Abductive Reasoning In Science

Well, they're completely different.(In talking about this, I'll just go with how WP defines abductive reasoning, it is possible that they are misdefining it, there are some strange confusions in the article.)Inductive reasoning goes like this: "I have a theory about how the world works. This theory predicts A, B, and C. I observe A, B, and C. So I shall take this theory to be true until and unless I find a counterexample to this general rule."Abductive reasoning goes like this: "I have a theory about how the world works (hopefully one confirmed by inductive reasoning). I observe Y. According to my theory, the only possible explanations for Y are X₁, X₂, or X₃. Therefore, one of X₁, X₂, or X₃ must be true."But as I say, the WP article is just odd. Sometimes it says one thing (which I would agree with completely) and sometimes it says another thing (which I would not endorse) and sometimes it says things that seem to me to be nonsensical (and I am not noted for being obtuse, so I think the fault lies with them and not me).

Quite right. That's my job.

I declare jihad.

quote:

---

If I can avoid a "what do words mean" type of conversation by saying so, then let me try to clarify - I guess what I'm saying is that it's difficult to reconcile formal logic with empiricism or abduction. Or, for that matter, induction. Many philosophers of science have construed this as a problem with empiricism and abduction - Hume's inductive fallacy, for instance.

---

But there isn't a real problem in reconciling them. Inductive and abductive modes of reasoning are fallible. That's all that declaring them "logically invalid" means.

quote:

---

I choose to construe the issue as being a problem with logic, one that largely relegates logic to the status of an amusing parlor game as opposed to a useful tool for grappling with the world

---

Formal deductive logic has the problem that it needs premises to reason from and can't produce them. And that's one reason why we need other modes of reasoning. Given premises, however, formal logic is useful and effective.

quote:

---

That may be, because I like to keep "valid" distinct from "true." A proposition in logic is valid when it descends from premises by means of logical transformations that preserve truth values. A proposition in logic is true when it descends validly from premises that are true. Most of the time. Of course, per Godel, under some circumstances a proposition can be valid but not true.

---

However, validity is not a property of a proposition. It is a property of the argument that leads from the premises to the proposition. Your definition is not only non-standard, it is useless since any proposition may be derived by valid logic given a free choice of premises.

quote:

---

I guess what I'm getting at is, I'm much less interested in whether abductive reasoning is logically valid, since it is useful and true, which is more important.

---

The point that abductive reasoning is fallible and should be carefully employed is important, I think. Aside from that I cannot say that validity in strict deductive logic is greatly important in any argument, unless it is presented as an argument of strict deductive logic.

Well, yeah. It's also difficult to reconcile being a bicycle with being an umbrella. 'Cos they're different things. It's not that there's a contradiction between bicycles and umbrellas, it's that they do different things.Now, the role of logic in science is this: If you think a theory is true, you are obliged to think that the logical consequences of that theory are also true. If you believe in the theory of gravity, then logic compels you to also think that planets must orbit the sun in (to a high degree of approximation) ellipses. You can't believe the theory but dispute the conclusion, because the conclusion is a logical consequence of the theory.Now, the role of logic as I have just explained it has, clearly, a huge and important role in science. Logic is what connects a theory to its predictions. Without that, there would be no such thing as science.But that is all that it does. When we have finished bowing down before logic and praising it, that's the only thing it does in science. It is not a valid critique of logic to complain that that's all it does, because this is all that (in science) it claims to do. It connects your theory to your observations.It is equally not a critique of science to complain that it is not a logical procedure. Science is logical just insofar as it uses logic to connect theories with observations. Whenever we want to do anything else, then science does not follow the rules of logic, it follows the rules of science. When science is scientific, when (for example) it is empirical, then it is not illogical, it is alogical. The question of whether it is logical simply doesn't occur, it's a category error, it's like asking an Orthodox Jew to pronounce whether his dietary laws permit me to ride a unicycle. The question doesn't even apply to the subject, since I am not asking if it is lawful to eat a unicycle.

Does anybody not from the "disagree just to be disagreeable" crowd have anything to say?

In order to reply to that, one would first have to identify the ""disagree just to be disagreeable" crowd".

I think you are over-complicating matters.Abductive reasoning in science is effectively about limiting explanations to those that have a record of success and tentatively discarding the infinite array of other possibilities until there is some evidential reason to think they might be relevant. If you see a painting of unknown origin it is reasonable to abduce that the painting in question was painted by a human rather than something that popped into existence ex-nihilo, or which was created by an alien, or which was painted by a particularly artistic badger or.....etc. etc. etc. etc. PREMISE: Humans have the capacity to produce paintings.
FACT: Exhibit A is a painting
CONCLUSION: Therefore exhibit A was produced by a human.This is an example of affirming the consequent. In deductive logical terms it is fallacious. That humans have the capacity to produce paintings doesn't mean that ALL paintings are necessarily created by humans. So the conclusion is potentially false and does not deductively follow from the premise.However given that we have no evidence of anything other than humans having created paintings (and lots of evidence that humans do create paintings) it is perfectly reasonable to abduce this deductively fallacious conclusion.Having accused you of over-complicating things I fear I am now doing the same even more so.....Edited by Straggler, : No reason given.

As has been pointed out by others - The reason for this is because deductive reasoning is not the same as abduction (or induction). They are different things. Well that is a fair point. Philosophers of science (most notably Popper I believe) have wrestled with the whole issue of science incorporating modes of reasoning that don't result in deductive certainty. It's part of the whole thorny issue of "truth" and what exactly it is that science is telling us about the world.The answers to these problems (in a simplistic nutshell) - Tentativity and verisimilitude. You go too far. Deductive logic is a crucial component of developing (and utilising) scientific theories. It just isn't the whole shebang!!

Well keep up the good work on that score.

Yes you are wrong.Can you explain how the hypothesis that ALL life on Earth, including as yet undiscovered species, shares a common ancestor can be derived purely by a process of deductive logic from the available evidence?You need to start understanding the role of different logical forms of reasoning if you are ever to understand science and how it actually works.

Hi Straggler, Logic (deductive, inductive, abductive) → hypothesishypothesis → Logic (deductive): if true, then it follows that {predictions} must be true: {predictions} → testingtesting → new informationnew information → analysis: it invalidates (contradicts) or it doesn't question (neither matches nor contradicts) predictionanalysis → reformulate hypothesis and predictions as needed to include new informationrepeatNote that failure to invalidate the {predictions} does not necessarily support the hypothesis: the new information could be just more of the {A} that is also {B} subset and there could still be {B} ≠ {A} that has not been detected yet (and which would then invalidate the hypothesis). You can deduce a cause → effect pattern from several observed instances,x causes y₁ (something causes a rock to fall)*
x causes y₂ (something causes a pencil to fall)
x causes y₃ (something causes a feather to fall)
...
x causes y_n (something causes object n to fall, where n is the number of observations involved)deduction: the pattern of cause (gravity) and effect (things falling) is the same for all n objects.then generalize (inductive\abductive) as a general application of that cause → effect pattern:x causes all y (something - gravity - causes all these objects to fall)while, in contrast, the general application is derived from a single instance (inductive) or guessed (abductive).Feel free to correct this, as I'm not wedded to it, but this is the way I read\see the process.Enjoy.* ie - fall from the hand when released and landing at, or relatively near, your feet (or the base of the leaning tower of Pisa)Edited by RAZD, : clrty, format, examples addedEdited by RAZD, : mo clrty

So would you agree that a theory is believed on inductive or abductive grounds ?

Ahhh. When in doubt pull out the Venn diagrams..... Yes you are wrong.Can you explain how the theory that ALL life on Earth, including as yet undiscovered species, shares a common ancestor can be derived purely by a process of deductive logic from the available evidence?If you can't do this just say so.....

Hi, Dr A. Oh, so my misunderstanding was more with what induction is: induction is not about explaining an observation at all, but about using observations to verify a theory, while abduction is about using a theory to explain observations.