| weak induction | strong induction | |
|---|---|---|
We proceed by (weak) induction on $k$. Our predicate $P(x)$
is that $x$ can be written as $0 + 1 + 1 + \cdots + 1$.
If $k=0$ then $k$ is 0 plus zero 1s, so $P(0)$ is true.
The inductive case is:
1: P(k) [Assumption A] 2: k = 0 + 1 + ... + 1 [definition of P(k)] 3: k + 1 = k + 1 [reflex. of equality] 4: k + 1 = 0 + 1 + ... + 1 + 1 [subst. of 3 using 2] 5: P(k+1) [definition of P] 5: P(k) => P(k+1) [close assumption]By the principle of induction $P(k)$ holds for all non-negative $k$. |
We proceed by (strong) induction on $k$. If $k=0$ then $k$ is 0 plus zero 1s. If $k>0$, then $k-1$ greater than or equal to zero by the Gap Theorem, and $k-1 \lt k$, so by the inductive hypothesis, $k-1$ is equal to $0$ plus $k-1$ 1s. So we get $$ k-1 + 1 = \underbrace{0+1+\cdots+1}_{=k-1\text{ by inductive hyp.}}+1 = k $$ and deduce that every non-negative integer $k$ can be written this way. |
ACTIVITY The following are results you should be able to prove without induction. They are all pretty straightforward.
1: 7 = 2*3 + 1 [Given] 2: ∃k[7 = 2*k + 1] [Existential introduction on 1] 3: ∀n[odd(n) <=> ∃k[n = 2*k + 1]] [definition of odd] 4: odd(7) <=> ∃k[7 = 2*k + 1] [specialize n to 7 in 3] 5: ∃k[7 = 2*k + 1] => odd(7) [equivalence splitting on 4] 6: odd(7) [modus ponens on 5 and 2]
Note: From now on, to prove a number is odd, we will take it as sufficient to write it as "2 * something + 1", since the remainder of the proof (lines 2-7 above) is boilerplate - i.e. it's the same regardless of what it is we are trying prove the oddness of. Similarly, if we can write a number as "2 * something" we will take that as sufficient for proving that the number is even. So, in the future, our proof that, for example, 11 is odd will look like:
Note: this "prose proof" is acceptable, because (hopefully!) all of you could "fill in the blanks" and produce a full first-order proof if you had to:
| prose proof lines | filling in the blanks |
| Assume $n_1$ and $n_2$ are even. | 1: even(n1) 2: even(n2) |
| Then by the definition of even, $n_1 = 2*k_1$ and $n_2 = 2*k_2$ for some $k_1$ and $k_2$. | 3: ∀n[even(n) <=> ∃k[n = 2*k]] [Definition of even] 4: even(n1) <=> ∃k[n1 = 2*k] [Specialization of 3 using n=n1] 5: even(n1) => ∃k[n1 = 2*k] [Equivalence splitting] 6: ∃k[n1 = 2*k] [Modus ponens 5, 1] 7: n1 = 2*k1 [Existential instantiation on 6] 8: even(n2) <=> ∃k[n2 = 2*k] [Specialization of 3 using n=n2] 9: even(n2) => ∃k[n2 = 2*k] [Equivalence splitting] 10: ∃k[n2 = 2*k] [Modus ponens 5, 1] 11: n2 = 2*k2 [Existential instantiation on 6] |
| So $n_1 + n_2 = 2*k_1 + 2*k_2 = 2*(k_1+k_2)$ | 12: n1 + n2 = n1 + n2 [Reflex. of equal] 13: n1 + n2 = 2*k1 + 2*k2 [Subst. of equal on 12 using 7,11] 14: n1 + n2 = 2*(k1 + k2) [Ring axiom 3.i (distributivity)] |
| so $n_1+n_2$ is even | 15: ∃k[n1 + n2 = 2*k] [Existential intro on 14, k = k1 + k2] 16: even(n1 + n2) <=> ∃k[ n1 + n2 = 2*k] [Spec. 3, n = n1 + n2] 17: ∃k[ n1 + n2 = 2*k] => even(n1 + n2) [Equiv. split. on 16] 18: even(n1 + n2) [Modus ponens 17, 15] |