Note. In the previous section, Section 9. Euler’s Theorem and Function we saw:
Theorem 9.1. Euler’s Theorem. Suppose that m ≥ 1 and (a, m) = 1. Then
aϕ(m) ≡ 1 (mod m).
Here ϕ(m) is Euler’s ϕ-function which is, for m is a positive integer, the number of positive integers less than or equal to m and relatively prime to m.
Note. If (a, m) = 1 then there is positive integer t such that at ≡ 1 (mod m); namely, t = ϕ(m). Of course there are infinitely many choices for t, since we could take t = kϕ(m) where k is any positive integer: akϕ(m) ≡ (aϕ(m) )k ≡ 1k ≡ 1 (mod m). We are particularly interested in the smallest value of t such that at ≡ 1 (mod m).
Definition. If (a, m) = 1 then the order of a modulo m is the smallest positive
integer t such that at ≡ 1 (mod m).
Note. Of course a = 1 is of order 1 for all m. When a = m − 1 we have
a2 ≡ (m − 1)2 ≡ m2 − 2m + 1 ≡ 1 (mod m) so that a = m − 1 has order 2. For m = 7 we have ϕ(m) = ϕ(7) = 6, and with a = 2 we have a3 ≡ 23 ≡ 1 (mod 7) so that the order of 2 modulo 7 is 3. So there are ample examples that the order of a modulo m can be less than ϕ(m). We will see below that if a is of order t modulo m then t | ϕ(m); see Theorem 10.2. The next result shows that the only exponents on a that produce a product of 1 modulo m are multiples of the order of a (from which Theorem 10.2 will easily follow). Section 10. Primitive Roots 2
Theorem 10.1. Suppose that (a, m) = 1 and a has order t modulo m. Then an ≡ 1 (mod m) if and only if n is a multiple of t.
Note. From Theorem 10.1, we easily get the following.
Theorem 10.2. If (a, m) = 1 and a has order t (mod m), then t | ϕ(m).
Exercise 10.2. What order can an integer have modulo 9? Find an example of each.
Solution. Since 1, 2, 4, 5, 7, 8 are relatively prime to 9, then ϕ(9) = 6. Since the
divisors of ϕ(9) = 6 are 1, 2, 3, and 6, the possible orders by Theorem 10.2 are also 1, 2, 3, and 6 . By the definition of “order,” we see that we only consider elements a that are relatively prime with 9. Element a = 1 is of order 1 modulo 9. Element a = 2 is of order 6 modulo 9 since 26 = 64 ≡ 1 (mod 9). Element a = 4 is of order 3 modulo 9 since 43 = 64 ≡ 1 (mod 9). Element a = 5 is of order 6 modulo 9 since 56 = 15,625 ≡ 1 (mod 9). Element a = 7 is of order 3 modulo 9 since 73 = 343 ≡ 1 (mod 9). Element a = 8 is of order 2 modulo 9 since 82 = 64 ≡ 1 (mod 9). So an element of order 1 is a = 1 , an element of order 2 is a = 8 , elements of order 3 are a = 4 and a = 7 , and elements of order 6 are a = 2 and a = 5 .
Note. We now explore odd prime divisors of powers of a, minus 1.
Section 10. Primitive Roots 3
Theorem 10.3. If p and q are odd primes and q | ap − 1, then either q | a − 1 or
q = 2kp + 1 for some integer k.
Note. With a = 2 (so that a − 1 = 1) in Theorem 10.3, we cannot have q | a − 1.
So if q | 2p − 1 then we must have that q = 2kp + 1 for some k. This is summarized in the next corollary.
Corollary 10.A. Any prime divisor of 2p −1 is of the form 2kp+1 for some integer k.
Note. We now return to a consideration of powers of a.
Theorem 10.4. If the order of a modulo m is t, then ar ≡ as (mod m) if and only
if r ≡ s (mod t).
Definition. For (a, m) = 1, if a is a least residue and the order of a modulo m is
ϕ(m), then a is a primitive root of m.
Note. The next theorem lets us use primitive roots to generate the ϕ(m) positive integers less than m that are relatively prime to m. Section 10. Primitive Roots 4
Theorem 10.5. If g is a primitive root of m, then the least residues modulo m of
g, g 2 , g 3 , . . . , g ϕ(m) are a permutation of the ϕ(m) positive integers less than m and relatively prime to it.
Note. To illustrate Theorem 10.5, with m = 9 and a = 2 we have that a is a
primitive root of m since ϕ(9) = 6 and 21 ≡ 2 (mod 9), 22 ≡ 4 (mod 9), 23 ≡ 8 (mod 9), 24 ≡ 7 (mod 9), 25 ≡ 5 (mod 9), and 26 ≡ 1 (mod 9). Now 21 ≡ 2 (mod 9), 22 ≡ 4 (mod 9), 23 ≡ 8 (mod 9), 24 ≡ 7 (mod 9), 25 ≡ 5, and 26 ≡ 1 and these are the positive integers less than m = 9 that are relatively prime to m = 9.
Note. Not every integer has a primitive root. For example, with m = 8 we have ϕ(8) = 4, but the order of a = 1 is 1, a = 3 has order 2 since 32 ≡ 1 (mod 8), a = 5 has order 2 since 52 ≡ 1 (mod 8), and a = 7 has order 2 since 72 ≡ 1 (mod 8); remember, we only consider those numbers less than m = 8 and relatively prime to m = 8. Our next goal is to show that each prime number has a primitive root (see Theorem 10.6). The proof requires three lemmas and the existence of a primitive root of a prime is given, though a technique of finding the primitive root is not part of the proof. Dudley comments (see page 77): “For these reasons, you do not lose too much if you take the result on faith.”
Lemma 10.1. Suppose that a has order t modulo m. Then ak has order t modulo m if and only if (k, t) = 1. Section 10. Primitive Roots 5
Corollary 10.B. Suppose that g is a primitive root of prime p. Then the least residue of g k is a primitive root of p if and only if (k, p − 1) = 1.
Note. The next result is reminiscent of the Fundamental Theorem of Algebra (that is, an n degree polynomial with complex coefficients has exactly n zeros, counting multiplicity). However, in considering a polynomial equivalence with an n-degree polynomial, we do not get exactly n zeros but instead at most n zeros.
Lemma 10.2. If f is a polynomial of degree n, then f (x) ≡ 0 (mod p) has at most
n solutions.
Note. Lemma 10.2 does not hold if the modulus is not prime. For example, the equation x2 + x ≡ 0 (mod 6) has more than n = 2 solutions, namely 0, 2, 3, and 5. This is because there are “zero divisors” modulo 6. Namely, 2 · 3 ≡ 0 (mod 6), yet neither 2 nor 3 is 0 (mod 6). For more on zero divisors, see my online notes for Introduction to Modern Algebra (MATH 4127/5127) on Section IV.19. Integral Domains; notice Definition 19.2.
Lemma 10.3. If d | p − 1, then xd ≡ 1 (mod p) has exactly d solutions.
Note. With Lemmas 10.1 to 10.3, we now have the equipment to prove that every prime number has a primitive root. In fact, we can also quantify the number of primitive roots. Section 10. Primitive Roots 6
Theorem 10.6. Every prime p has ϕ(p − 1) primitive roots.
Note. In the proof of Theorem 10.6, we introduced function ψ(t) as the number of integers 1, 2, . . . , p − 1 that have order t mod p. We showed that ψ(t) = ϕ(t) for each t a divisor of p − 1. Therefore, we have also proved the following.
Corollary 10.C. If p is a prime and t | (p − 1), then the number of least residues modulo p with order t is ϕ(t).
Note. We know by Theorem 10.6 that every prime has a primitive root. It is reasonable to consider other values of m for which a primitive root mod m exist. Such m are classified in the Primitive Root Theorem. A (lengthy) proof of it can be found in Amin Witno’s Theory of Numbers online book; see his Chapter 5 Primitive Roots.
Theorem 10.A. The Primitive Root Theorem.
Suppose m ≥ 2. Then primitive roots mod m exist if and only if m is 2 or 4 or of the form pα or 2pα for some odd prime p and some α ≥ 1. In particular, primitive roots mod p exist for every prime number p. Section 10. Primitive Roots 7
Note. Even though the Primitive Root Theorem lets us classify which numbers have primitive roots, it does not tell us how to find the primitive roots. Dudley comments (page 80): “No method is known for predicting what will be the smallest positive primitive root of a given prime p, nor is there much known about the distribution of the ϕ(p − 1) primitive roots among the least residues modulo p.”
Note. Recall that Wilson’s Theorem (Theorem 6.2) states: Positive integer p is prime if and only if (p − 1)! ≡ −1 (mod p). We can use primitive roots to easily prove one of the implications of Wilson’s Theorem
Theorem 10.B. If p is an odd prime then (p − 1)! ≡ −1 (mod p).
