The Calculus Paradox: Why ∫1/x dx Can Equal 1 = 0

The paradox where integration seems to prove 1 = 0 arises from treating an indefinite integral as a single function instead of a set of antiderivatives. Once you interpret the integral as an equivalence class of functions differing by a constant, and recognize that a disconnected domain (like ℝ minus zero) allows a separate constant on each piece, the contradiction disappears.

Key Takeaways

• The domain of 1/x is ℝ\{0}, which has two separate connected pieces (x > 0 and x < 0), so its antiderivative can carry two independent constants rather than one.
• The formula “if f’ = g’ then f = g + C” only holds when the domain is a single connected interval, because its proof relies on the mean value theorem.
• An indefinite integral is properly defined as the full set of functions sharing a given derivative, not one specific function plus a constant.
• The manipulation that appears to show 0 = 1 actually shows [0] = [1] as equivalence classes of functions, which is a true and unremarkable statement.
• Equivalence classes obey clean algebraic rules such as [f] + [g] = [f + g] and α[f] = [αf], which become confusing when written in the informal “f + C” notation.

Why Does the Antiderivative of 1/x Cause a Paradox?

The standard calculus result states:

$$\int \frac{1}{x} dx = \log |x|+C$$

This is taught in nearly every calculus course, but the usual interpretation is misleading. Students are typically told that every function whose derivative equals 1/x must be log|x| plus a single constant. That claim is false. The following function has derivative 1/x everywhere on its domain but does not fit the single-constant form:

$$f(x) = \left\{\begin{array}{ll} \log x & \text{if}~x>0\\ 1+\log(-x) & \text{if}~x<0\end{array}\right.$$

More generally, every function whose derivative is 1/x on ℝ\{0} has the form:

$$f(x) = \left\{\begin{array}{ll} \log(x) + C & \text{if}~x>0\\ \log(-x) +C’& \text{if}~x<0\end{array}\right.$$

There can be two independent arbitrary constants, one for each connected component of the domain, rather than a single shared constant.

Why Does the Standard f = g + C Theorem Fail for 1/x?

The familiar theorem “if f’ = g’ then f = g + C” depends on a hypothesis that is easy to overlook. The result follows directly from the mean value theorem, which is proven for a function f mapping a closed interval [a, b] to the real numbers, where f is continuous on [a, b] and differentiable on the open interval (a, b).

The natural domain of 1/x is ℝ\{0}, which is not a single interval since it splits into two disconnected pieces. Because of this, the formula:

$$\int f'(x)dx = f(x)+C$$

is only guaranteed to hold when the domain is a single interval, or more generally a connected set. On a disconnected domain, each connected component may carry its own separate integration constant.

How Does Integration by Parts Seem to Prove 1 = 0?

Consider the following formal manipulation:

$$\int \frac{1}{x} dx = 1 + \int \frac{1}{x} dx$$

Canceling the two integral expressions on both sides appears to leave 1 = 0. Resolving this requires returning to the precise definition of what an indefinite integral actually represents.

What Is an Indefinite Integral, Really?

An indefinite integral is not a single function. It is the complete set of all antiderivatives of a given function:

$$\int f(x)\,dx = \{F \mid F’ = f\}$$

The common notation “f(x) + C” should be understood as shorthand for an equivalence class of functions. Define an equivalence relation on functions by f ~ g if and only if f’ = g’. The equivalence class of f under this relation is:

$$[f] = \{g \mid f \sim g\}.$$

An indefinite integral is properly an equivalence class of functions, one that is habitually and imprecisely written as f + C.

How Do Equivalence Classes of Functions Behave Algebraically?

The operations on equivalence classes are natural and well-behaved:

• $$[f]+[g] = [f+g]$$
• $$\alpha [f] = [\alpha f]$$
• $$[\alpha] = [0]$$
• $$0[f] = [0]$$

with α a real number.

The same rules written in the informal “f + C” notation look considerably more awkward and are a common source of confusion:

• $$(f+C) + (g+C) = (f+g)+C$$
• $$\alpha (f+C) = (\alpha f) + C$$
• $$\alpha + C = C$$
• $$0(f+C) = C$$

How Does the Equivalence Class View Resolve the 1 = 0 Paradox?

Using the equivalence-class interpretation:

$$\int \frac{1}{x}dx = \log|x| + C = [\log|x|]$$

Therefore:

$$[\log|x|] = [1 + \log|x|] = [1]+[\log|x|].$$

The first step of the integration-by-parts manipulation in the paradox is valid, but it must be read as a statement about equivalence classes:

$$\int fg’ = [fg] + \int f’ g,$$

where fg denotes an equivalence class rather than one specific function. From this we get:

$$[\log|x|] – [\log|x|]= [1]+[\log|x|] – [\log|x|]$$

which implies:

$$[0] = [1].$$

This equality of equivalence classes is entirely correct. It does not mean that 0 equals 1 as functions. Two different functions can represent the same equivalence class once you factor out derivatives.

A Related Exercise to Try

Here is a related manipulative paradox worth working through:

$$0 = 0\int xdx = \int 0x dx = \int 0 dx = C$$

Try resolving this using the same equivalence-class formalism described above. The symbol C in this expression represents an arbitrary constant, or equivalence class, not a single fixed number.

Conclusion

An indefinite integral denotes a class of antiderivatives rather than one specific function. When the domain is disconnected, such as ℝ\{0}, each connected component may carry its own independent integration constant. Interpreting integrals and their algebraic manipulations at the level of equivalence classes, rather than as pointwise equalities between particular representative functions, removes both paradoxes discussed here.

Frequently Asked Questions

Why does 1/x have two integration constants instead of one?

Because the domain of 1/x, which is ℝ\{0}, splits into two separate connected pieces: all x greater than 0 and all x less than 0. The theorem that guarantees a single shared constant only applies on a connected domain, so each of the two pieces here can carry its own independent constant.

Is the equation [0] = [1] actually true?

Yes, but only as a statement about equivalence classes of functions under the relation f ~ g when f’ = g’. It does not mean the numbers 0 and 1 are equal. Two functions differing by a constant belong to the same equivalence class, which is what this notation expresses.

Where does the “f = g + C” rule come from?

It follows from the mean value theorem, which requires a function to be continuous on a closed interval and differentiable on the corresponding open interval. That proof assumes the domain is a single interval, which is why the rule can fail on domains with multiple disconnected pieces.

What is an equivalence class in this context?

An equivalence class groups together all functions that share the same derivative. Instead of treating “f(x) + C” as one function with an unspecified constant, it is more precise to treat it as the entire set of functions whose derivative matches f.

Does this paradox mean calculus is broken?

No. The paradox arises from an imprecise but widely taught shorthand, not from an actual flaw in calculus. Once the indefinite integral is treated correctly as a set of antiderivatives rather than a single function, the apparent contradiction disappears entirely.

For a related discussion, see Friends, Strangers, and Computers.