Difference between revisions of "logical induction and deduction"

From Lojban
Jump to navigation Jump to search
m
 
m (Text replace - "jbocre: ([L-Z])" to "$1")
 
(2 intermediate revisions by the same user not shown)
Line 1: Line 1:
  
"[[Robin's Palm Writings: l|la nicte cadzu]]" is both the title of a story (book, really) entirely in Lojban by [[jbocre: Robin Lee Powell|Robin Lee Powell]], and the name of the central organization in that story.
+
==  Deduction ==
  
Robin tends to make quite a few mistakes, both typographical and grammatical, so there is a [[jbocre: la nicte cadzu Errata|la nicte cadzu Errata]] page.
+
''I have a Case, and a Rule, and I infer a Result''
  
Creating LNC tends to involve a lot of vocabulary work, hence the [[jbocre: lo valsi po la nicte cadzu|lo valsi po la nicte cadzu]] page.
+
''Case:  Socrates is human''
 +
 
 +
''Rule:  All humans are mortal''
 +
 
 +
''Result: Therefore, Socrates is mortal''
 +
 
 +
nibli, ja'o ''(too often the longer equivalent .iseni'ibo)'', [[Resurrected Gismu idni|Resurrected Gismu idni]]
 +
 
 +
**Whence these strange definitions?  They don't work for a vast array of interesting cases:
 +
 
 +
''This or that. Not this Therefore that.''
 +
 
 +
''Socrates is human, Socrates is a philosopher, Therefore, Some philosopher is human''
 +
 
 +
----
 +
 
 +
====  Induction ====
 +
 
 +
''I have a Case and a Result, and I infer a Rule''
 +
 
 +
''Case:  Socrates is human''
 +
 
 +
''Result: Socrates is mortal''
 +
 
 +
''Rule:  Therefore, All humans are mortal''
 +
 
 +
sucta, su'a, [[Resurrected Gismu usna|Resurrected Gismu usna]]
 +
 
 +
'''Objection''':
 +
 
 +
''Case:  Socrates is human''
 +
 
 +
''Result: Socrates was a philosopher''
 +
 
 +
''Rule:  Therefore, All humans are philosophers''
 +
 
 +
''To do good induction you need a lot of case-result pairs.''
 +
 
 +
** This is more plausible in a way, but deals with only one type of induction and it the least useful.  Statistical induction and causal induction are more important and don't fit this pattern at all.
 +
 
 +
----
 +
 
 +
====  Abduction ====
 +
 
 +
''I have a Result and a Rule, and I infer a Case''
 +
 
 +
''Rule:  All humans are mortal''
 +
 
 +
''Result: Socrates is mortal''
 +
 
 +
''Case:  Therefore, Socrates is human''
 +
 
 +
tolsucta, su'anai
 +
 
 +
'''Objection''':
 +
 
 +
''Rule:  All tree frogs are mortal''
 +
 
 +
''Reslt:  Socrates is mortal''
 +
 
 +
''Case:  Therefore, Socrates is a tree frog.''
 +
 
 +
''i.e. Abduction is logically ''and'' scientifically silly; but as a (fallible) inferential mechanism it actually underlies much of human assumptions about the world.''
 +
 
 +
** Well, this pattern certainly is, but the usual abduction (in this sense of the word)is fairly sturdy: ''If H held, T would occur; If H does not hold, T is pretty unlikely to occur, T occurs, Therefore probably H holds.''
 +
 
 +
''What about all the rest of the inference types tht logic deals with? Interpretation, analogy, evaluative, not to mention again the ones under induction?'' They and abduction, too, often get buried away in "induction" but here there seems to be some sorting out.

Latest revision as of 14:46, 23 March 2014

Deduction

I have a Case, and a Rule, and I infer a Result

Case: Socrates is human

Rule: All humans are mortal

Result: Therefore, Socrates is mortal

nibli, ja'o (too often the longer equivalent .iseni'ibo), Resurrected Gismu idni

    • Whence these strange definitions? They don't work for a vast array of interesting cases:

This or that. Not this Therefore that.

Socrates is human, Socrates is a philosopher, Therefore, Some philosopher is human


Induction

I have a Case and a Result, and I infer a Rule

Case: Socrates is human

Result: Socrates is mortal

Rule: Therefore, All humans are mortal

sucta, su'a, Resurrected Gismu usna

Objection:

Case: Socrates is human

Result: Socrates was a philosopher

Rule: Therefore, All humans are philosophers

To do good induction you need a lot of case-result pairs.

    • This is more plausible in a way, but deals with only one type of induction and it the least useful. Statistical induction and causal induction are more important and don't fit this pattern at all.

Abduction

I have a Result and a Rule, and I infer a Case

Rule: All humans are mortal

Result: Socrates is mortal

Case: Therefore, Socrates is human

tolsucta, su'anai

Objection:

Rule: All tree frogs are mortal

Reslt: Socrates is mortal

Case: Therefore, Socrates is a tree frog.

i.e. Abduction is logically and scientifically silly; but as a (fallible) inferential mechanism it actually underlies much of human assumptions about the world.

    • Well, this pattern certainly is, but the usual abduction (in this sense of the word)is fairly sturdy: If H held, T would occur; If H does not hold, T is pretty unlikely to occur, T occurs, Therefore probably H holds.

What about all the rest of the inference types tht logic deals with? Interpretation, analogy, evaluative, not to mention again the ones under induction? They and abduction, too, often get buried away in "induction" but here there seems to be some sorting out.