You can type (-2)^3 into a calculator and get -8, no problem. You can type (-2)^{1/3} on many calculators and get -\sqrt[3]{2} \approx -1.26. But type (-2)^{1/2} and the calculator either beeps, returns "MATH ERROR," or — on a scientific calculator that handles complex numbers — gives you something like 0 + 1.414i.

Why does a fractional exponent go weird the moment the base is negative? The answer is that the innocent-looking extension from integer exponents to fractional ones was only half-honest. It worked perfectly for positive bases. For negative bases it runs into the fact that negative numbers don't have real square roots, and the whole structure wobbles.

Start with the rule that worked for positive bases

For positive a, the rule a^{m/n} = \sqrt[n]{a^m} is cleanly defined. Every step is a real number. 8^{2/3} = \sqrt[3]{64} = 4. 16^{3/4} = \sqrt[4]{4096} = 8. No drama.

The rule relied on two things you took for granted:

  1. Every positive real number a has a unique positive real n-th root.
  2. The laws of exponents continue to hold.

For positive bases, both assumptions are true, and the rule sails through.

The problem arrives with negative bases

Try a = -4, m = 1, n = 2. The rule says (-4)^{1/2} = \sqrt{-4}. But \sqrt{-4} is not a real number — no real number squared gives -4, because the square of any real number is non-negative.

So the rule has broken. (-4)^{1/2} has no real answer. If you insist on an answer, you must leave the real numbers and enter the complex numbers, where \sqrt{-4} = 2i (with i = \sqrt{-1}).

This is not a bug in your calculator — it is a genuine gap in the real-number system. The moment you raise a negative base to an exponent with an even denominator, you are outside the real line.

The divide: even vs odd denominators

The rule (-a)^{m/n} does work for negative bases when the denominator n is odd — because odd roots of negative numbers are real.

Why odd roots work: \sqrt[n]{-a} = -\sqrt[n]{a} when n is odd, because (-\sqrt[n]{a})^n = (-1)^n (\sqrt[n]{a})^n = -a. The (-1)^n is -1 when n is odd, preserving the negative sign through the root.

But for even denominators, the rule breaks:

So when someone writes (-a)^{m/n}, you need to inspect the denominator n:

The second problem: multiple values

It gets worse. Even when an answer does exist, fractional exponents of negative numbers are often multi-valued.

Consider (-1)^{1/3}. By one reading, this is \sqrt[3]{-1} = -1 — the real cube root. But -1 has three cube roots in the complex numbers: -1, and two complex conjugates. Depending on which branch of the cube-root function your calculator uses, you might get the real answer -1, or you might get one of the two complex answers \tfrac{1}{2} + \tfrac{\sqrt{3}}{2}i or \tfrac{1}{2} - \tfrac{\sqrt{3}}{2}i.

This is why some calculators disagree with each other on (-1)^{1/3}: Wolfram Alpha and Python's pow(-1, 1/3) give the complex value, but a scientific calculator or MATLAB's real-cube-root gives -1. Both are defensible. The ambiguity comes from the fact that a single symbol like (-a)^{m/n} does not, by itself, pin down which root to take.

What to actually do in school-level problems

For JEE / Class 9–12, the convention is simple and safe:

The rule (a)^{m/n} = \sqrt[n]{a^m} is defined only for a \geq 0.

If the base is negative, you do not apply the fractional-exponent rule. Instead:

  1. If the denominator n is odd, you can compute the n-th root directly: (-a)^{m/n} = -\sqrt[n]{a^m} (with appropriate sign care).
  2. If the denominator n is even, the expression is undefined in the real numbers.
  3. If the question explicitly invokes complex numbers, use (-a)^{1/2} = i\sqrt{a} and proceed from there.

In NCERT textbooks and JEE Main, negative-base fractional exponents are avoided except in specifically-marked complex-number problems.

The laws of exponents break here too

A subtle warning: the familiar laws of exponents stop working when you mix negative bases and fractional exponents carelessly.

Watch this apparent paradox:

-1 \;=\; (-1)^1 \;=\; (-1)^{2 \cdot 1/2} \;\stackrel{?}{=}\; \left((-1)^2\right)^{1/2} \;=\; 1^{1/2} \;=\; 1

So -1 = 1? Obviously not. The flaw is in the "=" marked with the question mark — the power of a power law (a^m)^n = a^{mn} silently fails when a < 0 and n is not an integer. The law was derived assuming a > 0. Applying it to a = -1 gave a false identity.

The takeaway is deep: laws derived under one set of conditions do not automatically extend. The laws of exponents work for all real exponents only when the base is positive. For negative bases, you must restrict to integer exponents (or odd-denominator fractions), or you must move to complex numbers where a proper theory of multi-valued exponentiation replaces the naive rules.

A quick decision chart

When you meet (a)^{m/n} with a possibly negative:

  1. Is a \geq 0? → All the usual rules apply. Use \sqrt[n]{a^m}.
  2. Is a < 0 and is n odd? → Apply the real n-th root, carrying the sign through. Check that the question is actually well-posed (no hidden multi-valuedness).
  3. Is a < 0 and is n even? → In real arithmetic, undefined. In complex arithmetic, write a = |a| \cdot e^{i\pi} and apply the polar-form rules.
  4. Is a < 0 and is the exponent irrational (like (-2)^{\sqrt{2}})? → Necessarily complex. You need complex exponentiation via a^x = e^{x \ln a} with a chosen branch of the complex logarithm. This is a calculus-II topic.

Why this is actually a good lesson

Your discomfort with (-2)^{1/2} is pointing at something important: the real number line is a more restricted structure than it looks, and fractional exponents are the first topic in school mathematics where that restriction becomes visible.

Most of the mathematics you learn for JEE — polynomials, roots, trigonometry, calculus — works cleanly because everyone has tacitly agreed to stay within the positive reals, or within cases where all the quantities behave. The moment a fractional exponent lands on a negative base, the structure shows its cracks, and you meet the complex numbers as the natural fix. Every physicist and every engineer eventually has to make this leap, because e^{i\theta} and related constructions are central to circuit analysis, wave mechanics, and signal processing. The discomfort you feel now is the first taste of why complex numbers eventually become indispensable — not a curiosity, but a necessity.

For now, in the world of school exponents, memorise the rule: fractional exponents on negative bases are trouble; avoid them or use odd-denominator roots.

Related: Exponents and Powers · Roots and Radicals · Complex Numbers Introduction · What Does a Fractional Exponent Like 2^{1/2} Actually Mean?