Q1: 20 points Q2: 20 points Q3: 10 points Q4: 20 points Q5: +10 points Q6: 30 points Total: 100+10 points
Question text is in black, solutions in blue.
Here are definitions of sets, predicates, and statements used on this exam.
Remember that the score of any quantifier is always to the end of the statement it is in.
Let D be a finite set of dogs consisting of exactly the four distinct dogs Cardie (c), Duncan (d), Mia (m), and Scout (s).
Let N be the set of natural numbers {0, 1, 2, 3,...}.
There are three owners O1, O2, and O3.
Let BT(u, n) be the relation from D to N defined so that BT(n, u) means "dog u belongs to owner On".
Let E be the binary relation on D defined so that E(u, v) means "dog u goes out earlier than dog v".
Let Y be the binary relation on D defined so that Y(u, v) means "dog u is younger than dog v". You may assume that no two dogs in D are exactly the same age.
Some of the true/false questions also refer to an arbitrary set D of dogs with predicates T(x) meaning "dog x is a terrier" and S(x, y) meaning "dog x is sillier than dog y".
If either Cardie or Duncan goes out earlier than the other, then Mia is younger than Duncan.
Y(m, c) ∧ ((E(c, d) ∨ E(d, c)) → ¬Y(m, d))
Many people misparsed this and made Y(m, c) part of the premise
of the if-then. The placement of the word "if" in the English
should make it clear that the statement is saying that Y(m, c) is
unconditionally true.
If it is not the case that Mia is younger than both Cardie and Duncan,
then either Cardie goes out earlier than Duncan or Duncan goes out
earlier than Cardie, but not both.
Again, many people placed the "if" incorrectly -- the
parentheses make it clear here that the ¬ modifies the premise
of the if-then and not the entire if-then. Also, the "but not
both" is necessary at the end to represent the ⊕. Given the
English meaning of "earlier", it is impossible for both to be
true, but we don't take acocunt of English meaning of predicates
when we do translations.
[∀x:¬Y(x, x)] ∧ [∀x:∀z:(Y(x, z) ∧
Y(z, x)) → (x = z)]
Many people correctly reused the quantifier between the two
definitions. Many people said "Y(x, z) → ¬Y(z, x)"
instead of the given implication -- this is not a correct
definition of "antisymmetric" in general, but is correct when the
relation is also antireflexive, so I accepted this answer.
For any two dogs, one goes out earlier than the other if and only if they each
have an owner such that the earlier dog's owner has a smaller number.
Note that the literal translation is "there exists naturals i
and j" rather than "there exists owners i and j", though I did not
take off for this type error.
[∀u:&exists;n: BT(u, n)] ∧ [∀u:∀n:∀n':
(BT(u, n) ∧ BT(u, n')) → (n = n')] ∧ [∃u: BT(u,
3)] ∧ [∀u:∀v: (BT(u, 3) ∧ BT(v, 3)) → (u =
v)]
This was really a pretty difficult translation, since in both
clauses you need to say both "at least one" and "at most one" in order
to convey the meaning of "exactly one". Many people said only one or
the other. It was possible to reuse quantifiers between clauses.
Any dog is different from Scout if and only if Scout both goes out earlier than it and is younger than it. (Equivalently, Scout is the earliest and the youngest dog, with no ties.)
∃u:∀v: Y(v, u) ↔ (v = s)
Most people got this right, perhaps because the syntax of the
English
is closer to the symbolic syntax. Many people had → in place of
the ↔.
FALSE. If p and q are false, but r is true, then the first statement is true but the second is false. You cannot use the Rule of Joining inside a statement becuase it is a deductive rule, not an equational rule. In this case it is effectively inside a negation because of the →.
TRUE. This is an instance of the Rule of Joining.
TRUE. Statement VI's first clause says that every dog has one owner, which is exactly what it means for BT to be a function.
TRUE. Since BT maps four dogs to three owners, it is clearly not one-to-one. But most people (I conjecture) overlooked the fact that BT is a function not to owners but to naturals, and it is not onto the naturals because no dog is mapped to 0, 4, 5, 6, etc.
TRUE. Y' is clearly reflexive, and it is antisymmetric because Y is. Its transitivity and totality come from the English meaning of "younger" and the fact that no two dogs are the same age.
FALSE. It is clearly reflexive by the definition, and it is and transitive by the English meaning. But we know from Statements I-III that Cardie and Duncan go out at the same time, so E'(c, d) and E'(d, c) are both true. This violates antisymmetry.
TRUE. Since u is a prefix of v, v = uw for some string w. By a fact we stated in lecture, vR = wRuR, so uR is a suffix of vR by the definition of suffix.
FALSE. It's certainly possible that there exists a terrier and no non-terriers, and in this case the given implication is true. The error comes in misparsing the statement by treating the quantifier as part of the premise, rather than referring to the entire statement. Remember that the scope of a quantifier is always to the end of its statement (which is the end of any parentheses enclosing it).
TRUE. We can prove this by cases: Let x be an arbitrary dog and specify both premises to x. We then know by cases that x likes snacks whether it is a terrier or not, so by generalization all dogs like snacks.
FALSE. Let x be 11 and y be 13. Clearly 143 divides neither of these, but since xy = 143 it does divide xy. If we replaced 143 with a prime number, the implication would be true by the Atomicity Lemma, but that lemma does not apply to composite numbers.
(Note: The point value of Questions 3 was given as 15 on the test paper, though it was correct on page 2.)
From Statements I, II, and III, there is only one possible setting of the truth values of the four propositions Y(m, c), Y(m, d), E(c, d), and E(d, c). Determine this setting and show that it is the only setting that works. You may use either a truth table or a deductive argument.
The deductive argument is easier. If we assume (E(c, d) ∨ E(d,
c)), we get Y(m, d) by Statement I and ¬Y(m, d) by the second
clause of Statement II. This is a contradiction, so (E(c, d) ∨
E(d, c)) must be false, and both E(c, d) and E(d, c) are false.
Y(m, c) is true by the first clause of Statement II, so the only
question is that of Y(m, d). Since E(c, d) and E(d, c) are both
false, the conclusion of Statement III is false, and so its premise
must be false. This makes Y(m, c) and Y(m, d) both true.
The truth table is greatly simplified if you note that Y(m, c)
must be true, so you need only eight lines rather than 16.
Many people got the wrong answer because of bad translations,
usually winding up with four out of ten points.
Prove, using any or all of Statements I through VII, that Statement VIII is true. Do not assume anything about the English meaning of the predicates, except what you are given in the statements. Make your use of quantifier proof rules clear.
Since the statement to be proved is existential, our last move is
going to be Existence. We need to determine, before beginning our
proof, which dog is the right choice for u. If Scout were u, she
would be required to be younger than herself, which is impossible. If
u is any other dog, Scout will be younger than u by Statement VII,
which is good. But we also need the other dogs who are not Scout to
not be younger than u. So u must be the second-youngest dog in
D.
We know from Statements I-III that Mia is younger than both Cardie and
Duncan, so Mia is the dog we need.
Note: If we were allowed to use the English meaning of
"younger", we would know that there must be a second-youngest dog and
we could use this for u. But we are explicitly told to use only the
statements, which don't (for example) say that Y is transitive.
Once we know that u should be Mia, we can let v be an arbitrary dog
and try to prove the statement "Y(v, m) ↔ (v = s)". We do this
by cases. We have proved ∀v: Y(v, m) ↔ (v = s), and our desired
statement follows by Existence.
The majority of purported proofs for this were very bad, not
even beginning to approach the problem correctly. Some people wrote
down the conclusion and began deriving consequences of it, which is
quite backward. I did give a lot of partial credit for the "there
must be a second-yougest dog" argument.
Assuming the truth of Statements I-VIII, determine which owner owns which dog, and determine the value of the predicate E(u, v) for all possible pairs of dogs u and v. Make your use of quantifier proof rules clear.
Clearly the key statement is VI, which says that a dog goes out
earlier than another if and only if it has an owner with a smaller
number. We know that O3 has exactly one dog, that
Cardie and Duncan go out at the same time, and that Scout goes out
earlier than any dog except herself.
Scout's owner cannot own any other dog because Scout would then
not be earlier than that dog. And Scout's owner cannot be
O3 because no dog has a larger-numbered owner than that.
Scout's owner must be O1 or O2, and if it is
O2 there is only one dog (O3's) who goes out
after Scout (but there need to be three).
The instances of E that are correct are E(s, m), E(s, c), E(s,
d), E(c, m), and E(d, m). That is, Scout is earlier than all three
other dogs, and Cardie and Duncan are earlier than Mia.
Readers of past exams might have been tempted to conclude that
Cardie and Duncan have the same owner because those are the names of
my two dogs. But such readers may also have noticed that there is
more than one dog named Duncan in my neighborhood, and I never said
that this Duncan was mine. In any case, Statements I-III legitimately
tell you that they have the same owner.
14 = 2*7, 16 = 2*2*2*2, 21 = 3*7, and 24 = 2*2*2*3. I deducted a point if you included 1 as a factor, because 1 is not a prime number.
Numbers have inverses modulo one another if and only if they are relatively prime. Thus 16 and 21 have inverses modulo one another, but every other pair of numbers share at least one prime factor and so are not relatively prime. I deducted two points if you said that 16 had an inverse modulo 21 but neglected to say that 21 has an inverse modulo 16. I did not take off if you used the Euclidean Algorithm rather than the factorizations to get the correct statements about relative primality.
The inverse of 39, modulo 28, is -5 (or 23).
The inverse of 28, modulo 39, is 7.
The Simple Form of the Chinese Remainder Theorem says that this pair
of congruences is equivalent to a single congruence of the form x
≡ c (mod 28*39). We established in part (c) that 28 and 39
are relatively prime. I gave four points for simply stating this
result, and took off those same four points if your answer was a
single number rather than a congruence class.
A naive, but in this case effective, way to find c is to list
numbers that are congruent to 25 mod 28: 25, 53, 81, 109, 137,
165,.., and numbers that are congruent to 20 mod 39: 20, 59, 98,
137, 176,..., and notice that 137 is on both lists. The set of x
such that x ≡ 137 (mod 28*39) is thus the answer.
The standard way to find c is to take the linear combination
-5*39 + 7*28 from part (c), which we know is congruent to 1 both mod
28 and mod 39. By multiplying the first term by 25, we do not
change the answer mod 39 but multiply the mod 28 answer by 25.
Similarly we get the right answer mod 39 by multiplying the second
term by 25. Our expression for c is thus -5*39*25 + 7*28*20. You
were not required to evaluate this, but it turns out to be -4875 +
3920 = -955. Since 28*39 = 1092, this is equivalent to -955 + 1092
= 137, the answer we got by our first method. Again, it was
important to say "x ≡ c (mod 28*39)" rather than just offering
the single number c as your answer. If you got incorrect inverses
in (c) but used them in (d) in such a way that you would have been
right with the right inverses, I gave you full credit for (d). If
you said "I need inverses" but didn't note that you had them in (c),
I took off points.
Last modified 6 March 2016