|
18904
|
'Predicable' terms come in charged pairs, with one the negation of the other [Sommers, by Engelbretsen]
|
|
Full Idea:
Sommers took the 'predicable' terms of any language to come in logically charged pairs. Examples might be red/nonred, massive/massless, tied/untied, in the house/not in the house. The idea that terms can be negated was essential for such pairing.
|
|
From:
report of Fred Sommers (Intellectual Autobiography [2005]) by George Engelbretsen - Trees, Terms and Truth 2
|
|
A reaction:
If, as Rumfitt says, we learn affirmation and negation as a single linguistic operation, this would fit well with it, though Rumfitt doubtless (as a fan of classical logic) prefers to negation sentences.
|
|
18895
|
Logic which maps ordinary reasoning must be transparent, and free of variables [Sommers]
|
|
Full Idea:
What would a 'laws of thought' logic that cast light on natural language deductive thinking be like? Such a logic must be variable-free, conforming to normal syntax, and its modes of reasoning must be transparent, to make them virtually instantaneous.
|
|
From:
Fred Sommers (Intellectual Autobiography [2005], 'How We')
|
|
A reaction:
This is the main motivation for Fred Sommers's creation of modern term logic. Even if you are up to your neck in modern symbolic logic (which I'm not), you have to find this idea appealing. You can't leave it to the psychologists.
|
|
18893
|
Translating into quantificational idiom offers no clues as to how ordinary thinkers reason [Sommers]
|
|
Full Idea:
Modern predicate logic's methods of justification, which involve translation into an artificial quantificational idiom, offer no clues to how the average person, knowing no logic and adhering to the vernacular, is so logically adept.
|
|
From:
Fred Sommers (Intellectual Autobiography [2005], Intro)
|
|
A reaction:
Of course, people are very logically adept when the argument is simple (because, I guess, they can test it against the world), but not at all good when the reasoning becomes more complex. We do, though, reason in ordinary natural language.
|
|
18903
|
Sommers promotes the old idea that negation basically refers to terms [Sommers, by Engelbretsen]
|
|
Full Idea:
If there is one idea that is the keystone of the edifice that constitutes Sommers's united philosophy it is that terms are the linguistic entities subject to negation in the most basic sense. It is a very old idea, tending to be rejected in modern times.
|
|
From:
report of Fred Sommers (Intellectual Autobiography [2005]) by George Engelbretsen - Trees, Terms and Truth 2
|
|
A reaction:
Negation in modern logic is an operator applied to sentences, typically writing '¬Fa', which denies that F is predicated of a, with Fa being an atomic sentence. Do we say 'not(Stan is happy)', or 'not-Stan is happy', or 'Stan is not-happy'? Third one?
|
|
18894
|
Predicates form a hierarchy, from the most general, down to names at the bottom [Sommers]
|
|
Full Idea:
We organise our concepts of predicability on a hierarchical tree. At the top are terms like 'interesting', 'exists', 'talked about', which are predicable of anything. At the bottom are names, and in between are predicables of some things and not others.
|
|
From:
Fred Sommers (Intellectual Autobiography [2005], 'Category')
|
|
A reaction:
The heirarchy seem be arranged simply by the scope of the predicate. 'Tallest' is predicable of anything in principle, but only of a few things in practice. Is 'John Doe' a name? What is 'cosmic' predicable of? Challenging!
|
|
13127
|
Categories can't overlap; they are either disjoint, or inclusive [Sommers, by Westerhoff]
|
|
Full Idea:
Fred Sommers, in his treatment of types, says that two ontological categories cannot overlap; they are either disjoint, or one properly includes the other. This is sometimes referred to as Sommers' Law.
|
|
From:
report of Fred Sommers (Types and Ontology [1963], p.355) by Jan Westerhoff - Ontological Categories §24
|
|
A reaction:
The 'types', of course, go back to Bertrand Russell's theory of types, which is important in discussions of ontological categories. Carnap pursued it, trying to derive ontological categories from grammatical categories. 85% agree with Sommers.
|
|
12887
|
A whole must have one characteristic, an internal relation, and a structure [Rescher/Oppenheim]
|
|
Full Idea:
A whole must possess an attribute peculiar to and characteristic of it as a whole; there must be a characteristic relation of dependence between the parts; and the whole must have some structure which gives it characteristics.
|
|
From:
Rescher,N/Oppenheim,P (Logical Analysis of Gestalt Concepts [1955], p.90), quoted by Peter Simons - Parts 9.2
|
|
A reaction:
Simons says these are basically sensible conditions, and tries to fill them out. They seem a pretty good start, and I must resist the temptation to rush to borderline cases.
|