ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Models (was The class of the planet Venus)

To: "[ontolog-forum] " <ontolog-forum@xxxxxxxxxxxxxxxx>, John F Sowa <sowa@xxxxxxxxxxx>
From: Pat Hayes <phayes@xxxxxxx>
Date: Sat, 14 Jul 2012 00:26:17 -0500
Message-id: <419E1955-208F-4A34-A6DC-7EC72D26D2D9@xxxxxxx>
(I had vowed to not get involved in this endlessly-repeated discussion yet 
again, but ...)    (01)

There are, as people have noted, two different senses of the English word 
"model" being used here. These really are different senses, and they do not 
naturally fall under a single general heading. They have almost nothing in 
common. One of them - the 'logical' usage of the word - is a technical use, and 
just like many other cases where a word has been co-opted to serve in a 
technical capacity , its technical meaning is essentially unrelated to its 
pre-formal meaning. (Other examples include "field" , "modular" and "wreath" in 
mathematics, which have nothing whatever to do with the English meanings of 
those words.)    (02)

The logical, technical, usage of "model" refers to a semantic interpretation of 
a sentence which makes the sentence true: a satisfying interpretation. A 
"model" in this sense *is* a satisfying intepretation of a sentence. In the 
rest of this email I will use "satisfying interpretation" for this sense and 
reserve the word "model" for its normal English sense, extended to such things 
as engineering models, architectural models of buildings, etc..     (03)

Saying X is a model of Y implies that Y is some part of reality, and (as Petri, 
quoted by John, says) that X bears some structural similarity to some aspect of 
Y. In contrast, to say that X is a satisfying interpretation of Y implies that 
Y is a set of sentences (*not* a part of reality) and X bears a particular 
algebraic relationship to it which renders it true according to the semantic 
rules of the language of Y. (This algebraic relationship is typically not a 
"structural similarity" - which in algebra would be a homomorphism - but in 
almost all cases is what is referred to as a Galois connection, which if 
anything is a kind of complement to a homomorphism: the interpretations get 
fewer as the sentences get larger.)      (04)

There are cases of models where the model is itself language-like or symbolic. 
A map or engineering diagram, or even a set of equations, can be called a 
"model". At this point, however, one has to adjust the Petri definition: if the 
model is a description, then the way it bears a 'structural similarity' to the 
world it describes is not through its structure as a set of sentences, ie its 
linguistic syntax, but rather through how that syntax is interpreted. The model 
is accurate to the extent that what it says about the part-of-reality it 
describes is in fact true. So we have come full circle: the modelling 
"similarity" is the linguistic/semantic truth. But notice that (as often 
happens when one goes in a circle) we have undergone a reversal of roles. The 
"model" now is the sentences, not their interpretation: When a model is a 
description, the X of [X models Y] is the Y of [X is a satisfying 
interpretation of Y]. And, even more potentially confusing, the satisfying 
interpretation of those sentences is not the model[1], but the part of the 
reality it is a model of. So in this situation, where a model is a description 
(in a language with a semantics), the logical sense of "model" and the 
pre-formal, or engineering, sense of "model" are *exactly opposite* from one 
another: X is an engineering model of Y just when Y is a logical "model" of X: 
that is, just when Y is a (piece of reality which is) a satisfying 
interpretation of X.       (05)

Allowing models to be descriptions, and then using the logical terminology to 
describe that situation,  does more than simply muddle two meanings of "model", 
it gets them exactly backwards from one another, an almost uniquely confusing 
clash of usage.     (06)

Pat    (07)

[1] Logically trained readers, recall I am NOT using "model" in the semantic 
sense here.     (08)

On Jul 13, 2012, at 12:07 PM, John F Sowa wrote:    (09)

> On 7/13/2012 12:43 PM, Burkett, William [USA] wrote:
>> A bit more to add to the discussion of "models", a "model"
>> may have several, likely overlapping,  purposes...
> 
> I agree.
> 
> But all those purposes  -- Inquisitive, Descriptive, Predictive,
> Informative, Prescriptive, and Representative -- are consistent
> with and can be supported by models that fit the basic definition:
> 
> JFS
>> Petri... observed that when you say X is a model for Y, that means
>> X has some structural similarity Z to an important aspect of Y.
> 
> The reason why people develop models is that they are usually easier
> to build, modify, and analyze than the real-world systems they model.
> 
> That property makes them more convenient for all the above purposes.
> 
> John
> 
> 
> 
> 
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/ 
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> 
>     (010)

------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes    (011)






_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (012)

<Prev in Thread] Current Thread [Next in Thread>