Example: strictly upper triangular matrices

Strictly upper triangular matrices form a nilpotent Lie algebra under commutator; commutators move entries further above the diagonal.
Example: strictly upper triangular matrices

Let nngln(R)\mathfrak n_n\subset \mathfrak{gl}_n(\mathbb R) be the vector space of strictly upper triangular matrices (zeros on and below the diagonal), with Lie bracket the commutator [A,B]=ABBA[A,B]=AB-BA.

This is a standard example of a .

Concrete bracket computation in the matrix-unit basis

For i<ji<j, let EijE_{ij} be the matrix unit. A direct multiplication gives

EijEkl=δjkEil, E_{ij}E_{kl}=\delta_{jk}E_{il},

so

[Eij,Ekl]=EijEklEklEij=δjkEilδliEkj. [E_{ij},E_{kl}] = E_{ij}E_{kl}-E_{kl}E_{ij} = \delta_{jk}E_{il}-\delta_{li}E_{kj}.

In particular, if i<ji<j and k<lk<l, then [Eij,Ekl][E_{ij},E_{kl}] is either 00 or another strictly upper triangular matrix unit.

Lower central series (explicit nilpotency mechanism)

Define the “height” of EijE_{ij} as ji1j-i\ge 1. From the formula above, any nonzero commutator of strictly upper triangular matrices increases height: the product EijEjk=EikE_{ij}E_{jk}=E_{ik} has height (ki)=(ji)+(kj)(k-i)=(j-i)+(k-j).

Consequently, iterated commutators eventually vanish: the

γ1=nn,γr+1=[nn,γr] \gamma_1=\mathfrak n_n,\qquad \gamma_{r+1}=[\mathfrak n_n,\gamma_r]

satisfies γn=0\gamma_{n}=0. Thus nn\mathfrak n_n is nilpotent of class at most n1n-1, hence solvable (compare and ).

Small case n=3n=3 (fully explicit)

A general element is

(0ab00c000)=aE12+bE13+cE23. \begin{pmatrix} 0 & a & b\\ 0 & 0 & c\\ 0 & 0 & 0 \end{pmatrix} = aE_{12}+bE_{13}+cE_{23}.

Using [E12,E23]=E13[E_{12},E_{23}]=E_{13} and other brackets 00, we get

[n3,n3]=span{E13},[n3,[n3,n3]]=0, [\mathfrak n_3,\mathfrak n_3]=\mathrm{span}\{E_{13}\}, \qquad [\mathfrak n_3,[\mathfrak n_3,\mathfrak n_3]]=0,

which matches the pattern.