Date
|
What we discussed/How we spent our time
|
Aug 21
|
Syllabus. Policies. Text.
The use of AI in this class.
|
Aug 24
|
Read Chapter 0 (pages 1-10).
We discussed the connection between logic
and language, and the role of
predicates in particular.
We practiced formalizing three of the
axioms of set theory (the axioms
of extensionality, empty set, and pairing).
|
Aug 26
|
We discussed:
- How finite sequences are defined in set theory.
- The fact that if $A$ is an alphabet of size $\leq \kappa$,
$\kappa$ infinite, then there are $\leq \kappa$
finite sequences that are constructible from the alphabet $A$.
This implies that there are only countably many
formulas in a given countable alphabet.
- The
types of symbols needed
to express the Goldbach Conjecture: variables,
logical symbols, equality, nonlogical symbols
(constant symbols, operation symbols, predicate symbols),
and punctuation symbols.
- The
types of symbols needed
to express the continuity of a function: variables,
logical symbols, equality, nonlogical symbols
(constant symbols, operation symbols, predicate symbols),
and punctuation symbols.
- The construction of formulas
in a 1-sorted language, $\mathcal{L}$. We completed the
description of $\mathcal{L}$-terms.
|
Aug 28
|
We discussed:
- Terms.
- Atomic formulas.
- Formulas and sentences.
- Formula trees.
Then we spent time working in groups
trying solve Problem 1 from HW1. (You do not have to turn this
problem in.)
|
Aug 31
|
Read Chapter 1, Sections 1-5 (pages 11-54).
We discussed:
- The language of Sentential Logic
(= propositional logic = 0th-order logic).
(Introduced by Stoic philosopher
Chryssipus ~250 BCE.)
- Sentential formulas. Wffs.
- Formula trees.
- Truth assignments.
Tautologies and contradictions.
Quiz 1 solutions.
Latex source.
|
Sep 2
|
We discussed:
- Unique readability of propositional formulas.
- Completeness of sets of connectives.
|
Sep 4
|
We discussed:
- Completeness of sets of connectives.
The first major topic was Post's Classification.
Then we discussed a corollary to Post's Classification:
a set $\Omega$
of logical connectives is complete iff the following five
statements are true:
- $\Omega$ contains an operation that fails to preserve the relation $\{0\}$.
(I.e., some $f(x_1,\ldots,x_n)\in\Omega$ satisfies $f(0,\ldots,0)\neq 0$.)
- $\Omega$ contains an operation that fails to preserve the relation $\{1\}$.
(I.e., some $f(x_1,\ldots,x_n)\in\Omega$ satisfies $f(1,\ldots,1)\neq 1$.)
- $\Omega$ contains an operation that fails to preserve the relation $\leq$.
(I.e., for some $f(x_1,\ldots,x_n)\in\Omega$
and some $a_i\leq b_i$, $i=1,\ldots,n$,
$f(a_1,\ldots,a_n)\not\leq f(b_1,\ldots,b_n)$.)
- $\Omega$ contains an operation that fails to commute with negation.
(I.e., some $f(x_1,\ldots,x_n)\in\Omega$ fails to satisfy
$\neg f(x_1,\ldots,x_n)=f(\neg x_1,\ldots,\neg x_n)$.)
- $\Omega$ contains an operation that fails to commute with ternary
addition modulo $2$.
(I.e., there is some $f(x_1,\ldots,x_n)\in\Omega$
and tuples $\mathbf{a}, \mathbf{b},\mathbf{c}$ such that
$f(\mathbf{a}\oplus \mathbf{b}\oplus\mathbf{c})\neq
f(\mathbf{a})\oplus f(\mathbf{b})\oplus f(\mathbf{c}))$. )
- Disjunctive Normal Form..
(We briefly mentioned Conjunctive Normal Form, too.)
|
Sep 7
|
Labor Day! No class.
|
Sep 9
|
We reviewed Posts' work on completeness of connectives,
including a discussion of the five relations that are
barriers to completeness.
We then started discussing Truth ($\models$) versus Provability ($\vdash$).
We applied the definition of $\models$ to show that
$\{(p\to q), (q\to r)\}\models (p\to r)$.
Newly defined terms included valuation and proof.
|
Sep 11
|
We reviewed Truth versus Provability
in the context of propositional logic.
We examined three silly proof systems,
where each failed one of the following
properties:
- Soundness.
- Completeness.
- Decidability.
We selected a proof system where
Axioms = all tautologies and
Deductive Rules = only Modus Ponens.
We explained why this proof system
is Sound, finitely Complete, and Decidable.
We ran out of time to explain why this proof system
is fully Complete. For this, we need the Compactness Theorem (page 59).
|
Sep 14
|
We discussed the Compactness Theorem following
these slides.
Quiz 2 solutions.
|
Sep 16
|
We proved the Compactness Theorem for Propositional Logic.
(The two main stages of the proof
involved solving Exercises 1 and 2 of
Section 1.7.)
|
Sep 18
|
We worked on a group quiz for most of the hour.
In the last 10 minutes we discussed
alternative proof systems for Propositional Logic
where we decrease the number of axioms or
increase the number of deductive rules.
|
Sep 21
|
We practiced ``think, pair, share'' at our seats
on Problems 1(e) and 1(f) of Section 2.1.
Then we discussed how to assign meaning
to a sentence in first-order logic following
these slides.
Quiz 3 solutions.
|
Sep 23
|
We started by discussing a potential ambiguity
in the HW4 problem Exercise 6 of Section 2.1.
The ambiguity was: what does little $a$ represent?
Is it a variable, a constant, or a predicate symbol?
The leading suggestion was: it is a constant representing
a single person whose last name is ``Adams''.
Next, we began a discussion of Models and Truth.
We discussed the possibility of multisorted models, but
I proposed that we avoid multisortedness by introducing extra
predicates. Then we introduced valuations and
explained how they can be used to decide truth in a model.
We reviewed this handout
from August 26 (skipping the exercises), and the first two pages
of this handout.
|
Sep 25
|
Read Section 2.2.
We worked on this worksheet.
Then we discussed how to assign meaning
to arbitrary terms and formulas using valuations
in a structure.
|
Sep 28
|
Read Section 2.2.
We continued our Sep 25 discussion about
how to assign meaning
to arbitrary terms and formulas using valuations
in a structure.
Quiz 4 solutions.
|
Sep 30
|
Read Section 2.2.
HW5 due date changed to October 2.
We continued our Sep 25-28 discussion about
how to assign meaning
to arbitrary terms and formulas using valuations
in a structure.
Today we discussed the consequence relation
following these slides.
|
Oct 2
|
Read Section 2.2.
We discussed (i) definability of a class
of structures and (ii) definability within a structure.
For Item (i), we discussed how Galois Connections ecode
binary relations following these notes.
Terminology associated to this point of view include
the concept of an elementary class
and the concept of a theory. We also defined
the concept of elementary equivalence.
For Item (ii) we discussed how to define new
constants, operations, and predicates in a structure.
In one example, we explained how to define
$\{0\}$, $\{1\}$, $\{2\}$, $[0,\infty)$, $\leq$, and
$[1,2]\cup [3,4]$ in $\langle \mathbb{R}; \cdot,+\rangle$.
|
Oct 5
|
We discussed how to use automorphisms to
establish the undefinability
of a relation in a structure.
We followed these slides.
Midterm Review Sheet.
Quiz 5 solutions.
|
Oct 7
|
We discussed the
Midterm Review Sheet.
The midterm will be held on Friday, October 9, in class,
from 12:20-1:10pm.
|
Oct 9
|
Midterm Exam.
No quiz next Monday. No HW assignment due next Wednesday.
|
Oct 12
|
We discussed the Midterm Answer Key.
No quiz today. No HW assignment due Wednesday.
|
Oct 14
|
We introduced and discussed the following concepts.
- homomorphism
- strong homomorphism
- isomorphism
- automorphism
- embedding
- substructure
We explained why an isomorphism $h\colon \mathbf{A}\to \mathbf{B}$
induces a satisfaction-preserving bijection between
valuations in $A$ and valuations in $B$. We derived that
$\mathbf{A}\cong \mathbf{B}\to \mathbf{A}\equiv \mathbf{B}$.
We also derived that any relation definable in $\mathbf{A}$
must be closed under automorphisms. Finally, we stated the
Łoś-Tarski Preservation Theorem,
which can be formulated as: embeddings
preserve
existential sentences and reflect universal sentences.
|
Oct 16
|
We discussed universal sentences versus existential
formulas. Then we defined $\Pi_m$- and $\Sigma_n$-formulas.
We then discussed prenex form to
explain why every formula is equivalent
to a $\Pi_m$- and/or a $\Sigma_n$-formula.
We practiced with these exercises.
|
Oct 19
|
Read Section 2.4.
We discussed a proof calculus for first-order logic. (Page 112.)
The only deductive rule is modus ponens.
The sets of axioms is the closure under $\forall$
of the set of formulas that are:
- First-order tautologies,
- Universal specification,
- Quantifier distribution,
- Quantifier introduction,
- $(x=x)$, and
- $(x=y)\to (\alpha=\alpha')$ if $\alpha'$ is obtained
from $\alpha$ by replacing zero or more $x$'s with $y$'s.
Quiz 6 solutions.
|
Oct 21
|
We discussed Quiz 6 and HW 7.
After this, we compared the structure of
Enderton's proof calculus
to Monk's proof calculus.
In the final ten minutes we discussed
Universal Specification and Variable Capture.
|
Oct 23
|
We discussed meaning of Axiom Group 2 and
the validity of Axiom Group 3.
Then we
spent time on this worksheet.
|