Lie Bracket of two vector fields

1. The definition of a Lie bracket.

Let M be an n-manifold. Recall that if X and Y are smooth vector fields
on M then X and Y are 1-st order differential operators on smooth functions
M → R. Thus, for a smooth function f : M → R, Xf and Y f are again
smooth functions M → R.
We now define a differential operator [X, Y ], called the Lie bracket or the
commutator of X and Y as
[X, Y ] := XY − Y X,
that is, for a smooth function f : M → R,
[X, Y ](f ) := X(Y f ) = Y (Xf ).
By its definition, [X, Y ] is a 2-nd order differential operator. However, by
a kind of a miracle, it turns out that [X, Y ] is actually a 1-st order differential
operator and, in fact, a vector field:
Theorem 1. For any smooth vector fields X and Y on an n-manifold
M , the Lie bracket [X, Y ] is again a smooth vector field on M .
We will not give a detailed proof of Theorem 1 but instead will only verify
that Theorem 1 holds for the basic vector fields.
Let φ = (x1 , . . . , xn ) : U → Rn be a coordinate patch on M . Let X =
∂ n ∂
a(x1 , . . . , xn ) ∂x 1
i and Y = b(x , . . . , x ) ∂xj for some 1 ≤ i, j ≤ n and for

some smooth functions a = a(x , . . . , xn ) and b = b(x1 , . . . , xn ).

1

Let f : M → R be a smooth function. We now compute [X, Y ](f ). By

definition:
∂ ∂ ∂ ∂f ∂ ∂
[X, Y ](f ) = [a i , b j ]f = a i (b j ) − b j (a i ) =
∂x ∂x ∂x ∂x ∂x ∂x
using the product rule
∂b ∂f ∂2f ∂a ∂f ∂2f
a + ab − b − ba =
∂xi ∂xj ∂xi ∂xj ∂xj ∂xi ∂xj ∂xi
∂b ∂f ∂a ∂f ∂b ∂ ∂a ∂
a i j − b j i = (a i j − b j i )f
∂x ∂x ∂x ∂x ∂x ∂x ∂x ∂x
Therefore
∂ ∂ ∂b ∂ ∂a ∂
(†) [X, Y ] = [a i
,b j] = a i j − b j i.
∂x ∂x ∂x ∂x ∂x ∂x
∂b ∂a
Note that if a = b ≡ 1 then ∂xi = ∂xj = 0. Therefore
∂ ∂
, ] = 0.
∂xi ∂xj
2. General properties of the Lie bracket
We have:
(1) [X, Y ] = −[Y, X].
(2) [X1 + X2 , Y ] = [X1 , Y ] + [X2 , Y ] and [X, Y1 + Y2 ] = [X, Y1 ] + [X, Y2 ].
1
2

(3) For any smooth functions a, b : M → R

[aX, bY ] = ab[X, Y ] + a(Xb)Y − b(Y a)X.
(4) [“Jacobi Identity”]
[[X, Y ], Z] + [[Y, Z], X] + [[Z, X], Y ] = 0.
Note that usingP property (2) above together with formula (†), it is not
n ∂ Pn
hard to compute [ i=1 ai (x) ∂xi , j=1 bj (x) ∂x∂ j ] in a coordinate chart φ =
(x1 , . . . , xn ) : U → Rn on M :

n n
X
i ∂ X j ∂ XX ∂ ∂
[ a (x) i , b (x) j ] = [ai (x) i , bj (x) j ] =
∂x ∂x ∂x ∂x
i=1 j=1 i j
XX ∂bj ∂ i
j ∂a ∂
(‡) ai − b
∂xi ∂xj ∂xj ∂xi
i j

However, as a practical matter, it is better to remember formula (†) to-

gether with properties (1),(2),(3),(4) above, rather than to try to memorize
formula (‡).
∂
Example. Let M = R3 with coordinates (x, y, z) and let X = 2xz ∂x +
yz ∂ 2 2 2 ∂ ∂
e ∂y and Y = (x + y + z ) ∂x + 5 ∂z .
We want to compute [X, Y ]. We have
[X, Y ] =
∂ ∂ ∂ ∂ ∂ ∂ ∂ ∂
[2xz , (x2 + y 2 + z 2 ) ] + [2xz , 5 ] + [eyz , (x2 + y 2 + z 2 ) ] + [eyz , 5 ].
∂x ∂x ∂x ∂z ∂y ∂x ∂y ∂z
Using (‡), we get
∂ ∂ ∂ ∂
[2xz , (x2 + y 2 + z 2 ) ] = 2xz2x − (x2 + y 2 + z 2 )2z =
∂x ∂x ∂x ∂x
∂
(2x2 z − 2y 2 z − 2z 3 ) ,
∂x
and
∂ ∂ ∂ ∂
[2xz , 5 ] = 0 − 10x = −10x ,
∂x ∂z ∂x ∂x
and
∂ ∂ ∂ ∂ ∂
[eyz , (x2 + y 2 + z 2 ) ] = eyz 2y − (x2 + y 2 + z 2 ) · 0 = eyz 2y ,
∂y ∂x ∂x ∂y ∂x
and
∂ ∂ ∂ ∂ ∂
[eyz , 5 ] = eyz · 0 − 5yeyz = −5yeyz .
∂y ∂z ∂z ∂y ∂y
Therefore

∂ ∂
[X, Y ] = 2x2 z − 2y 2 z − 2z 3 − 10x + 2yeyz − 5yeyz .
∂x ∂y