Non-Classical Logics
Review for the Midterm Exam
Spring, 2007
The midterm will be an open-book exam: you can bring Priest,
An Introduction to Nonclassical Logics, and use it however you wish during
the exam. I'll try to mostly write questions for which you can't simply look up
the answers, though! But making the exam open-book will allow you to look up
tableau rules, specific facts about the individual systems, etc. (Warning: the
index of the book isn't particularly good!)
Important Systems:
- Classical propositional logic
- basic modal logic (no restrictions on the accessibility relation)
- normal modal logics (restrictions: rho, sigma, tau, eta, and
combinations thereof)
- non-normal modal logics
- conditional logics: C (no restrictions), C+, S, C1,
C2
- intuitionist(ic) logic
- many-valued logics: K3, LP (don't worry about L, RM3)
- FDE
- K4, N4, K*, N*
- Fuzzy Logic
Sample questions:
These are only samples, but hopefully they give the general idea of the sorts
of questions I will ask. I want (mostly) questions that will require you to
think about and apply the ideas we have discussed.
- What does it mean to say that one logic is an extension of another
logic? What does it mean to say that one logic is a sub-logic of
another? If logic L1 is an extension of logic L2,
which one is a sub-logic of the other? If L2 is an extension
of L1, and T is a theorem of L2, must it also be a
theorem of L1? If T is a theorem of L1, must it
also be a theorem of L2? Explain your answer.
- Consider the following relation R between logics: R(L1, L2)
iff L2 is an extension of L1. Is this relation
reflexive? symmetric? transitive? extendable? Explain.
- Suppose we wanted to a modal logic for reasoning about space. What
would "worlds" correspond to? If we use K_upsilon (no accessibility
relation: every "world" is accessible from every "world"), what would
the box mean? What would the diamond mean?
- Present and evaluate a philosophical motivation for logic X (where X
is a variable ranging over K, C, K4, N4, fuzzy
logic, etc.).
- proofs in the various logics using the tableau method and/or finding
countermodels.
- Stalnaker advocated condition 6, p. 88; Lewis recommended replacing
it with condition 7, p. 90. What do these conditions mean, in ordinary
English? What do they have in common? How are they different? Give an
example that shows that counterfactual conditionals in English do not
seem to be compatible with condition 6.
- Invent a tableau method for fuzzy logic, and prove that it is sound
and complete.
- What is the definition of a relevant logic?
- Prove that K4 is not a relevant logic. (You can do this
by proving, in K4, a theorem that is incompatible with the
definition of a relevant logic.)
- As you know, in K3, p horseshoe p is not a logical truth. (Countermodel:
v(p) = i.) Now, keep the same truth tables as K3, but revise the
definition of logical truth as follows: p is logically true iff p is
true in every supervaluation of every interpretation. Show that p
horseshow p is a logical truth in the new system.
- A different question with a content similar to the previous one.
Define logical truth as in the previous question, and consider the truth
tables of K3. Consider the sentence: p horseshoe (q v ~q). List every
distinct interpretation of this sentence. (That is, give all the
possible distinct functions v from the atomic sentences p and q to truth
values.) (There should be nine.) For each of these functions, list every
possible supervaluation of that function. (There should be a minimum of
one and a maximum of four.) Is the sentence true in every supervaluation
of every interpretation?
- Suppose that A
⊨
B (in a given logic). What property of the tableau method for that logic
guarantees that every branch of a tableau that begins with A and ~B will
close?
- Suppose that, in some logic, I build a tableau whose first node
contains the sentences A and ~A, and I find that every branch of the
tableau closes. What property of the tableau method for that logic
guarantees that A
⊨
B ?
- The characteristic axiom of K_eta is:
□p
⊃
◊p.
(a) Explain why, if this axiom is a logical truth, the accessibility
relation R must be extendable. (You can do this by proving the
contrapositive, i.e. by showing that if R is not extendable, there is a
countermodel.) (b) Explain why, if R is extendable, the axiom must be
true in every world of every interpretation, i.e. must be a logical
truth.
- I could ask the same question as 14 for the axioms related to other
restrictions on R. For example, reflexivity is related to the logical
truth of
□p
⊃
p; symmetry is related to the logical truth of p
⊃
□◊p;
and transitivity is related to the logical truth of
□p
⊃
□□p.