Talk:Analogy

The first paragraph of this article made me cringe. It was so overly complex and difficult to understand:

I changed the first paragraph to:

The rest of the article probably reads just as badly, and is overly complex. I added the {{Copyedit}} tag.

As a law student, I read complex/nasty legislation and cases everyday. My job will be to interpret complex ideas and text. This paragraph was even difficult for me to understand.

Travb (talk) 14:48, 4 March 2007 (UTC)

--Agreed, I am not sure if it was changed back or what, but the first paragraph is horrible. Even for those looking for more than a definition, that is quite a confusing and seemingly random way to start an article about a word commonly used to mean "comparison" (slightly simplified, but basically true).Mfergason 14:06, 8 June 2007 (UTC)

I'm sure the wording can be improved, but please notice that analogy isn't comparison nor similarity. The word analogy is subject to Wiktionary. In Wikipedia the philosophical and cognitive science subjects of analogy must be addressed. Velho 13:44, 9 June 2007 (UTC)

Yes, Analogy is not necessarily a simple comparison A is like B. There is analogy of "proportionality" which means that "the relationship between A and B is like the relationship between X and Y, in some way": for example if you said "the relationship of monarch and subject is like the relation between parent and child": it is not even saying the monarch is directly like a parent or parent is like a monarch but the two relationships are (in some way) alike in that they have the same proportionality.Rick Jelliffe (talk) 06:32, 11 June 2023 (UTC)

At some stage the opening sentence was reverted back, I think correctly. But it was still a mess. I have re-jigged it in two ways, without attempting a better definition:

• the first is editorial: instead of giving the broad then the narrow definition, I give the narrow then the broader definition. This is less vexing to the grey matter, I think.
• The second fixes the central problem with the defintion, that it is internally contradictory: it says that analogy is a comparison between two particulars, where one of them is not particular but general! So I have pulled out the "where" sentence and softened it to say that analogies can also have general terms, which is the only thing the original sentence can have meant.

Rick Jelliffe (talk) 06:32, 11 June 2023 (UTC)

OK, I bit the bullet and added a new first sentence, see #Definition above.Rick Jelliffe (talk) 16:33, 13 June 2023 (UTC)

In logic, it is pretty obvious that analogy is simply two logical inferences: induction followed by deduction.

Suppose an analogy inference gives p(b) from p(a), this can be reduced to two inferences:

By induction: p(a) -> p(x)   (constant a is generalised to a variable x)
By deduction: p(x) -> p(b)   (variable x is specialised to a constant b)

where p is a predicate symbol, a and b are constants and x is a variable. Clearly there also has to be some additional knowledge that specifies the domain. Presumably the same approach can be extended to more complex logical formulae. Pgr94 14:52, 20 June 2007 (UTC)

Charles Sanders Peirce wrote something to this effect. Pgr94 15:46, 20 June 2007 (UTC)

I guess the article already makes clear that sometimes we can go from p(a) to p(b) without accepting (or being able to accept) p(x). See the cited article by Juthe. Besides, what you mean is that analogy gives the same results as induction + deduction give, not that analogy is induction + deduction. Velho 01:52, 22 June 2007 (UTC)

The article currently has subsections on reducing analogy to induction ("special case of induction") and reducing analogy to deduction ("Hidden deduction"). Both inferences are inadequate on their own, as you need both. Not sure about the Juthe paper as it completely fails to mention Peirce who considered the problem nearly a century earlier! Pgr94 07:30, 22 June 2007 (UTC)

Analogy is both the cognitive process of transferring information from a particular subject (the analogue or source) to another particular subject (the target), and a linguistic expression corresponding to such a process. In a narrower sense, analogy is an inference or an argument from one particular to another particular, as opposed to deduction, induction, and abduction, where at least one of the premises or the conclusion is general. The word analogy can also refer to the relation between the source and the target themselves, which is often, though not necessarily, a similarity, as in the biological notion of analogy.

Some of the language and grammar used within this article needs simplifying. I'm quite highly educated and the first paragraph required me to read over it once or twice to understand fully what it was conveying.--GateKiller (talk) 11:39, 1 September 2008 (UTC)

Completely agree with this. I know what an analogy is, but if I didn't then the introductory paragraph would be incredibly difficult to understand. If you happened to be a student trying to brush up on your English then it would be almost incomprehensible. This sentence in particular is extremely poor:

"In a narrower sense, analogy is an inference or an argument from one particular to another particular, as opposed to deduction, induction, and abduction, where at least one of the premises or the conclusion is general."

In other words, it's a type of argument in which you use a generally accepted example as a comparison to prove a point. Why not simply write that instead of over-complicating a frankly very simple concept. Bandanamerchant (talk) 09:39, 3 August 2013 (UTC)

I have rearranged and split the definition, corrected a contradiction, and adjusted some prepositions, without attempting a much better definition. I think this is an improvement (though improvement would be welcome.) Rick Jelliffe (talk) 06:35, 11 June 2023 (UTC)

Article refers to 'analogia', translates that as 'proportion'. Does that mean 'analogy' is related to neither 'ana' (as opposed to 'kata'), nor to 'logos'? Or does it simply not make sense to analyze it any further? --84.177.94.31 (talk) 09:52, 13 May 2011 (UTC)

No, it means that 'ana' and 'logos' have wider/different meanings than any English term, so when combined they mean something that surprises us. I am surprised by it too, but it is correct that it means 'proportion'. has an explanation.Rick Jelliffe (talk) 08:22, 11 June 2023 (UTC)

Analogy is both the cognitive process of transferring information from a particular subject (the analogue or source) to another particular subject (the target).

I cannot see in what sense a source and a target can be a subject, and I cannot see how meaning and information can be transferred without knowing at all what information and meaning are, whether they are possible to transfer at all. Like I assume that you are talking about discourse, texts written in a natural language and infomration and meaning are related to those tools.

It is horrible. Information is not "transferred" at all, in any idiomatic English sense of "transferring". "Applying" or "asserting" or "arguing" or "positing" or "recognizing" would be better. Rick Jelliffe (talk) 06:40, 11 June 2023 (UTC)

In an attempt to be less confusing, I have replaced "from" with "on" and "to" with "onto". This is because transfer denotes that the thing transferred has been removed from its source, which is not the case. I would prefer to use transfer in some way. (I still don't like information/meaning.) Also separated the confusing long sentence by using semi-colon. Please revert if this has made things worse, of course! Rick Jelliffe (talk) 06:53, 11 June 2023 (UTC)

OK, I think I can see where the "transferring" terminology comes from: it belongs to the "meta" in "metaphor". Which raises the possibility that it is bogus when describing analogy (outside a direct comparison with metaphor, I guess)...Rick Jelliffe (talk) 08:25, 11 June 2023 (UTC)

I have added as section on religion, and given some material in a Catholic section: analogy plays a essential role in theology (according to various recent Popes and theologians) but there is very little material for laymen or on the WWW on it. I am working on a full article on the analogia entis which I will link to when it is up: it is a real gap in Wikipedia, probably because it is so often mentioned only in passing or is an under-the-hood assumption in Catholic material. Rick Jelliffe (talk) 02:36, 27 April 2018 (UTC)

In section Analogy#Analogy and complexity, both Kolmogorov complexity and minimum description length is mentioned. In the text, analogical reasoning is implemented by searching for a description (or program) that minimizes some sum of Kolmogorov complexities.

Kolmogorov complexity depends on the description/programming language to be used to describe the various parts of source and target.^[1] The existence of a fixed bound |K₁(p) - K₂(p)| ≤ c_1,2, for arbitrary universal programming languages 1 and 2, an appropriate constant c_1,2 and every program p,^[2] is of no help in establishing that the same program will lead to minimum complexity, no matter in which language it is expressed.

As an example, consider a Turing-complete language ("language 0") that is even more verbose than Cobol, let language 1 be obtained from 0 by adding a very short language construct (say, just "+", as usual) to express addition, and let language 2 be a similar extension of language 0 to express multiplication briefly (say "*"). Then consider the analogy "2 is to 4 as 3 to ?". In language 1, the shortest program will be a description of "x is to x+x" for both source ("2 is to 4") and target ("3 is to ?"), so "?" will be answered as "6". In language 2, the shortest program will amount to "x is to x*x" for both, leading to "? = 9".

So I wonder if the Kolmogorov approach to analogical reasoning depends on the programming language, or not. Likewise, when minimum description length (apparently slightly different from Kolmogorov complexity) is used instead, what about language dependence then? - Jochen Burghardt (talk) 19:42, 6 November 2019 (UTC)

I have marked an uncited paragraph which claims there was no systematic treatment of analogy since Cajetan as dubious. (... perhaps I should have marked it as uncited.) Rick Jelliffe (talk) 05:52, 11 June 2023 (UTC)

There were two very large sections related to Structure Mapping Theory.

• One set in the Shared Structure section related really to AI/Computer science, so I moved those paragraphs to that subsection.
• Another large set under Psychology really was unnecessary details on Structure Mapping, with little concerning analogy per se: I have moved it all to the page on Structure mapping. I have kept some paragraphs, which are therefore duplicates, so that the gist/overview is still on the analogy page.

I did these moves boldly, because of the tags that the article had too much tangential etc information. I think it allows the reader to be less bogged down by information that is not quite related to the topic. Rick Jelliffe (talk) 12:06, 11 June 2023 (UTC)