Deduction and induction

[lpetrich]

The validity of induction is an old conundrum of philosophy. Everybody accepts the validity of deduction, so how does induction fit in? If one treats induction as a kind of deduction, one finds it to be a fallacy, Affirming the Consequent alongside Denying the Antecedent. Or at least so it seems at first sight.

I once saw the counterargument that one's experience validates induction, but that is itself an inductive argument.

Has anyone gotten anywhere on this question?

 

[steve_bnk]

MH370 demonstrates is inductive vsdeductive analysis.

Trying to reason out causes ad reachconclusions going from the specific to the general and viceversa.

The plane disappears, the s[specific. Assigning causeswithout hard data, top down deductivereasoning.

http://en.wikipedia.org/wiki/Deductive_reasoning

A car crash occirs and two teams investigate. One starts atthe crash end poinytand tries to determine causation. Top down.

Theother team looks at the road and conditions trying to determinecausations. Bottoms up analysis.

There is no special inductiveor deductive logic. The difference is the starting point.

Working backwards from the conclusion to detailsor working forward to a conclusion from details.

For anaxiomatic system like Boolean Logic, given a set of conditions, theconclusions are unambiguously defined.


In general problem solving like thecar crash example, the problem is induction and deduction may notalways agree.


You can't have one without the oter.


From the book How To Read And DoMathematical Proofs, the backwards-forward method is alternatingbetween deduction and induction.


In a closed completely defined logicalsystem, one only has to apply the rules.

The application of logical reasoning does not infer a valid conclusion. A logally vvalid syllogism can lead to a false conclusion. In CS parlance, garbage in garbage out.

 

[Bomb#20]

I once saw the counterargument that one's experience validates induction, but that is itself an inductive argument.

Has anyone gotten anywhere on this question?

Deduction and induction are both subsets of pattern recognition. Pattern recognition isn't guaranteed to give right answers, but we believe in its results anyway, because pattern nonrecognition is guaranteed to prevent us from passing on whichever genes made us pattern nonrecognizers. Does that count as validation?

 

[ficino]

I always thought that talk about induction started with the followers of Socrates. They would portray him as reasoning from a string of examples to some general conclusion. I.e. from a string of examples in which members of a class x are F, he would conclude by induction that x's are F. This is called an epagogic procedure. Aristotle says it's one of the things Socrates came up with.

Today I see the tendency to call "inductive" an argument that uses one or more probable premises. Its conclusion is only as strong as the least probable of the premises.

To me, the connection seems to be that by an epagogic procedure like Socrates', we can only reach a probable conclusion. There might be a black swan out there somewhere. So when we use that conclusion as a premise in a further argument, it's only a probable premise.

Anyone with more expertise is invited to correct me!

 

[steve_bnk]

Deduction and induction are both subsets of pattern recognition. Pattern recognition isn't guaranteed to give right answers, but we believe in its results anyway, because pattern nonrecognition is guaranteed to prevent us from passing on whichever genes made us pattern nonrecognizers. Does that count as validation?

I'd answer you but I can't decide to analyze your postdeductively or inductively..



To me the basic difference is where you startanalyzing. Sherlock Holmes was mostly deductive, he worked top down or backwards from the crime.



In one of the old movies Watson is pasting newspaperclippings into an album. He loses track of one and says out loud 'what would Holmes do?'. Watson then commences to work his motions in reverse finding the clipping stuck to his elbow.



If you pay attention as to how you analyze things Ithink you will find it is almost always a combination of the two.








...

 

[fast]

If something is an argument, and if it's not a deductive argument, then it's a non-deductive argument. No inductive argument is a deductive argument, but not all non-deductive arguments are inductive arguments. Either way, no non-deductive argument (and thus no inductive argument) are valid arguments, for only deductive arguments can be valid. In fact, that's why there is a genuine distinction (despite purported necessary truths to the contrary) that some arguments can be not valid yet not invalid. For instance, a deductive argument that is not valid is in fact invalid, but an inductive argument that is not valid is not therefore invalid, for a non-deductive argument is not the kind of argument that can be valid or invalid.

 

[Cheerful Charlie]

The validity of induction is an old conundrum of philosophy. Everybody accepts the validity of deduction, so how does induction fit in? If one treats induction as a kind of deduction, one finds it to be a fallacy, Affirming the Consequent alongside Denying the Antecedent. Or at least so it seems at first sight.

I once saw the counterargument that one's experience validates induction, but that is itself an inductive argument.

Has anyone gotten anywhere on this question?

The answer is, induction (abduction) is a necessary part of life but not always sure. You may for example have incomplete or misleading facts to work with or may only play the odds which may leave errors possible. That is why in science we have peer review, tests, replication, et al. If because of this we reject induction, we will fail far more than be mislead. And, like math, some people are good at this and some not so good.

 

[Speakpigeon]

No inductive argument is a deductive argument, but not all non-deductive arguments are inductive arguments. Either way, no non-deductive argument (and thus no inductive argument) are valid arguments, for only deductive arguments can be valid. In fact, that's why there is a genuine distinction (despite purported necessary truths to the contrary) that some arguments can be not valid yet not invalid. For instance, a deductive argument that is not valid is in fact invalid, but an inductive argument that is not valid is not therefore invalid, for a non-deductive argument is not the kind of argument that can be valid or invalid.

Can you give an example of what you see as a typical inductive argument?
EB

 

[steve_bnk]

Rather than argue in the abstract, itis much easier to talk to tangible examples.

If you believe inevolution then us humans can never at any time have or have hada-priori knowledge of anything.


How did ancient Zog learn to predictthe weather? IF the sky looks a certain way and the air has a certainsmell THEN a storm may be imminent. IF and THEN are abstractions that try to describe a process in the brain, but it is not the process.


The conundrum occurs when you try toreduce it all to a an abstract logical system and discover it gets more fuzzy the more you try to define it.


The brain naturally makes correlationsand selects behavior based on predictions of the future. Cats, dogs,birds, and chimps do it.


Given something happen like getting burned by fire , the brain correlates the conditions that led to theevent. Cats, dogs, birds, and chimps do it.


In real life experience neitherinduction nor deduction are infallible.


In teal life we do both everyday. Lookfor it in how you analyze new situations and events.


The analysis of flight 370 has been both deductive and inductive. With the North Korea - South Korea tensions over unification and the China-Japan tensions over islands, is predicting various future scenarios inductive or deductive?

 

[Speakpigeon]

Rather than argue in the abstract, itis much easier to talk to tangible examples.

Can you put your example into some sort of formal shape like if X then Y or something?

If you believe inevolution then us humans can never at any time have or have hada-priori knowledge of anything.

DNA qualifies as a store of a priori knowledge. If only Kant had known of it!

Applying the principle of charity, even knowledge provided by our senses can be regarded, if not as a priori knowlege per se, but at least as what people thought about when they talked about a priori knowledge.

The brain naturally makes correlationsand selects behavior based on predictions of the future. Cats, dogs,birds, and chimps do it. Given something happen like getting burned by fire , the brain correlates the conditions that led to theevent. Cats, dogs, birds, and chimps do it.

Bees do it to. They have a brain.

I always liked bees.
EB

 

[Speakpigeon]

The validity of induction is an old conundrum of philosophy. Everybody accepts the validity of deduction, so how does induction fit in? If one treats induction as a kind of deduction, one finds it to be a fallacy, Affirming the Consequent alongside Denying the Antecedent. Or at least so it seems at first sight.

I once saw the counterargument that one's experience validates induction, but that is itself an inductive argument.

Has anyone gotten anywhere on this question?

Both deduction and induction are inferences and as such can both be said to be either valid or invalid although not quite for the same reasons.

A valid deduction is definitely very compelling and the truth of the consequent (conclusion) seems to depend only on the truth of the antecedent (premise).

A valid induction is much less compelling. First, the premise can be wrong just as in the case of deduction. But even if the premise be true we understand that the conclusion can still be wrong, and often is. For example, all prejudices are more often wrong inferences than true ones. In science, which relies both on deduction for making prediction and induction for conceiving models or theories, new observations regularly prove even entrenched theories wrong, which shows that induction is limited in principle in that scientists cannot usually observe all relevant events, not least those that will take place in the future. But the reality is that we have nothing else for now. That may change though, but that depends on the nature of reality not just on how bright we are.

In deductions, however, premises are often just plain wrong. This is because we build most of our premises on inductive reasonning, for example, "all men are mortal", which may or may not be true.

Still, you can look at induction and deduction as means of convincing each other of our views to arrive at a consensus decision and collective action through dialogue and debate rather through the murder of those we disagree with. I think it beats the totalitarian principle, the law of the jungle, God knows best etc.
EB

 

[steve_bnk]

On one side of an impassible river arifle is placed on a stand and fired into woods across the river.

The bullet passes through the woodsnicking branches and leaves ending embedded in a tree.

The people firing the rifle assess the rifle accuracy, ballistics, wind, air density, and develop a model ofthe woods predicting where the bullet ends up.

The people in the woods find the bulletin a tree. Using ballistics and tracing a path back through the woodsthey develop a predictive model of where the bullet came from..

Which approach is inductive and which deductive reasoning, Why would either analysis be any more valid than the other?

 

[Speakpigeon]

In formal logic, validity is a technical term. It just means that an implication is true in all logical cases.

From that, we can deduce that if the premises are true of the material world then the conclusion has to be true of the material world too.

The thing is, we rarely know that any premise is really true of the material world.

In:

All men are mortal;
Socrates is a man;
Therefore Socrates is mortal.

Maybe premise 1 is wrong. Maybe not all men are mortal (how would you know?).

Maybe all men except Socrates are mortal. You want to say that Socrates therefore is not a man? Fine, then it is premise 2 which is wrong.

In any case, the reasoning is still logically valid but it no longer apply to the material world because one or both premise are not true of it.
EB

 

[fast]

In formal logic, validity is a technical term. It just means that an implication is true in all logical cases.

From that, we can deduce that if the premises are true of the material world then the conclusion has to be true of the material world too.

The thing is, we rarely know that any premise is really true of the material world.

In:

All men are mortal;
Socrates is a man;
Therefore Socrates is mortal.

Maybe premise 1 is wrong. Maybe not all men are mortal (how would you know?).

Maybe all men except Socrates are mortal. You want to say that Socrates therefore is not a man? Fine, then it is premise 2 which is wrong.

In any case, the reasoning is still logically valid but it no longer apply to the material world because one or both premise are not true of it.
EB

i'm not sure why you're bringing up all this talk about the material world, and whether the premises are or are not true hasn't much to do with the underlying issue, and the issue has to do with the form of the argument. If (if, i say) the premises are true, then it's not merely the case that a conclusion in a valid argument is (i said, "is") true, but it MUST be true.

 

[fast]

No inductive argument is a deductive argument, but not all non-deductive arguments are inductive arguments. Either way, no non-deductive argument (and thus no inductive argument) are valid arguments, for only deductive arguments can be valid. In fact, that's why there is a genuine distinction (despite purported necessary truths to the contrary) that some arguments can be not valid yet not invalid. For instance, a deductive argument that is not valid is in fact invalid, but an inductive argument that is not valid is not therefore invalid, for a non-deductive argument is not the kind of argument that can be valid or invalid.

Can you give an example of what you see as a typical inductive argument?
EB

Every time I’ve walked by that dog, he hasn’t tried to bite me. So, the next time I walk by that dog he won’t try to bite me.

 

[Perspicuo]

Induction is probably the instinctive theory-producing strategy, and is probably the reason people believe in God or any other unempirical entity handed down within the individual's tribe/troop.

 

[fast]

Induction is probably the instinctive theory-producing strategy, and is probably the reason people believe in God or any other unempirical entity handed down within the individual's tribe/troop.

Acting only on conclusions guaranteed to follow from their respective premises while ignoring all conclusions only strongly suggestive to be true can be mighty restrictive.

 

[no-one-particular]

deduction tells you whether something is true or false.
Induction tells you whether it is reasonable or unreasonable

to know that something is true beyond a reasonable doubt requires both

 

[Perspicuo]

Induction is probably the instinctive theory-producing strategy, and is probably the reason people believe in God or any other unempirical entity handed down within the individual's tribe/troop.

Acting only on conclusions guaranteed to follow from their respective premises while ignoring all conclusions only strongly suggestive to be true can be mighty restrictive.

Definitely. The reason we are smart is also the reason we believe in bull. Being scaredy cats about anything that goes bump in the night, and also heeding to and being mightily interested in the horror stories we are told, has given our species strong survival advantages.

 

[steve_bnk]

deduction tells you whether something is true or false.
Induction tells you whether it is reasonable or unreasonable

to know that something is true beyond a reasonable doubt requires both

Deduction and induction use the same logical reasoning. Explain how deduction can be more reliable than induction.

The computer science cliché 'garbage in garbage pout' applies.

Logic alone regardless of mode is never a guarantor of truth in any real world sense.