advanced39 min studyLesson 578 in Mathematics

Completions and p-adic fields

Complete the rational numbers using a prime-adic metric and distinguish valuation normalization from finite residue precision.

On this page

What you’ll understand

  • Calculate rational p-adic valuations and absolute values.
  • Interpret convergence through growing divisibility rather than real magnitude.
  • Distinguish normalized extension valuations from those extending v_p.

Completion depends on the notion of distance. In a p-adic metric, a number is small when it is divisible by a large power of a chosen prime, even if its ordinary real size is enormous. Completing Q in this metric produces a field that retains exact arithmetic while allowing infinite prime-power expansions. Finite congruence calculations give approximations to its integers, but a residue ring at one precision is not the whole field. We fix valuation conventions carefully because ramified extensions introduce a second natural normalization.

A prime exponent defines the metric

For nonzero rational x, write x=pku/vx=p^ku/v with integers u,v not divisible by p and define vp(x)=kv_p(x)=k. Set vp(0)=∞v_p(0)=\infty and ∣x∣p=p−vp(x)|x|_p=p^{-v_p(x)}, with ∣0∣p=0|0|_p=0. The valuation is additive under multiplication and satisfies vp(x+y)≥min⁡(vp(x),vp(y))v_p(x+y)\ge\min(v_p(x),v_p(y)). Hence the absolute value is nonarchimedean: the size of a sum is at most the larger input size. If the two input valuations differ, cancellation cannot raise the smaller one, so equality holds.

Cauchy sequences for the resulting distance can fail to converge in Q, just as real Cauchy sequences can. Their completion is Q_p, and the field operations extend continuously. The p-adic integers Z_p consist of elements of nonnegative valuation. Its units have valuation zero, its maximal ideal is pZ_p, and its residue field is F_p. This local ring is different from the rational localization Z_(p), which is not complete, even though both have the same rational valuation and residue field.

∣x∣p=p−vp(x),∣x+y∣p≤max⁡(∣x∣p,∣y∣p),Zp={x∈Qp:vp(x)≥0}|x|_p=p^{-v_p(x)},\qquad|x+y|_p\le\max(|x|_p,|y|_p),\qquad\mathbb Z_p=\{x\in\mathbb Q_p:v_p(x)\ge0\}
The base normalization is v_p(p)=1. Valuation infinity and absolute value zero are used at x=0.↔ Scroll wide equations horizontally.

Digits record compatible congruences

Every p-adic integer has a unique expansion ∑j≥0ajpj\sum_{j\ge0}a_jp^j with digits from zero through p−1. Truncating after N digits determines its residue modulo pNp^N. The inverse-limit description uses compatible residues at every precision, not independent choices of residues at unrelated moduli. A p-adic field element can have finitely many negative powers as well. Infinite positive powers converge because their valuations grow without bound, rather than because their real values become small.

For a convergent geometric series, the usual finite-sum identity still determines the limit when interpreted in this metric. The same formal series may diverge over the reals. Negative rational numbers can therefore have endless nonnegative p-adic digit expansions, and that does not make them positive in an ordered-field sense; the p-adic field has no ordering compatible with this digit intuition. Congruence stability explains the approximation process more reliably than comparing the real sizes of the successive integers.

Zp≅lim←⁡NZ/pNZ,∑j=0∞pj=11−p in Qp\mathbb Z_p\cong\varprojlim_N\mathbb Z/p^N\mathbb Z,\qquad\sum_{j=0}^{\infty}p^j=\frac1{1-p}\text{ in }\mathbb Q_p
The residues must reduce compatibly. The geometric identity is a p-adic convergence statement, not a real one.↔ Scroll wide equations horizontally.

Extension valuations have two useful normalizations

For a finite extension L of Q_p, one often normalizes its discrete valuation vLv_L by vL(πL)=1v_L(\pi_L)=1 for a uniformizer. Then vL(p)=ev_L(p)=e, the ramification index. The valuation extending the original vpv_p is instead vL/ev_L/e, which can have fractional values. Confusing these conventions changes the stated value of a root of p and can make the identity degree equals ramification times residue degree appear inconsistent. State which valuation is in use before comparing base and extension elements.

The residue field of L has p to the f elements, where f is its residue degree, and the extension degree is ef. The extended absolute value with the base p-adic normalization is p raised to minus v_L divided by e. A residue field of p squared elements is not generally the ring Z/p²Z: the latter has a nonzero nilpotent class p, while the former is a field. Distinguish extension degree, residue-field size, and finite-precision residue rings throughout local computations.

Different valuations prevent cancellation

In Q5 let a=75/14 and b=125/9. Find their valuations, absolute values, and the valuation of a+b.

  1. The numerator 75 has two factors of five and denominator fourteen has none, so v5(a)=2 and |a|5=1/25. The numerator 125 has three factors, giving v5(b)=3 and |b|5=1/125.

  2. Because the valuations differ, the smaller valuation survives addition: v5(a+b)=2. The nonarchimedean bound is an equality in absolute value, with |a+b|5=1/25.

  3. Direct addition verifies this: a+b=(675+1750)/126=2425/126=25 times 97/126. Neither 97 nor 126 is divisible by five, confirming exactly two factors in the numerator after reduction.

The result

A term of higher valuation is smaller p-adically and cannot cancel the leading lower-valuation contribution.

Explore the model

Interactive lab

Interactive lab · modular

Finite precision is a quotient, not the entire p-adic field

Pause and Reset, then set both controls to modulus m=3 and additive step a=1; predict return to zero after three Steps. Reset and set m=9 and a=1; predict return after nine Steps. These show the additive quotients Z3/3Z3 and Z3/9Z3. Reducing the second residues modulo three recovers the first precision, but the instrument has a finite modulus and does not represent all compatible p-adic digits or a field extension.

Finite precision is a quotient, not the entire p-adic fieldAddition modulo m starts at zero: rₙ = na mod m. Playback advances two additions per second. The orbit has period m/gcd(m,a), including period one for a = 0. This is not a cryptographic security test or a primality algorithm. Current readings: Residue r0: 0; Orbit period · m/gcd(m,a): 12; Distinct residues visited: 1 of 12. All parameters are adjustable with the labeled controls below.Repeated addition modulo 12: the current residue is 0, and the orbit period is 12.01234567891011+ 5mod 12An additive orbitn = 0rₙ = 0Period = 12Cycle starts at 0:0 → 5 → 10 → 3 →8 → 1 → 6 → 11 →4 → 9 → 2 → 7 → 0Exact residues · two additions per second
Additive orbitCurrent residue / next edge
Residue r0
0
Orbit period · m/gcd(m,a)
12
Distinct residues visited
1 of 12
t = 0 s

Addition modulo m starts at zero: rₙ = na mod m. Playback advances two additions per second. The orbit has period m/gcd(m,a), including period one for a = 0. This is not a cryptographic security test or a primality algorithm. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.

Look forP-adic integers retain compatible residues at every precision; one finite cyclic quotient is only one approximation level.

Check your understanding

If v_L is normalized by v_L(pi_L)=1 in a ramified finite extension of Q_p with ramification index e, which valuation extends the base v_p?

Try it yourself

In Q3 find the sum of 1+3+3²+... and the exact 3-adic error after retaining terms through 3^(N−1). Explain why the real behavior differs.

Take a moment to make a prediction or work it out on paper. Open a hint when you need one.

Give me a hint

Use the finite geometric-sum formula first, then calculate the valuation of its difference from the proposed limit.

Compare with the solution

The N-term sum is (3^N−1)/2. Its difference from −1/2 is 3^N/2, whose valuation is N because two is a 3-adic unit. The error absolute value is therefore 3^(−N), tending to zero, so the sum is −1/2 in Q3. Over the reals the partial sums grow without bound. The algebraic finite-sum identity is the same; the topology deciding which error tends to zero is different.

Take it with you

  • P-adic smallness means high prime divisibility.
  • Compatible finite residues encode p-adic integer precision.
  • Uniformizer-normalized valuations and base-extending valuations differ by the ramification index.

Read further

Original sources and deeper explanations to follow at your own pace.

Optional progress, saved in this browser when available. Revisit any lesson at any time.

Lesson not marked complete.