A class XI student stares at the board. Her teacher has just written \sqrt{-1} = i, matter-of-factly, as if this were an ordinary thing to do. She raises her hand. "But no real number squared is negative. So \sqrt{-1} doesn't exist. You just made up the letter i to pretend it does. How is that allowed?"
It is a great question, and the honest answer is: yes, we invented i. But we invented it for the same reason we invented zero, the same reason we invented negative numbers, and the same reason we invented fractions. Each time, mathematics hit a wall inside an old number system, and each time the solution was to enlarge the system until the wall disappeared. The imaginary unit is just the latest enlargement — not a dodge, but a genuine piece of mathematics that does real work.
The pattern: every number system was once invented
To feel why i is legitimate, step back and notice how every kind of number you already accept was once "invented to make arithmetic work."
Natural numbers (1, 2, 3, …). Counting pebbles and goats. These are "natural" in the sense that you could not live without them; no one debates whether 3 is real.
Zero. Natural numbers alone cannot answer "what is 3 - 3?" Subtraction breaks out of the natural numbers. So we invented zero — a number for "nothing" — and made subtraction closed on \{0, 1, 2, \ldots\}. For centuries, European mathematicians debated whether zero was a "real" number. Today it seems obvious. But it was invented.
Negative numbers. Even with zero, you cannot answer "3 - 5." So we invented the negative integers. Again, there were centuries of pushback — Indian mathematicians like Brahmagupta accepted negatives by 628 CE, but European mathematics took until the 1500s to grudgingly admit them. They were called "false numbers" as late as the 1700s.
Fractions. Integers alone cannot answer "3 \div 5." So we invented fractions — and made division (by non-zero things) closed on \mathbb{Q}.
Irrational numbers. Rationals cannot answer "what is \sqrt{2}?" So we invented the irrationals. The Pythagoreans, famously, had an emotional crisis when they discovered \sqrt{2} was not a ratio of integers; legend has it they drowned the discoverer. But we got over it and enlarged the system.
Imaginary numbers. Real numbers cannot answer "what is \sqrt{-1}?" So we invented the imaginaries, and then the complex numbers a + bi.
Every step is the same move: find an operation that breaks out of the current system; invent new numbers to close the gap. There is nothing philosophically special about i — it is the next brick in the same wall.
Why the extension is forced by the algebra
You might object: "We invented zero because counting needed it. What does i solve?"
Quadratic equations. Consider x^2 + 1 = 0. Solving gives x^2 = -1, so x = \pm\sqrt{-1}. Inside the real numbers, this equation has no solution — not because of a technicality, but because the graph y = x^2 + 1 never touches the x-axis. The real numbers cannot supply an answer.
But when you allow i, the equation has two perfectly clean solutions, x = i and x = -i, and the quadratic formula works uniformly for every quadratic, no exceptions. More generally, the fundamental theorem of algebra (a nineteenth-century result) says that every polynomial equation of degree n has exactly n roots, counted with multiplicity, when you allow complex numbers. With only the real numbers, some polynomials have fewer than n roots; with the complex numbers, the count is always exact.
This is a huge payoff. A whole universe of irregular, exception-ridden behavior inside the reals becomes regular and uniform in the complex numbers. The imaginary unit is not a dodge; it is the key that unlocks a cleaner theory.
Are imaginary numbers "real"? (A quibble about the word real.)
Here is where the question gets linguistic. The word real in "real numbers" is unfortunate — it makes it sound like other numbers are unreal. But "real" in mathematics just means "a number on the number line," and "imaginary" means "a number involving the i axis perpendicular to the number line." Both words are historical labels; neither is a claim about what exists.
A better way to think about it: both real and imaginary numbers are ideas. We use the real numbers to model physical length, time, temperature, velocity. We use the complex numbers to model rotation, AC circuit impedance, wave amplitudes in quantum mechanics, and the stability of aircraft wings. Whether either kind of number "physically exists" in the way a rock exists is a question for philosophers. What both kinds of number do exist for is computation — and by that measure, i is as real as \pi.
The physicist's vote
The strongest practical argument for the legitimacy of i: it shows up everywhere in physics.
- Alternating-current circuit analysis uses complex impedance. Every electrical engineer in India computes with i (or j, as engineers call it to avoid confusion with current) every working day. Your phone, your laptop, the power grid that runs Delhi — all designed using i.
- Quantum mechanics is fundamentally complex-valued. The wave function \psi takes complex values, and the Schrödinger equation has an explicit i in it. Without i, there is no quantum theory — and without quantum theory, there is no semiconductor physics, no transistors, no modern electronics.
- Signal processing uses the Fourier transform, which is naturally complex-valued. Every time you stream a video or make a phone call, a complex-number computation is happening.
None of this proves that i is more than a convenient fiction — but if i were a pure fiction, you would expect physical predictions built on it to fail. They do not. They match experiment to many decimal places. Whatever i is, it is not an arbitrary decoration.
A quick visualisation of where i lives
Real numbers sit on a line — one dimension. Complex numbers sit on a plane — two dimensions. The real axis is horizontal; the imaginary axis is vertical. A complex number like 3 + 2i is a point at the coordinates (3, 2) in that plane.
Multiplication by i in this plane is a 90-degree rotation around the origin. Start at 1 on the real axis; multiply by i to land at i on the imaginary axis; multiply by i again to land at -1 on the real axis; again to reach -i; and one more time back to 1. Four multiplications by i, one full rotation. This is why i^2 = -1 is geometrically the same as "rotate by 90 degrees twice equals a half-turn."
"So we just defined it into existence?"
Yes, and that is how all of mathematics works. Zero was defined into existence. Negative numbers were defined into existence. Irrationals were defined into existence — \pi is defined as "the ratio of a circle's circumference to its diameter," and \sqrt{2} is defined as "the positive number whose square is 2." Each of these is a definition, a stipulation, a "let us agree to call this a number."
Imaginary numbers are built the same way. We define i by the rule i^2 = -1, and then we check that arithmetic with i is self-consistent — that you can add, subtract, multiply, and divide complex numbers without running into contradictions. It is self-consistent. So the complex numbers are mathematically as legitimate as the real numbers.
The only real difference is that you cannot use i to measure a physical length, because lengths are inherently non-negative real numbers. But you can use i to encode rotation, phase, oscillation, and many other things that do show up physically. Different numbers, different jobs.
The takeaway
\sqrt{-1} has no answer inside the real numbers. We invented an answer, called it i, and checked that the arithmetic still works. The invention is legitimate for exactly the same reasons zero, negatives, and irrationals are legitimate — each one filled a gap the previous system could not fill, and each one produced a richer, cleaner theory in the process. i is made up, in the same sense that every number is made up. That does not make it any less useful, or any less beautiful.
Related: Roots and Radicals · √(−1): Imaginary or Undefined? Both, Depending on Where You Stand · Why We Need Irrational Numbers · Number Systems