Why the Form of an Identity Matters
A trigonometric expression can often be simplified by replacing a group of terms with an equal expression. The Pythagorean identities provide several such replacements. The useful choice depends on which functions appear: an expression involving sine and cosine may be handled directly, while one involving tangent and secant may call for a different form.
The goal is not to change an expression at points where it is undefined. A simplified formula may be defined at more inputs than the original, so restrictions from denominators must be recorded as part of the answer.
The Pythagorean Identities
For every real angle \(x\), sine and cosine satisfy the basic Pythagorean identity. The other two forms follow by dividing this identity by a square, but each division requires that its divisor be nonzero.
The conditions on the last two identities are essential. Tangent and secant are defined only when \(\cos x\ne 0\); cotangent and cosecant are defined only when \(\sin x\ne 0\). These are also the conditions required to divide by \(\cos^2 x\) or \(\sin^2 x\).
Why the Three Forms Are Equivalent
The unit-circle identity \(\sin^2 x+\cos^2 x=1\) is the starting point. To obtain a form involving tangent and secant, divide each term by \(\cos^2 x\), which is permitted only when \(\cos x\ne 0\):
Using \(\tan x=\frac{\sin x}{\cos x}\) and \(\sec x=\frac{1}{\cos x}\), this becomes \( \tan^2 x+1=\sec^2 x\). The corresponding derivation for cotangent and cosecant divides by \(\sin^2 x\), so it requires \(\sin x\ne 0\):
This explains both why the identities are true and why their domain conditions matter. A division that is valid only under a condition cannot justify an identity at inputs that violate that condition.
A Practical Simplification Method
When simplifying, first notice which functions and combinations are present. Match them to an identity, substitute an equal expression, and then use ordinary algebra. Before cancelling a factor or dividing by an expression, check whether it can be zero in the original expression.
Find inputs that make an original denominator zero or a trigonometric function undefined.
Use the identity whose functions and structure already appear in the expression.
Show the algebraic steps, preserving the original restrictions.
Confirm the replacement using the identity and, when useful, evaluate both expressions at an allowed input.
Worked Examples
Worked Example: Replace a Difference of Squares
Simplify \(\frac{1-\cos^2 x}{\sin x}\), and state its domain.
The denominator in the original expression requires \(\sin x\ne 0\). The sine-cosine identity gives \(1-\cos^2 x=\sin^2 x\). Therefore,
The simplified formula is \(\sin x\), but the original expression is still restricted to inputs where \(\sin x\ne 0\). To check the algebra on that domain, replace \(\sin^2 x\) in the final quotient by \(1-\cos^2 x\); the quotient is exactly the original expression. As a numerical check, at \(x=\frac{\pi}{6}\), \(\sin x=\frac12\) and \(\cos x=\frac{\sqrt3}{2}\), so
Worked Example: Use the Tangent-Secant Form
Simplify \(\sec^2 x-\tan^2 x\), and state where the original expression is defined.
Both secant and tangent require \(\cos x\ne 0\). Rearranging \(1+\tan^2 x=\sec^2 x\) gives \(\sec^2 x-\tan^2 x=1\). Thus,
The identity verifies the simplification at every input where the expression is defined. For a numerical check, take \(x=\frac{\pi}{4}\), where \(\cos x\ne0\), \(\sec^2 x=2\), and \(\tan^2 x=1\). The original expression evaluates to \(2-1=1\), matching the simplified value. The restriction \(\cos x\ne0\) remains part of the result even though the constant \(1\) by itself is defined for every real \(x\).
Worked Example: Use the Cotangent-Cosecant Form
Simplify \(\frac{1+\cot^2 x}{\csc x}\), and state its domain.
The numerator uses the cotangent-cosecant identity. Both \(\cot x\) and \(\csc x\) require \(\sin x\ne0\); also, the denominator \(\csc x\) is never zero where it is defined. Substitute \(1+\cot^2 x=\csc^2 x\), then cancel one factor of \(\csc x\):
To check, multiply the simplified value by the original denominator: \(\csc x\cdot\csc x=\csc^2 x=1+\cot^2 x\), as required. For \(x=\frac{\pi}{6}\), \(\cot x=\sqrt3\) and \(\csc x=2\), so the original expression is \(\frac{1+3}{2}=2\), equal to the simplified value \(2\).
Worked Example: Factor Before Cancelling
Simplify \(\frac{\sec^2 x-1}{\sec x+1}\), taking care with the domain.
The original expression requires \(\cos x\ne0\), so that secant is defined. It also requires \(\sec x+1\ne0\). The Pythagorean identity gives \(\sec^2 x-1=\tan^2 x\), while factoring the numerator as a difference of squares makes the cancellation visible:
Cancellation is valid only when \(\sec x+1\ne0\), which is already required by the original denominator. The simplified formula alone would also be defined when \(\sec x=-1\), but the original expression is not. As a check at \(x=\frac{\pi}{3}\), \(\sec x=2\), so the original value is \(\frac{4-1}{2+1}=1\); the simplified value is \(2-1=1\).
Common Pitfall: Losing a Domain Restriction
A common error is to simplify correctly and then claim that the new formula is valid everywhere. For instance, one might write
The cancellation in the second step divides by \(\sin x\). It is valid only if \(\sin x\ne0\). At \(x=0\), the original expression has denominator \(\sin 0=0\) and is undefined, even though the simplified formula gives \(\sin 0=0\). The correct statement is that the expressions agree for inputs satisfying \(\sin x\ne0\), which is the domain of the original expression.
Choosing and Checking a Replacement
A useful choice is usually the identity that directly matches the terms. Use \(\sin^2 x+\cos^2 x=1\) when the expression contains sine and cosine; use \(1+\tan^2 x=\sec^2 x\) when it contains tangent and secant; and use \(1+\cot^2 x=\csc^2 x\) when it contains cotangent and cosecant. Rearranging an identity is allowed, provided its original conditions are respected.
For example, the tangent-secant identity can be rearranged as \(\sec^2 x-1=\tan^2 x\), but this form is valid only where \(\cos x\ne0\). Likewise, \(1-\sin^2 x=\cos^2 x\) follows from the sine-cosine identity without dividing, so it holds for all real \(x\). Distinguishing these cases helps prevent accidental extensions beyond the original domain.
A final check should answer two questions: does the substitution follow from an identity under the stated conditions, and have all restrictions from the original denominators been retained? A test at a valid angle can catch arithmetic mistakes, but the identity and the domain conditions justify the simplification generally.
Check Your Understanding
For each expression, simplify it and state any restrictions inherited from the original expression.
- Simplify \(\frac{\sin^2 x+\cos^2 x}{\cos x}\).
- Rewrite \(\csc^2 x-1\) using cotangent, and state the condition needed.
- Simplify \(\frac{\sec^2 x}{\sec x}\), stating where the original expression is defined.
- Why must \(\sin x\ne0\) be retained when simplifying \(\frac{1-\cos^2 x}{\sin x}\)?
- Explain why \(1+\cot^2 x=\csc^2 x\) cannot be used at an input where \(\sin x=0\).
