如何不使用Wronskian证明对象组是否线性无关？

Got it, let's ditch the Wronskian and go back to the core of linear independence—this is actually way more intuitive once you lean into the fundamental definition. Below are the most reliable methods I use regularly, with concrete examples to make it stick:

1. Lean into the definition of linear independence (the most universal approach) #

At its core, a set of objects ( {f_1, f_2, ..., f_n} ) (whether they're functions, vectors, matrices, etc.) is linearly independent if the only solution to the equation:
c₁f₁ + c₂f₂ + ... + cₙfₙ = 0
(where 0 is the "zero object" for that set—like the zero function, zero vector, etc.) is when all constants ( c₁, c₂, ..., cₙ = 0 ).

Example: Proving ( {1, x, x²} ) is linearly independent over ℝ #

Suppose there exist constants ( c₁, c₂, c₃ ) such that:
c₁ + c₂x + c₃x² = 0 for all real ( x ).

• Plug in ( x = 0 ): We get ( c₁ = 0 ).
• Plug in ( x = 1 ): ( 0 + c₂ + c₃ = 0 ) → ( c₂ = -c₃ ).
• Plug in ( x = -1 ): ( 0 - c₂ + c₃ = 0 ). Substitute ( c₂ = -c₃ ): ( c₃ + c₃ = 0 ) → ( c₃ = 0 ), so ( c₂ = 0 ).

All constants are zero, so the set is linearly independent. Alternatively, for polynomials, we can use the fact that a non-zero polynomial can have at most ( n ) roots—since this polynomial equals zero for all ( x ), all coefficients must be zero directly.

2. Use linear transformations to map to a simpler set #

If you can find a linear transformation ( T ) such that ( {T(f₁), T(f₂), ..., T(fₙ)} ) is linearly independent, then the original set is also linearly independent (linear transformations preserve linear independence in this direction).

Example: Proving ( {e^x, xe^x, x²e^x} ) is linearly independent #

Take the linear transformation ( T(f) = f/e^x ) (this works because ( e^x ) is never zero). Applying ( T ) to each element gives ( {1, x, x²} )—which we already proved is linearly independent. Therefore, the original set must also be linearly independent.

Another trick for differentiable functions: evaluate derivatives at a specific point to get a set of vectors. For the same set above, evaluate at ( x=0 ):

• ( f₁(0) = 1 ), ( f₁'(0) = 1 ), ( f₁''(0) = 1 ) → vector ( (1,1,1) )
• ( f₂(0) = 0 ), ( f₂'(0) = 1 ), ( f₂''(0) = 2 ) → vector ( (0,1,2) )
• ( f₃(0) = 0 ), ( f₃'(0) = 0 ), ( f₃''(0) = 2 ) → vector ( (0,0,2) )

These three vectors are linearly independent (they form an upper triangular matrix with non-zero diagonal entries), so the original function set is linearly independent.

3. Proof by contradiction #

Assume the set is linearly dependent, then show this leads to an impossible conclusion.

Example: Proving ( {sin(x), cos(x)} ) is linearly independent #

Suppose there exist non-zero constants ( c₁, c₂ ) such that c₁sin(x) + c₂cos(x) = 0 for all ( x ). If ( c₁ ≠ 0 ), we can rearrange to get ( tan(x) = -c₂/c₁ )—a constant value for all ( x ). But ( tan(x) ) clearly isn't constant; it oscillates between ( -∞ ) and ( ∞ ) as ( x ) changes. This contradiction means our initial assumption of linear dependence is wrong.

Quick note for vector/matrix sets #

For finite-dimensional vectors, arrange them into a matrix and compute its rank. If the rank equals the number of vectors, they're linearly independent. For example, vectors ( (1,0,0) ), ( (0,1,0) ), ( (0,0,1) ) form an identity matrix with rank 3—so they're independent.

Hope these methods click for you! The Wronskian is just a convenient shortcut for differentiable functions, but going back to basics often clears up confusion when you can't use it.

内容的提问来源于stack exchange，提问作者Stuy