Understanding through Discussion


Welcome! You are not logged in. [ Login ]
EvC Forum active members: 114 (8734 total)
Current session began: 
Page Loaded: 03-28-2017 2:12 AM
420 online now:
AZPaul3, DrJones*, Minnemooseus (Adminnemooseus), PaulK (4 members, 416 visitors)
Chatting now:  Chat room empty
Newest Member: timtak
Post Volume:
Total: 802,056 Year: 6,662/21,208 Month: 2,423/2,634 Week: 86/525 Day: 1/60 Hour: 0/0


Thread  Details

Email This Thread
Newer Topic | Older Topic
  
Prev1234
5
678Next
Author Topic:   Abductive Reasoning In Science
RAZD
Member
Posts: 18241
From: the other end of the sidewalk
Joined: 03-14-2004
Member Rating: 3.0


Message 61 of 120 (672453)
09-08-2012 12:18 PM
Reply to: Message 57 by PaulK
09-08-2012 2:56 AM


Re: theory acceptance and as a good working model
Hi PaulK

I like "firstky" ... sounds Slavicishky

Curiously, I don't think we are really in disagreement here.

... Falsification is deductive, ...

That is my point, the analysis of the new data is deductive: does it falsify the hypothesis or not (presumably such analysis has already been done for any old data)? If the new data does not falsify the hypothesis then you can move on to the next step.

quote:
deductive element:
  • the hypothesis was tested for (new) predicted situations
  • the hypothesis was not invalidated by any of the new situations
  • therefore the hypothesis is valid for the new predicted situations as well as (all) old situations

Here, you have run headfirst into the problem of induction. Testing a hypothesis under new conditions is certainly useful but not for your "deduction" (which is not valid). Indeed it would seem to be an abduction i.e. success in the test is better explained by the hypothesis applying, rather than chance, or some more convoluted explanation.

Sorry about my lack of clarity here, and I'll clarify my conclusion in my previous post as follows:

• therefore the hypothesis is valid for the new predicted* tested situations as well as (all) old situations.

I did NOT mean future or untested predictions, only the ones that had been tested thus far.

K?

Enjoy

Edited by RAZD, : No reason given.


we are limited in our ability to understand
by our ability to understand
Rebel American Zen Deist
... to learn ... to think ... to live ... to laugh ...
to share.


Join the effort to solve medical problems, AIDS/HIV, Cancer and more with Team EvC! (click)

This message is a reply to:
 Message 57 by PaulK, posted 09-08-2012 2:56 AM PaulK has responded

Replies to this message:
 Message 62 by PaulK, posted 09-08-2012 12:50 PM RAZD has responded

  
PaulK
Member
Posts: 12443
Joined: 01-10-2003
Member Rating: 2.5


Message 62 of 120 (672456)
09-08-2012 12:50 PM
Reply to: Message 61 by RAZD
09-08-2012 12:18 PM


Re: theory acceptance and as a good working model
quote:

That is my point, the analysis of the new data is deductive: does it falsify the hypothesis or not (presumably such analysis has already been done for any old data)? If the new data does not falsify the hypothesis then you can move on to the next step.

Then you are only seeing the half of it. The failure to falsify does not deductively justify any use of the theory as a model at all. Even under the tested circumstances. For that you need other reasoning.

quote:

Sorry about my lack of clarity here, and I'll clarify my conclusion in my previous post as follows:

• therefore the hypothesis is valid for the new predicted* tested situations as well as (all) old situations.


But that is what I understood you to be saying - and that conclusion can only be justified inductively or abductively. My point stands.


This message is a reply to:
 Message 61 by RAZD, posted 09-08-2012 12:18 PM RAZD has responded

Replies to this message:
 Message 65 by RAZD, posted 09-08-2012 1:36 PM PaulK has responded

    
RAZD
Member
Posts: 18241
From: the other end of the sidewalk
Joined: 03-14-2004
Member Rating: 3.0


Message 63 of 120 (672458)
09-08-2012 1:11 PM
Reply to: Message 56 by bluegenes
09-08-2012 2:00 AM


hypothesis and data analysis
Hi bluegenes,

RAZD writes:

This seems to me to be especially true for the initial formation of the theory of natural selection by Darwin, as he listed a number of situations where selection (cause) appeared to affect population traits (effect).

That's observation. From those observations in particular species, he could make the inductive hypothesis that natural selection effected all organisms around the world in both present and past

Let me set up the logic here before we go further:

Initial formation of the theory of natural selection as I see it:

P1: here is a list of observed situations regarding selection and population traits
P2: none of these situations invalidate the concept that selection affects population traits
C: therefore it is not invalid\irrational to conclude that selection affects population traits in these specific situations.

The deductive element is the analysis of the data in relation to the hypothetical concept, and, IF the hypothetical concept is not invalidated, THEN you can move on to the "inductive hypothesis that natural selection effected* all organisms around the world in both present and past."

* - (don't you mean affected ?)

Of course this only applies to the situations listed by Darwin, and other situations could have existed but not be known by Darwin that would have invalidated the initial concept: I don't believe he would have ignored or cherry picked only situations that fit his concept, based on all his comments about what would have invalidated it.

Then, he could make an explanatory abductive hypothesis: that natural selection is the cause or driving force behind the "origin of species".

Deductive reasoning comes in at this point, and is very important. It is how we decide what would necessarily follow from the hypotheses that have been made. That's how we establish what their predictions are.

Of course he was working with information available at the time, but I would have said " ... is a major cause ..." as there could be (and we now know there are) other causes (neutral drift for instance, and even "major" may be an overstatement), but otherwise we are in agreement.

Enjoy.


we are limited in our ability to understand
by our ability to understand
Rebel American Zen Deist
... to learn ... to think ... to live ... to laugh ...
to share.


Join the effort to solve medical problems, AIDS/HIV, Cancer and more with Team EvC! (click)

This message is a reply to:
 Message 56 by bluegenes, posted 09-08-2012 2:00 AM bluegenes has responded

Replies to this message:
 Message 70 by bluegenes, posted 09-08-2012 7:55 PM RAZD has acknowledged this reply

  
RAZD
Member
Posts: 18241
From: the other end of the sidewalk
Joined: 03-14-2004
Member Rating: 3.0


Message 64 of 120 (672459)
09-08-2012 1:14 PM
Reply to: Message 59 by nwr
09-08-2012 11:29 AM


Re: Before Abductive, Inductive, or Deductive reasoning ...
Hi nwr,

No, no, and no (in that order).

Well then you will have to explain further. Just saying no here is, curiously, not enough for me.

Enjoy


we are limited in our ability to understand
by our ability to understand
Rebel American Zen Deist
... to learn ... to think ... to live ... to laugh ...
to share.


Join the effort to solve medical problems, AIDS/HIV, Cancer and more with Team EvC! (click)

This message is a reply to:
 Message 59 by nwr, posted 09-08-2012 11:29 AM nwr has responded

Replies to this message:
 Message 71 by nwr, posted 09-08-2012 11:43 PM RAZD has acknowledged this reply

  
RAZD
Member
Posts: 18241
From: the other end of the sidewalk
Joined: 03-14-2004
Member Rating: 3.0


Message 65 of 120 (672461)
09-08-2012 1:36 PM
Reply to: Message 62 by PaulK
09-08-2012 12:50 PM


Re: theory acceptance and as a good working model
Hi PaulK

Then you are only seeing the half of it. The failure to falsify does not deductively justify any use of the theory as a model at all. Even under the tested circumstances. For that you need other reasoning.
...
But that is what I understood you to be saying - and that conclusion can only be justified inductively or abductively. My point stands.

Curiously, you still do not seem to understand my point. Let me see if this helps:

Does new information from the test 1 falsify the hypothetical concept Yes No
Does new information from the test 2 falsify the hypothetical concept Yes No
...
Does new information from the test n falsify the hypothetical concept Yes No
(where n is the number of tests made)

Analysis of the data:

None of the new information from the tests falsifies the hypothetical concept Correct Incorrect

Is this not a deductive analysis of the new information?

Then you are only seeing the half of it. ...

Correct: the analysis of the new data half, not the inductive hypothesis half, where you induce that it applies outside the tested set of data. That is the "next step" after the analysis.

capicheky?

quote:
... If the new data does not falsify the hypothesis then you can move on to the next step.

Then you are only seeing the half of it. The failure to falsify does not deductively justify any use of the theory as a model at all. Even under the tested circumstances. For that you need other reasoning.

So my lack of clarity is from the use of "hypothesis" as opposed to "hypothetical concept" or conjecture here? ... implying that the hypothesis is already fully formed at this point?

Enjoy

Edited by RAZD, : clrty


we are limited in our ability to understand
by our ability to understand
Rebel American Zen Deist
... to learn ... to think ... to live ... to laugh ...
to share.


Join the effort to solve medical problems, AIDS/HIV, Cancer and more with Team EvC! (click)

This message is a reply to:
 Message 62 by PaulK, posted 09-08-2012 12:50 PM PaulK has responded

Replies to this message:
 Message 66 by PaulK, posted 09-08-2012 1:42 PM RAZD has responded

  
PaulK
Member
Posts: 12443
Joined: 01-10-2003
Member Rating: 2.5


Message 66 of 120 (672462)
09-08-2012 1:42 PM
Reply to: Message 65 by RAZD
09-08-2012 1:36 PM


Re: theory acceptance and as a good working model
quote:

Is this not a deductive analysis of the new information?

As I said falsification is deductive. I don't see anything significant in what you are saying here.

quote:

Correct: the analysis of the new data half, not the inductive hypothesis half, where you induce that it applies outside the tested set of data. That is the "next step" after the analysis.

You're still missing it. The idea that it applies within the tested data is not deductively true. You need induction or abduction to even conclude that. Indeed, the conclusion that you achieved the expected result in the test because the theory worked is an abductively conclusion.


This message is a reply to:
 Message 65 by RAZD, posted 09-08-2012 1:36 PM RAZD has responded

Replies to this message:
 Message 67 by RAZD, posted 09-08-2012 1:50 PM PaulK has responded

    
RAZD
Member
Posts: 18241
From: the other end of the sidewalk
Joined: 03-14-2004
Member Rating: 3.0


Message 67 of 120 (672463)
09-08-2012 1:50 PM
Reply to: Message 66 by PaulK
09-08-2012 1:42 PM


Re: theory acceptance and as a good working model
Hi again PaulK

You're still missing it. The idea that it applies within the tested data is not deductively true. ...

It applies to the analysis of the data not "within the tested data" sheesh.

The hypothetical concept\conjecture (see edit to previous post) is not falsified by the new information True Not true

Deductive: the conclusion is true if the premises are true.

Enjoy

Edited by RAZD, : clrty


we are limited in our ability to understand
by our ability to understand
Rebel American Zen Deist
... to learn ... to think ... to live ... to laugh ...
to share.


Join the effort to solve medical problems, AIDS/HIV, Cancer and more with Team EvC! (click)

This message is a reply to:
 Message 66 by PaulK, posted 09-08-2012 1:42 PM PaulK has responded

Replies to this message:
 Message 69 by PaulK, posted 09-08-2012 6:48 PM RAZD has acknowledged this reply
 Message 80 by Straggler, posted 09-09-2012 5:05 PM RAZD has responded

  
Straggler
Member (Idle past 100 days)
Posts: 10188
From: London England
Joined: 09-30-2006


(1)
Message 68 of 120 (672478)
09-08-2012 4:53 PM
Reply to: Message 49 by RAZD
09-07-2012 6:34 PM


Facts and Theories
RAZD writes:

Curiously, I thought evolution was developed from deductive reasoning from examples where it is known to occur (likewise the original Theory of Natural Selection), am I wrong?

Straggler writes:

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?

RAZD writes:

I observe P to be true in case A
I observe P to be true in case B
...
I observe P to be true in case N (where N is the total number of observations involve)
Therefore I always observe P to be true (... so far).

No. You are conflating theory and fact. This does not come as any surprise to those of us aware of your track record on these matters.

That all observed and analysed instances of life on Earth are consistent with common descent is the data that supports the theory. It isn't the theory itself. Obviously. The theory is (tentatively) that ALL life on Earth, including as yet undiscovered species, share a common ancestor.

RAZD writes:

You will forgive me if I don't just take your - or anyone else's - word, but I'm willing to entertain an explanation of it.

It is not possible to derive a generalised conclusion from specific instances via deductive logic alone. Deductive logic can only ever tell you that which was contained in your starting assumptions/premises. It is a reductive process. Scientific theories are falsifiable generalised conclusions based on necessarily limited specific cases. Thus scientific theories cannot be derived by deduction alone. It just isn't possible by the very nature of science. Falsifiable and tentative as scientific theories are they necessarily contain non-deductive (i.e. inductive and abductive) components.

If you still think that the the theory that ALL life on Earth, including as yet undiscovered species, share a common ancestor: Can be derived deductively - Try again. This time try to bear in mind the difference between theories and the facts that support them.


This message is a reply to:
 Message 49 by RAZD, posted 09-07-2012 6:34 PM RAZD has acknowledged this reply

  
PaulK
Member
Posts: 12443
Joined: 01-10-2003
Member Rating: 2.5


Message 69 of 120 (672487)
09-08-2012 6:48 PM
Reply to: Message 67 by RAZD
09-08-2012 1:50 PM


Re: theory acceptance and as a good working model
quote:

It applies to the analysis of the data not "within the tested data" sheesh.

Only if the "analysis of the data" is restricted to determining if the theory has been falsified or not.

But that is not enough for "acceptance of the theory" even tentatively "as a working model". For that you need to apply induction or abduction as I keep telling you.


This message is a reply to:
 Message 67 by RAZD, posted 09-08-2012 1:50 PM RAZD has acknowledged this reply

    
bluegenes
Member
Posts: 2967
From: U.K.
Joined: 01-24-2007


Message 70 of 120 (672498)
09-08-2012 7:55 PM
Reply to: Message 63 by RAZD
09-08-2012 1:11 PM


Re: hypothesis and data analysis
RAZD writes:


P1: here is a list of observed situations regarding selection and population traits
P2: none of these situations invalidate the concept that selection affects population traits
C: therefore it is not invalid\irrational to conclude that selection affects population traits in these specific situations.

It would be deductively invalid to conclude that selection affects those populations in those situations from the information you've given there. You could conclude that by observation if you had observed it to be the case in all of the situations (your P1 didn't say that), but then you wouldn't need to deduce anything. Or you could conclude it by induction if you had observed it to be the case in some examples and there was no observed reason it couldn't be in the others. In which case, it's tentative of course.

RAZD writes:

The deductive element is the analysis of the data in relation to the hypothetical concept, and, IF the hypothetical concept is not invalidated, THEN you can move on to the "inductive hypothesis that natural selection effected* all organisms around the world in both present and past."

What's the "hypothetical concept" at this point, if you've already established the existence of natural selection as a phenomenon by observation? You seem to be describing the testing of an inductive hypothesis. Your only deduction is working out what would logically follow if your hypothesis was correct (its predictions, and therefore its falsifications). That may be (sort of) what you're meaning to say. But comparing data to the predictions is really testing by observation after the deductive reasoning has given you those predictions.

I think one way to describe the basis of Darwin's original theory briefly is that he abduces common descent (as an explanation of the apparent relationships seen in classification), induces the generality of the observed phenomenon of natural selection from particular examples, and abduces that hereditary variations and selection on them are the processes that lead to the origin of new features and species (as an explanation of how and why it came about that there's lots of organisms around, not just one or several originals).

Deduction comes in when it's used to determine what predictions those three hypotheses and the overall theory make. Because we determine the predictions (what should necessarily follow if the hypotheses are true) by deduction from the hypotheses, it follows that falsifications (any observations that directly contradict predictions) are also determined by deduction from the hypotheses.

I don't think that last part is really too different from what you were meaning to say.


This message is a reply to:
 Message 63 by RAZD, posted 09-08-2012 1:11 PM RAZD has acknowledged this reply

  
nwr
Member
Posts: 5504
From: Geneva, Illinois
Joined: 08-08-2005
Member Rating: 5.9


(1)
Message 71 of 120 (672513)
09-08-2012 11:43 PM
Reply to: Message 64 by RAZD
09-08-2012 1:14 PM


Re: Before Abductive, Inductive, or Deductive reasoning ...
I am forever puzzled that people cannot see what should be staring them in the face.

Philosophers (from whom induction and abduction theses come), see observations as abstract proposition. And they see induction and abduction as a ways of inferring a more general abstract propositions from rather specific abstract propositions.

It is all completely solipsistic. Reality is not needed. It is just a matter of operations on abstract proposition.

If you look at science, it is not at all like that. It is all about reality.

I see a scientific theory as what connects those abstract propositions to reality. The theory is prerequisite to being able to make the observations from which the theory is allegedly induced or abduced.

I'm saying that induction and abduction are the creation myths of philosophy. They are every bit as true as the Adam and Eve story, but not one iota more true.


Jesus was a liberal hippie

This message is a reply to:
 Message 64 by RAZD, posted 09-08-2012 1:14 PM RAZD has acknowledged this reply

Replies to this message:
 Message 72 by bluegenes, posted 09-09-2012 2:46 AM nwr has responded
 Message 73 by PaulK, posted 09-09-2012 4:37 AM nwr has responded

  
bluegenes
Member
Posts: 2967
From: U.K.
Joined: 01-24-2007


Message 72 of 120 (672518)
09-09-2012 2:46 AM
Reply to: Message 71 by nwr
09-08-2012 11:43 PM


Re: Before Abductive, Inductive, or Deductive reasoning ...
nwr writes:

Philosophers (from whom induction and abduction theses come), see observations as abstract proposition.

When you say philosophers, do you mean "all philosophers"?

When a philosopher or a scientist claims that science uses those forms of reasoning, they do so based on observation. Philosophers have merely defined the forms of reasoning used and given them names. Scientists (and laypeople) were reasoning in these ways before the definitions were given. Humans are obliged to use inductive and abductive reasoning when exploring the unknown. We all do it regularly.

nwr writes:

I see a scientific theory as what connects those abstract propositions to reality. The theory is prerequisite to being able to make the observations from which the theory is allegedly induced or abduced.

I see observation as what connects a scientific theory to reality.

You can't hypothesise anything about anything if you are a brain in a void and have never observed anything.

Science is based on both observation and reasoning. An observation based hypothesis, theory or law can certainly guide you to new observations, if that's what you mean. If it predicted the existence of something, it might lead you to look for it, as with physicists searching for direct evidence of the Higgs boson or dark matter, and biologists searching for fish/amphibian transitionals.


This message is a reply to:
 Message 71 by nwr, posted 09-08-2012 11:43 PM nwr has responded

Replies to this message:
 Message 77 by nwr, posted 09-09-2012 3:49 PM bluegenes has responded

  
PaulK
Member
Posts: 12443
Joined: 01-10-2003
Member Rating: 2.5


(1)
Message 73 of 120 (672521)
09-09-2012 4:37 AM
Reply to: Message 71 by nwr
09-08-2012 11:43 PM


Re: Before Abductive, Inductive, or Deductive reasoning ...
The thing striking me in the face is the obvious fact that your description of philosophers better fits yourself than anyone else.

To raise one simple example the question of which reasoning modes are employed in science is best decided by looking at actual scientific reasoning rather than by abstract reasoning. (Even good abstract reasoning).

Please bow out of this thread. I don't want to see another train wreck like your attempt to "prove" that scientists can't use induction.


This message is a reply to:
 Message 71 by nwr, posted 09-08-2012 11:43 PM nwr has responded

Replies to this message:
 Message 74 by Dr Adequate, posted 09-09-2012 7:28 AM PaulK has responded
 Message 78 by nwr, posted 09-09-2012 3:54 PM PaulK has not yet responded

    
Dr Adequate
Member
Posts: 15482
Joined: 07-20-2006
Member Rating: 3.9


Message 74 of 120 (672526)
09-09-2012 7:28 AM
Reply to: Message 73 by PaulK
09-09-2012 4:37 AM


Re: Before Abductive, Inductive, or Deductive reasoning ...
To be fair, he does have a point about induction, though he overstates it. Science is not inductive, it's hypothetico-deductive. At best, induction is a description of a psychological phenomenon: if we see the sun rising in the east often enough, we may formulate a general law that that's what it does. But how we get the law is not really part of the scientific method, it's just a fact about how our brains tend to work.

What is curious about his objection is that he blames the concept of induction on philosophers, whereas it is minimizing the role of induction that is really an abstruse philosophical concept. It is completely obvious that science is inductive: it requires careful reasoning to discover that it isn't.

About abduction he is of course completely wrong.


This message is a reply to:
 Message 73 by PaulK, posted 09-09-2012 4:37 AM PaulK has responded

Replies to this message:
 Message 75 by PaulK, posted 09-09-2012 9:23 AM Dr Adequate has not yet responded
 Message 76 by Blue Jay, posted 09-09-2012 10:13 AM Dr Adequate has responded

  
PaulK
Member
Posts: 12443
Joined: 01-10-2003
Member Rating: 2.5


Message 75 of 120 (672538)
09-09-2012 9:23 AM
Reply to: Message 74 by Dr Adequate
09-09-2012 7:28 AM


Re: Before Abductive, Inductive, or Deductive reasoning ...
Even if that was correct he'd only be right by accident, having no valid argument or understanding of the reasoning used by scientists. However in my view the whole idea of a hypothesis being corroborated by repeated failures to falsify it is an example of inductive reasoning, and any claim that science doesn't use induction at all would certainly need to provide an alternative explanation of the reasoning used in that case.
This message is a reply to:
 Message 74 by Dr Adequate, posted 09-09-2012 7:28 AM Dr Adequate has not yet responded

    
Prev1234
5
678Next
Newer Topic | Older Topic
Jump to:


Copyright 2001-2015 by EvC Forum, All Rights Reserved

™ Version 4.0 Beta
Innovative software from Qwixotic © 2017