bayesian argument is that a test of P co-occurring with Q lends support to the belief that P causes Q (depending on temporal and other considerations; more strictly it just lends support to the belief that they are not conditionally independent). It’s the old black swan thing isn’t it? However many white swans us Europeans observe, we can’t prove that all swans are white. First order logic is for first year philosophers. Not for the real world.
Well, I was rather convinced by Nelson Goodman that it is induction, not deduction, which is the basic mode of thought. And induction is all about statistical re-evaluation. And deduction is the troublesome step-son of our thoughts.
I agree. Deduction is a limit case of ampliative reasoning. I am somewhat bemused by some of the commentators above who think I was making the error of thinking deduction is the basic form of reasoning. They clearly did not read carefully.
‘Fraid I don’t get the point at all. OM writes about induction as a mode of thought and JW replies about deduction and reasoning. Thought, reasoning? I’m not sure what those terms mean and which one is talking about. The human brain generates its hypotheses by intuition and emotion, doing a load of pattern matching with a bit of extrapolation that looks like induction. That ain’t logic. The awful “trolley scenario” in moral philosophy excites the philosophers because the subjects come up with apparently inconsistent conclusions but that is completely unsurprising because brains don’t work logically. But the strength of science is that it says “fie!” to intuitive reasoning and says “show me!” and demands a solid argument to justify any hypothesis. It is precisely because that is unnatural for the human mind that science has to be so insistent on the scientific method and good justifications. If that argument is “reasoning” then of course it’s not deductive, it’s always constructed post hoc to support the conjecturer’s conjecture… even if it’s presented to the reader in a deductive format. So a scientist’s reasoning “hmmm… I see this, that looks a possible explanation, yes, that other data supports it”, is not the same as the formal experiments and arguments that are required to test and support the conjecture. The process of scientific thought and reasoning in constructing a hypothesis is not the same as the structure of scientific argument used to support it.
You observe a phenomenon (ordinary thought). It doesn’t make sense to you, so you give up and declare it to be senseless. Well, that is the epitome of bad philosophy. Goodman’s claim is that deduction cannot be the basis for scientific thought. He has compelling evidence. Check “Fact, Fiction and Forecast” – chapters 3,4. It is a great read. Our claim here is that deduction is actually not the basic tool of science, but a dangerous tool, secondary to induction. This means that the outcrop of science is never certain. Deduction in the extreme is Mathematics – which won’t take you anywhere without the inductive input from reality. Goodman actually goes much further than that. You should really read him. He’s great.
How can we absolutely know something about the past? That is my assumption that helps me with practical living skills, but how can we know? 🙂
I mean that I assume that I know some things about the past, which helps me with practical living skills. For example, assuming I know nothing about the past would make the simplest decisions way too difficult : -) But how can I justify knowing anything about the past other than its practicalness : -)
“But how can I justify knowing anything about the past” As a simple minded historian I think the answer to that is easy. History is a real subject, despite the difficulties in dealing with the dead we can’t escape the conclusion that we have a past and despite bias or the limitations of our brains, it is possible to be guided by actual historical events and to shape worthwhile questions about them. Its this reality that gives the subject its power and means it cannot be ignored. At least that is the perspective of one historian who I agree with.
I agree that history is a real subject, and that there is a past sequence of events, unlike eternalist theory. Moreover, we make worthwhile conclusions about the past that helps our decision making in the present. But perhaps this would be nothing but circular reasoning if we claim that such worthwhile conclusions about the past in some way justifies knowledge of the past.
Maybe you can give us a concrete example of a historical hypothesis, how it has been affirmed, and how this has been criticised as fallacious. We probably can then get over the stumbling block that you claim science to be as fallacious and see what you meant to say about history. After all, everybody seems to have his own opinion on the logic of science, here.
Hi There Evolvingthoughts, Along the same lines,, Since most people here probably like to think about Science, read this the whole way through and tell me what you think about it. In Philosophy of Science and some Research Methodology classes, you learn that Science says that given the evidence, such and such theory (explanation) is most probable. You can never prove or disprove, or even say “probably true”. Proof is not logical, disproof is logical but not practical. In the Philosophy of Science you learn that this is because the first is the logical fallacy of “affirming the consequent” and the second is using “modus tollens” which is genuine deductive logic and deductive logic is only as true as your starting assumptions are. To illustrate what you learn in a Philosophy of Science class: If there’s a universal rule saying if “If A then always B”, does that necessarily mean if B then A? There could also be a rule saying “If C then B” and “If D then B”, so you can’t even say “If B then probably A” (logical error of affirming the consequent). However, if you find that there’s not B, then there can’t be A (using modus tollens is genuine deductive logic). When testing a scientific theory (explanation that’s testable), you can’t observe it directly, so you form a hypothesis where you say “If Theory A is true then B will happen.” After you conduct the research, even if it’s 100% true that “If Theory A then B” and you find that you have B, that doesn’t prove Theory A is correct, because countless others could also produce the same test results even if you don’t brainstorm them. If you have “strep throat you have a sore throat”, but not necessarily vice versa, most of the time when people have a sore throat they don’t have strep. If you find that B didn’t happen, then that only disproves Theory A if you can know for sure that there wasn’t B and that “If Theory A then B” is 100% true. However, in real life that’s not possible and so disproof is logical but not practical. Deductive logic only guarantees a correct answer if the starting assumptions entered in are correct, and even if those assumptions are derived using deductive logic, somewhere down the line inductive logic was used to create starting assumptions. So our example: Gravity. When you drop something and it falls, no one denies that it falls towards the earth. However, “Gravity” is just an explanation (theory) for “why” it happens. No one has ever been able to observe or measure gravity directly. Scientists believe in gravitation because it’s testable and hasn’t been disproved. However, the concept of what gravitation is will sometimes change. Newton said gravitation is an actual force. Then Einstein came along and said gravitation is not really a force, but rather spacetime curvatures that mass creates with his Theory of Relativity, and that’s why two bodies of mass are attracted to each other. Experiments find that the Theory of Relativity works in situations where Newton’s theory of gravitation actually being a force fails. However, physicists and mechanical engineers usually use Newton’s theories because they usually work and have less complicated mathematics than Relativity does. Then later, physicists found that even Relativity Theory has problems and that other theories work where Relativity and force theory don’t, but still use them because they’re less complicated and usually work. Basically, no amount of experiments could ever ever prove gravity/gravitation, nor exactly what it is. All they do is test it, see that they fail to disprove it, and test alternative theories. Seeing that those alternative theories don’t seem to work, they claim that they have more confidence in gravitation, but don’t ever come to absolute truth. So in Science you never prove or disprove, nor even say that a theory is probably true, but instead “Given the evidence, this theory is the most probable.” BTW great blogpost
Well said. However one of the points I was trying to make is that I don’t think science (as in your (Lucia Castro’s) description) corresponds to ‘If P then Q, Q therefore P’ in a strictly logical sense. Science is all about the most useful explanation given the circumstances, not about whether a theory is right or wrong.
Okay I have been a bit bemused by the reactions to this one. Let me say a few things. 1. The point of this was to argue that logic is not the same thing as science, and that falsificationism is based on a merely logical point. Also, I am a great fan of Nelson Goodman, and a proponent of ampliative inference (in fact, I am in print doing so). 2. The underlying motivation was to argue that creationist objections to historical inferences about evolution and geology, etc., are not warranted (this was obvious where I first posted it, in a forum for discussing these things). 3. Of course logic is about abstraction. But this is true of all inference. So what? How else can we make inferences if not on abstractions? 4. Andrew Hobbs’ comment is wrong, I think. If P then Q says only that if P obtains, Q will obtains. It does not mean that if Q obtains, P did, though. That’s the whole point about the fallacy. The modal necessity is built into the definition of if-then conditionals. 5. Sam C assumes I made the logical fallacy. Oy. 6. James, we can’t absolutely know anything about anything (except in the views of some people, analytic truths, which I think are rare to nonexistent). Hope this clarifies things. It was meant to be a teaching moment. I am a very poor teacher, it seems.
Sorry, please don’t assume that my tangents in any way reflect on your teaching skill. I sometimes cannot resist going in a tangent.
“I have been a bit bemused by the reactions ” Perhaps everyone got the same Christmas gift as my son. It’s all the rage I understand. Science Putty “Bend it, shape it, anyway you want it! Bounce, shatter, tear, shape and play (warning may stain garments, carpets and upholstery)”
Hi Jeb, You appear to support that we can know worthwhile things about the past. And then you take jabs at science. Do you reject that the scientific method can teach us worthwhile things about the past?
How do Microsoft and the owners of Firefox make money? Google is planning to create a browser also to compete with them called “chrome”. I was wondering where does the source of revenue come in? There are no ads, etc. These programs are free..