Lie algebra extension

In the theory of Lie groups, Lie algebras and their representation theory, a Lie algebra extension is an enlargement of a given Lie algebra by another Lie algebra . Extensions arise in several ways. There is the trivial extension obtained by taking a direct sum of two Lie algebras. Other types are the split extension and the central extension. Extensions may arise naturally, for instance, when forming a Lie algebra from projective group representations. Such a Lie algebra will contain central charges.

Starting with a polynomial loop algebra over finite-dimensional simple Lie algebra and performing two extensions, a central extension and an extension by a derivation, one obtains a Lie algebra which is isomorphic with an untwisted affine Kac–Moody algebra. Using the centrally extended loop algebra one may construct a current algebra in two spacetime dimensions. The Virasoro algebra is the universal central extension of the Witt algebra.

Central extensions are needed in physics, because the symmetry group of a quantized system usually is a central extension of the classical symmetry group, and in the same way the corresponding symmetry Lie algebra of the quantum system is, in general, a central extension of the classical symmetry algebra. Kac–Moody algebras have been conjectured to be symmetry groups of a unified superstring theory. The centrally extended Lie algebras play a dominant role in quantum field theory, particularly in conformal field theory, string theory and in M-theory.

A large portion towards the end is devoted to background material for applications of Lie algebra extensions, both in mathematics and in physics, in areas where they are actually useful. A parenthetical link, ( background material), is provided where it might be beneficial.

History

Due to the Lie correspondence, the theory, and consequently the history of Lie algebra extensions, is tightly linked to the theory and history of group extensions. A systematic study of group extensions was performed by the Austrian mathematician Otto Schreier in 1923 in his PhD thesis and later published.
Otto Schreier (1901 - 1929) was a pioneer in the theory of extension of groups. Along with his rich research papers, his lecture notes were posthumously published (edited by Emanuel Sperner) under the name ''Einführung in die analytische Geometrie und Algebra'' (Vol I 1931, Vol II 1935), later in 1951 translated to English i
Introduction to Modern Algebra and Matrix Theory
See for further reference. The problem posed for his thesis by Otto Hölder was "given two groups and , find all groups having a normal subgroup isomorphic to such that the factor group is isomorphic to ".

Lie algebra extensions are most interesting and useful for infinite-dimensional Lie algebras. In 1967, Victor Kac and Robert Moody independently generalized the notion of classical Lie algebras, resulting in a new theory of infinite-dimensional Lie algebras, now called Kac–Moody algebras. They generalize the finite-dimensional simple Lie algebras and can often concretely be constructed as extensions.

Notation and proofs

Notational abuse to be found below includes for the exponential map given an argument, writing for the element in a direct product ( is the identity in ), and analogously for Lie algebra direct sums (where also and are used interchangeably). Likewise for semidirect products and semidirect sums. Canonical injections (both for groups and Lie algebras) are used for implicit identifications. Furthermore, if , , ..., are groups, then the default names for elements of , , ..., are , , ..., and their Lie algebras are , , ... . The default names for elements of , , ..., are , , ... (just like for the groups!), partly to save scarce alphabetical resources but mostly to have a uniform notation.

Lie algebras that are ingredients in an extension will, without comment, be taken to be over the same field.

The summation convention applies, including sometimes when the indices involved are both upstairs or both downstairs.

Caveat: Not all proofs and proof outlines below have universal validity. The main reason is that the Lie algebras are often infinite-dimensional, and then there may or may not be a Lie group corresponding to the Lie algebra. Moreover, even if such a group exists, it may not have the "usual" properties, e.g. the exponential map might not exist, and if it does, it might not have all the "usual" properties. In such cases, it is questionable whether the group should be endowed with the "Lie" qualifier. The literature is not uniform. For the explicit examples, the relevant structures are supposedly in place.

Definition

Lie algebra extensions are formalized in terms of short exact sequences. A short exact sequence is an exact sequence of length three,

such that is a monomorphism, is an epimorphism, and . From these properties of exact sequences, it follows that (the image of) is an ideal in . Moreover,
:$\mathfrak g \cong \mathfrak e/\operatorname i = \mathfrak e/ \operatorname s,$
but it is not necessarily the case that is isomorphic to a subalgebra of . This construction mirrors the analogous constructions in the closely related concept of group extensions.

If the situation in prevails, non-trivially and for Lie algebras over the same field, then one says that is an extension of by .

Properties

The defining property may be reformulated. The Lie algebra is an extension of by if

is exact. Here the zeros on the ends represent the zero Lie algebra (containing the null vector only) and the maps are the obvious ones; maps to and maps all elements of to . With this definition, it follows automatically that is a monomorphism and is an epimorphism.

An extension of by is not necessarily unique. Let denote two extensions and let the primes below have the obvious interpretation. Then, if there exists a Lie algebra isomorphism such that
:$f \circ i = i', \quad s' \circ f = s,$

then the extensions and are said to be equivalent extensions. Equivalence of extensions is an equivalence relation.

Extension types

Trivial

A Lie algebra extension
:$\mathfrak h \; \overset i \hookrightarrow \; \mathfrak t \; \overset s \twoheadrightarrow \; \mathfrak g,$
is trivial if there is a subspace such that and is an ideal in .

Split

A Lie algebra extension
:$\mathfrak h \; \overset i \hookrightarrow \; \mathfrak s \; \overset s \twoheadrightarrow \; \mathfrak g,$
is split if there is a subspace such that as a vector space and is a subalgebra in .

An ideal is a subalgebra, but a subalgebra is not necessarily an ideal. A trivial extension is thus a split extension.

Central

Central extensions of a Lie algebra by an abelian Lie algebra can be obtained with the help of a so-called (nontrivial) 2-cocycle (background) on . Non-trivial 2-cocycles occur in the context of projective representations (background) of Lie groups. This is alluded to further down.

A Lie algebra extension
:$\mathfrak h \; \overset i \hookrightarrow \; \mathfrak c \; \overset s \twoheadrightarrow \; \mathfrak g,$
is a central extension if is contained in the center of .

Properties
*Since the center commutes with everything, in this case is abelian.
*Given a central extension of , one may construct a 2-cocycle on . Suppose is a central extension of by . Let be a linear map from to with the property that , i.e. is a section of . Use this section to define by
:
\epsilon(G_1, G_2) = l( _1, G_2 - (G_1), l(G_2) \quad G_1, G_2 \in \mathfrak g.

The map satisfies
:
\epsilon(G_1, _2, G_3 + \epsilon(G_2, _3, G_1 + \epsilon(G_3, _1, G_2 = 0 \in \mathfrak e.

To see this, use the definition of on the left hand side, then use the linearity of . Use Jacobi identity on to get rid of half of the six terms. Use the definition of again on terms sitting inside three Lie brackets, bilinearity of Lie brackets, and the Jacobi identity on , and then finally use on the three remaining terms that and that so that brackets to zero with everything.
It then follows that satisfies the corresponding relation, and if in addition is one-dimensional, then is a 2-cocycle on (via a trivial correspondence of with the underlying field).

A central extension
:$0 \; \overset \iota \hookrightarrow \mathfrak h \; \overset i \hookrightarrow \; \mathfrak e \; \overset s \twoheadrightarrow \; \mathfrak g \; \overset \sigma \twoheadrightarrow \; 0$
is universal if for every other central extension
:$0 \; \overset \iota \hookrightarrow \mathfrak h' \; \overset \hookrightarrow \; \mathfrak e' \; \overset \twoheadrightarrow \; \mathfrak g \; \overset \sigma \twoheadrightarrow \; 0$
there exist ''unique'' homomorphisms $\Phi : \mathfrak e \to \mathfrak e'$ and $\Psi : \mathfrak h \to \mathfrak h'$ such that the diagram

commutes, i.e. and . By universality, it is easy to conclude that such universal central extensions are unique up to isomorphism.

Construction

By direct sum

Let $\mathfrak g$, $\mathfrak h$ be Lie algebras over the same field $F$. Define
:$\mathfrak e = \mathfrak h \times \mathfrak g,$
and define addition pointwise on $\mathfrak e$. Scalar multiplication is defined by
:$\alpha(H, G) = (\alpha H, \alpha G), \alpha \in F, H \in \mathfrak h, G \in \mathfrak g.$
With these definitions, $\mathfrak h \times \mathfrak g \equiv \mathfrak h \oplus \mathfrak g$ is a vector space over $F$. With the Lie bracket:

$\mathfrak e$ is a Lie algebra. Define further
:$i:\mathfrak h \hookrightarrow \mathfrak e; H \mapsto (H, 0), \quad s:\mathfrak e \twoheadrightarrow \mathfrak g; (H, G) \mapsto G.$
It is clear that holds as an exact sequence. This extension of $\mathfrak g$ by $\mathfrak h$ is called a trivial extension. It is, of course, nothing else than the Lie algebra direct sum. By symmetry of definitions, $\mathfrak e$ is an extension of $\mathfrak h$ by $\mathfrak g$ as well, but $\mathfrak h \oplus \mathfrak g \neq \mathfrak g \oplus \mathfrak h$. It is clear from that the subalgebra $0 \oplus \mathfrak g$ is an ideal (Lie algebra). This property of the direct sum of Lie algebras is promoted to the definition of a trivial extension.

By semidirect sum

Inspired by the construction of a semidirect product (background) of groups using a homomorphism , one can make the corresponding construct for Lie algebras.

If is a Lie algebra homomorphism, then define a Lie bracket on $\mathfrak e = \mathfrak h \oplus \mathfrak g$ by

With this Lie bracket, the Lie algebra so obtained is denoted and is called the semidirect sum of and .

By inspection of one sees that is a subalgebra of and is an ideal in . Define by and by . It is clear that . Thus is a Lie algebra extension of by .

As with the trivial extension, this property generalizes to the definition of a split extension.

Example
Let be the Lorentz group and let denote the translation group in 4 dimensions, isomorphic to , and consider the multiplication rule of the Poincaré group
:$(a_2, \Lambda_2)(a_1, \Lambda_1) = (a_2 + \Lambda_2a_1, \Lambda_2\Lambda_1), \quad a_1, a_2 \in \mathrm T \subset \mathrm P, \Lambda_1, \Lambda_2 \in \mathrm O(3,1) \subset \mathrm P,$
(where and are identified with their images in ). From it follows immediately that, in the Poincaré group, . Thus every Lorentz transformation corresponds to an automorphism of with inverse and is clearly a homomorphism. Now define
:$\overline \mathrm P = \mathrm T \otimes_S \mathrm O(3, 1),$
endowed with multiplication given by . Unwinding the definitions one finds that the multiplication is the same as the multiplication one started with and it follows that . From follows that and then from it follows that .

By derivation

Let be a derivation (background) of and denote by the one-dimensional Lie algebra spanned by . Define the Lie bracket on byTo show that the Jacobi identity holds, one writes everything out, uses the fact that the underlying Lie algebras have a Lie product satisfying the Jacobi identity, and that .

: _1 + H_1, G_2 + H_2= lambda\delta + H_1, \mu\delta + H_2= _1, H_2+ \lambda \delta(H_1) - \mu \delta(H_2).

It is obvious from the definition of the bracket that is and ideal in in and that is a subalgebra of . Furthermore, is complementary to in . Let be given by and by . It is clear that . Thus is a split extension of by . Such an extension is called extension by a derivation.

If is defined by , then is a Lie algebra homomorphism into . Hence this construction is a special case of a semidirect sum, for when starting from and using the construction in the preceding section, the same Lie brackets result.

By 2-cocycle

If is a 2-cocycle (background) on a Lie algebra and is any one-dimensional vector space, let (vector space direct sum) and define a Lie bracket on by
:
mu H + G_1, \nu H + G_2= _1, G_2+ \epsilon(G_1, G_2)H, \quad \mu, \nu \in F.

Here is an arbitrary but fixed element of . Antisymmetry follows from antisymmetry of the Lie bracket on and antisymmetry of the 2-cocycle. The Jacobi identity follows from the corresponding properties of and of . Thus is a Lie algebra. Put and it follows that . Also, it follows with and that . Hence is a central extension of by . It is called extension by a 2-cocycle.

Theorems

Below follows some results regarding central extensions and 2-cocycles.

Theorem

Let and be cohomologous 2-cocycles on a Lie algebra and let and be respectively the central extensions constructed with these 2-cocycles. Then the central extensions and are equivalent extensions.

Proof

By definition, . Define
:$\psi: G + \mu c \in \mathfrak_1 \mapsto G + \mu c + f(G)c \in \mathfrak_2.$
It follows from the definitions that is a Lie algebra isomorphism and holds.

Corollary

A cohomology class defines a central extension of which is unique up to isomorphism.

The trivial 2-cocycle gives the trivial extension, and since a 2-coboundary is cohomologous with the trivial 2-cocycle, one has

Corollary

A central extension defined by a coboundary is equivalent with a trivial central extension.

Theorem

A finite-dimensional simple Lie algebra has only trivial central extensions.

Proof

Since every central extension comes from a 2-cocycle , it suffices to show that every 2-cocycle is a coboundary. Suppose is a 2-cocycle on . The task is to use this 2-cocycle to manufacture a 1-cochain such that .

The first step is to, for each , use to define a linear map . But the linear maps are elements of . This suffices to express in terms of , using the isomorphism . Next, a linear map is defined that turns out to be a derivation. Since all derivations are inner, one has for some . An expression for in terms of and is obtained. Thus set, trusting that is a derivation,
:\varphi(G_1, G_2) \equiv \rho_(G_2) = K(\nu^(\rho_), G_2) \equiv K(d(G_1), G_2) = K(\mathrm_(G_1), G_2) = K( _d, G_1 G_2) = K(G_d, _1, G_2.

Let be the 1-cochain defined by
:$f(G) = K(G_d,G).$
Then
:\delta f(G_1, G_2) = f( _1, G_2 = K(G_d, _1, G_2 = \varphi(G_1, G_2),
showing that is a coboundary. By the previous results, any central extension is trivial.

To verify that actually is a derivation, first note that it is linear since is, then compute
:\beginK(d( _1, G_2, G_3)) &= \varphi( _1, G_2, G_3)) = \varphi(G_1, _2, G_3 + \varphi(G_2, _3, G_1\\
&= K(d(G_1), _2, G_3 + K(d(G_1), (G_3, G_1)) = K( (G_1),G_2 G_3) + K( _1, d(G_2) G_3))\\
&= K( (G_1), G_2+ _1, d(G_2) G_3).\end
By appeal to the non-degeneracy of , the left arguments of are equal on the far left and far right.

The observation that one can define a derivation , given a symmetric non-degenerate associative form and a 2-cocycle , by
:$K(\nu^(\rho_), G_2) \equiv K(d(G_1), G_2),$
or using the symmetry of and the antisymmetry of ,
:$K(d(G_1), G_2) = -K(G_1, d(G_2)),$
leads to a corollary.

Corollary

Let be a non-degenerate symmetric associative bilinear form and let be a derivation satisfying
:$L(d(G_1), G_2) = -L(G_1, d(G_2)),$
then defined by
:$\varphi(G_1, G_2) = L(d(G_1), G_2)$
is a 2-cocycle.

Proof
The condition on ensures the antisymmetry of . The Jacobi identity for 2-cocycles follows starting with
:\varphi( 1, G_2 G_3) = L(d 1, G_2 G_3) = L( (G1), G_2 G_3) + L( 1, d(G_2) G_3),
using symmetry of the form, the antisymmetry of the bracket, and once again the definition of in terms of .

If is the Lie algebra of a Lie group and is a central extension of , one may ask whether there is a Lie group with Lie algebra . The answer is, by Lie's third theorem affirmative. But is there a ''central extension'' of with Lie algebra ? The answer to this question requires some machinery, and can be found in .

Applications

The "negative" result of the preceding theorem indicates that one must, at least for semisimple Lie algebras, go to infinite-dimensional Lie algebras to find useful applications of central extensions. There are indeed such. Here will be presented affine Kac–Moody algebras and Virasoro algebras. These are extensions of polynomial loop-algebras and the Witt algebra respectively.

Polynomial loop-algebra

Let be a polynomial loop algebra (background),
:\mathfrak g = \mathbb lambda, \lambda^\otimes \mathfrak g_0,
where is a complex finite-dimensional simple Lie algebra. The goal is to find a central extension of this algebra. Two of the theorems apply. On the one hand, if there is a 2-cocycle on , then a central extension may be defined. On the other hand, if this 2-cocycle is acting on the part (only), then the resulting extension is trivial. Moreover, derivations acting on (only) cannot be used for definition of a 2-cocycle either because these derivations are all inner and the same problem results. One therefore looks for derivations on . One such set of derivations is
:$d_k \equiv \lambda^\frac, \quad k \in \mathbb Z.$

In order to manufacture a non-degenerate bilinear associative antisymmetric form on , attention is focused first on restrictions on the arguments, with fixed. It is a theorem that ''every'' form satisfying the requirements is a multiple of the Killing form on . This requires
:$L(\lambda^m \otimes G_1, \lambda^n \otimes G_2) = \gamma_K(G_1, G_2).$
Symmetry of implies
:$\gamma_=\gamma_,$
and associativity yields
:$\gamma_=\gamma_.$
With one sees that . This last condition implies the former. Using this fact, define . The defining equation then becomes
:$L(\lambda^m \otimes G_1, \lambda^n \otimes G_2) = f(m+n)K(G_1, G_2).$
For every the definition
:$f(n) = \delta_ \Leftrightarrow \gamma_=\delta_$
does define a symmetric associative bilinear form
:$L_i(\lambda^m \otimes G_1, \lambda^n \otimes G_2) = \delta_K(G_1, G_2).$
These span a vector space of forms which have the right properties.

Returning to the derivations at hand and the condition
:$L_i(d_k(\lambda^l \otimes G_1), \lambda^m \otimes G_2) = -L_i(\lambda^l \otimes G_1, d_k(\lambda^m \otimes G_2)),$
one sees, using the definitions, that
:$l\delta_ = -m\delta_,$
or, with ,
:$n\delta_ = 0.$
This (and the antisymmetry condition) holds if , in particular it holds when .

Thus choose and . With these choices, the premises in the corollary are satisfied. The 2-cocycle defined by
:$\varphi(P(\lambda) \otimes G_1), Q(\lambda) \otimes G_2)) = L(\lambda\frac \otimes G_1, Q(\lambda) \otimes G_2)$
is finally employed to define a central extension of ,
:$\mathfrak e = \mathfrak g \oplus \mathbb CC,$
with Lie bracket
: (\lambda) \otimes G_1 + \mu C, Q(\lambda) \otimes G_2 + \nu C= P(\lambda)Q(\lambda)\otimes _1, G_2+ \varphi(P(\lambda) \otimes G_1,Q(\lambda) \otimes G_2)C.
For basis elements, suitably normalized and with antisymmetric structure constants, one has
:\begin lambda^l \otimes G_i + \mu C, \lambda^m \otimes G_j + \nu C&= \lambda^\otimes _i, G_j+ \varphi(\lambda^l \otimes G_i,\lambda^m \otimes G_j)C\\
&= \lambda^\otimes ^kG_k + L(\lambda \frac \otimes G_i, \lambda^m \otimes G_j)C\\
&=\lambda^\otimes ^kG_k + lL(\lambda^l \otimes G_i, \lambda^m \otimes G_j)C\\
&=\lambda^\otimes ^kG_k + l\delta_K(G_i, G_j)C\\
&=\lambda^\otimes ^kG_k + l\delta_^m^kC = \lambda^\otimes ^kG_k + l\delta_\delta^C.
\end

This is a universal central extension of the polynomial loop algebra.

;A note on terminology
In physics terminology, the algebra of above might pass for a Kac–Moody algebra, whilst it will probably not in mathematics terminology. An additional dimension, an extension by a derivation is required for this. Nonetheless, if, in a physical application, the eigenvalues of or its representative are interpreted as (ordinary) quantum numbers, the additional superscript on the generators is referred to as the level. It is an additional quantum number. An additional operator whose eigenvalues are precisely the levels is introduced further below.

Current algebra

As an application of a central extension of polynomial loop algebra, a current algebra of a quantum field theory is considered (background). Suppose one has a current algebra, with the interesting commutator being

with a Schwinger term. To construct this algebra mathematically, let be the centrally extended polynomial loop algebra of the previous section with
: lambda^l \otimes G_i + \mu C, \lambda^m \otimes G_j + \nu C= \lambda^\otimes ^kG_k + l\delta_\delta_C
as one of the commutation relations, or, with a switch of notation () with a factor of under the physics convention,
: ^m_a, T^n_b= i^cT^_c + m\delta_\delta_C.

Define using elements of ,
:$J_a(x) = \frac\sum_^e^T_a^, x \in \mathbb R.$
One notes that
:$J_a(x+L) = J_a(x)$
so that it is defined on a circle. Now compute the commutator,
:$\begin[] [J_a(x),J_b(y)]&= \left(\frac\right)^2\left[\sum_^e^T_a^, \sum_^e^T_b^\right]\\ &=\left(\frac\right)^2 \sum_^ e^ e^ [T_a^,T_b^].\end$

For simplicity, switch coordinates so that and use the commutation relations,
:$\begin[] [J_a(z),J_b(0)] &= \left(\frac\right)^2\sum_^ e^[i^cT^_c + m\delta_\delta_C]\\ &=\left(\frac\right)^2\sum_^ e^\sum_^\infty ie^^cT^_c + \left(\frac\right)^2\sum_^e^m\delta_\delta_C\\ &=\left(\frac\right)\sum_^ e^i^cJ_c(z) - \left(\frac\right)^2\sum_^e^n\delta_C\end$

Now employ the Poisson summation formula,
:$\frac\sum_^\infty e^ = \frac\sum_^\infty \delta(z + nL) = \delta(z)$
for in the interval and differentiate it to yield
:$-\frac\sum_^\infty ne^ = \delta'(z),$
and finally
: _a(x-y),J_b(0)= i\hbar ^cJ_c(x-y)\delta(x-y) + \frac\delta_C\delta'(x-y),
or
: _a(x),J_b(y)= i\hbar ^cJ_c(x)\delta(x-y) + \frac\delta_C\delta'(x-y),
since the delta functions arguments only ensure that the arguments of the left and right arguments of the commutator are equal (formally ).

By comparison with , this is a current algebra in two spacetime dimensions, ''including a Schwinger term'', with the space dimension curled up into a circle. In the classical setting of quantum field theory, this is perhaps of little use, but with the advent of string theory where fields live on world sheets of strings, and spatial dimensions are curled up, there may be relevant applications.

Kac–Moody algebra

The derivation used in the construction of the 2-cocycle in the previous section can be extended to a derivation on the centrally extended polynomial loop algebra, here denoted by in order to realize a Kac–Moody algebra (background). Simply set
:$D(P(\lambda) \otimes G + \mu C) = \lambda \frac \otimes G.$
Next, define as a vector space
:$\mathfrak e = \mathbb Cd + \mathfrak g.$
The Lie bracket on is, according to the standard construction with a derivation, given on a basis by
:\begin lambda^m \otimes G_1 + \mu C + \nu D, \lambda^n \otimes G_2 + \mu' C + \nu' D&= \lambda^ \otimes _1, G_2+ m\delta_K(G_1, G_2)C + \nu D(\lambda^n \otimes G_1) - \nu'D(\lambda^m \otimes G_2)\\
&= \lambda^ \otimes _1, G_2+ m\delta_K(G_1, G_2)C + \nu n\lambda^n \otimes G_1 - \nu'm \lambda^m \otimes G_2.\end

For convenience, define
:$G_i^m \leftrightarrow \lambda^m \otimes G_i.$
In addition, assume the basis on the underlying finite-dimensional simple Lie algebra has been chosen so that the structure coefficients are antisymmetric in all indices and that the basis is appropriately normalized. Then one immediately through the definitions verifies the following commutation relations.

:\begin _i^m,G_j^n&= ^kG_k^ + m\delta_\delta^C,\\
,G_i^m&= 0, \quad 1 \le i, j, N,\quad m,n \in \mathbb Z\\
, G_i^m&= mG_i^m\\
,C&= 0.\end
These are precisely the short-hand description of an untwisted affine Kac–Moody algebra. To recapitulate, begin with a finite-dimensional simple Lie algebra. Define a space of formal Laurent polynomials with coefficients in the finite-dimensional simple Lie algebra. With the support of a symmetric non-degenerate alternating bilinear form and a derivation, a 2-cocycle is defined, subsequently used in the standard prescription for a central extension by a 2-cocycle. Extend the derivation to this new space, use the standard prescription for a split extension by a derivation and an untwisted affine Kac–Moody algebra obtains.

Virasoro algebra

The purpose is to construct the Virasoro algebra (named after Miguel Angel Virasoro) Miguel Angel Virasoro, born 1940 is an Argentine physicist. The Virasoro algebra, named after him, was first published in as a central extension by a 2-cocycle of the Witt algebra (background). The Jacobi identity for 2-cocycles yields

Letting and using antisymmetry of one obtains
:$(m+p)\eta_=(m-p)\eta_.$

In the extension, the commutation relations for the element are
: _0 + \mu C, d_m + \nu C\varphi = -md_m + \eta_C = -m(d_m - \fracC).
It is desirable to get rid of the central charge on the right hand side. To do this define
:$f:W \to \mathbb C; d_m \to \frac = \frac.$
Then, using as a 1-cochain,
:\eta'_ = \varphi'(d_0, d_n) = \varphi(d_0, d_n) + \delta f( _0, d_n = \varphi(d_0, d_n) -n \frac= 0,
so with this 2-cocycle, equivalent to the previous one, one hasThe same effect can be obtained by a change of basis in .
: _0 + \mu C, d_m + \nu C = -md_m.
With this new 2-cocycle (skip the prime) the condition becomes
:$(n+p)\eta_ = (n-p)\eta_=0,$
and thus
:$\eta_=a(m)\delta_, \quad a(-m) = -a(m),$
where the last condition is due to the antisymmetry of the Lie bracket. With this, and with (cutting out a "plane" in ), yields
:$(2m+p)a(p) + (m-p)a(m+p) + (m+2p)a(m) = 0,$

that with (cutting out a "line" in ) becomes
:$(m-1)a(m+1) - (m+2)a(m) + (2m+1)a(1) = 0.$
This is a difference equation generally solved by
:$a(m) = \alpha m + \beta m^3.$
The commutator in the extension on elements of is then
: _l, d_m= (l-m)d_ + (\alpha m + \beta m^3)\delta_C.
With it is possible to change basis (or modify the 2-cocycle by a 2-coboundary) so that
: '_l, d'_m= (l-m)d_,
with the central charge absent altogether, and the extension is hence trivial. (This was not (generally) the case with the previous modification, where only obtained the original relations.) With the following change of basis,
:$d'_l = d_l + \delta_\fracC,$
the commutation relations take the form
: '_l, d'_m= (l-m)d'_ + (\gamma m + \beta m^3)\delta_C,
showing that the part linear in is trivial. It also shows that is one-dimensional (corresponding to the choice of ). The conventional choice is to take and still retaining freedom by absorbing an arbitrary factor in the arbitrary object . The Virasoro algebra is then
:$\mathcal V = \mathcal W + \mathbb C C,$
with commutation relations

Bosonic open strings

The relativistic classical open string (background) is subject to quantization. This roughly amounts to taking the position and the momentum of the string and promoting them to operators on the space of states of open strings. Since strings are extended objects, this results in a continuum of operators depending on the parameter . The following commutation relations are postulated in the Heisenberg picture.
:\begin ^I(\tau, \sigma), \mathcal P^(\tau, \sigma)&= i\eta^\delta(\sigma-\sigma'),\\
_0^-(\tau),p^+(\tau)&= -i.\end
All other commutators vanish.

Because of the continuum of operators, and because of the delta functions, it is desirable to express these relations instead in terms of the quantized versions of the Virasoro modes, the Virasoro operators. These are calculated to satisfy
: alpha_m^I, \alpha_n^J= m\eta^\delta_
They are interpreted as creation and annihilation operators acting on Hilbert space, increasing or decreasing the quantum of their respective modes. If the index is negative, the operator is a creation operator, otherwise it is an annihilation operator. (If it is zero, it is proportional to the total momentum operator.) In view of the fact that the light cone plus and minus modes were expressed in terms of the transverse Virasoro modes, one must consider the commutation relations between the Virasoro operators. These were classically defined (then modes) as
:$L_n = \frac\sum_\alpha_^I\alpha_p^I.$

Since, in the quantized theory, the alphas are operators, the ordering of the factors matter. In view of the commutation relation between the mode operators, it will only matter for the operator (for which ). is chosen normal ordered,
:$L_0 = \frac\alpha_0^I\alpha_0^I + \sum_^\infty \alpha_^I\alpha_p^I, = \alpha' p^Ip^I + \sum_^\infty p \alpha_^\alpha_p^I + c$
where is a possible ordering constant. One obtains after a somewhat lengthy calculation the relations
: _m, L_n= (m-n)L_, \quad m+n\ne 0.
If one would allow for above, then one has precisely the commutation relations of the Witt algebra. Instead one has
: _m, L_n= (m-n)L_ + \frac(m^3-m)\delta_,\quad \forall m,n \in \mathbb Z.
upon identification of the generic central term as times the identity operator, this is the Virasoro algebra, the universal central extension of the Witt algebra.

The operator enters the theory as the Hamiltonian, modulo an additive constant. Moreover, the Virasoro operators enter into the definition of the Lorentz generators of the theory. It is perhaps the most important algebra in string theory. The consistency of the Lorentz generators, by the way, fixes the spacetime dimensionality to 26. While this theory presented here (for relative simplicity of exposition) is unphysical, or at the very least incomplete (it has, for instance, no fermions) the Virasoro algebra arises in the same way in the more viable superstring theory and M-theory.

Group extension

A projective representation of a Lie group (background) can be used to define a so-called group extension .

In quantum mechanics, Wigner's theorem asserts that if is a symmetry group, then it will be represented projectively on Hilbert space by unitary or antiunitary operators. This is often dealt with by passing to the universal covering group of and take it as the symmetry group. This works nicely for the rotation group and the Lorentz group , but it does not work when the symmetry group is the Galilean group. In this case one has to pass to its central extension, the Bargmann group, which is the symmetry group of the Schrödinger equation. Likewise, if , the group of translations in position and momentum space, one has to pass to its central extension, the Heisenberg group.

Let be the 2-cocycle on induced by . DefineIf the 2-cocycle takes its values in the abelian group , i. e. it is a phase factor, which will always be the case in the contezt of Wigner's theorem, then may be replaced with in the construction.
:$G_ = \mathbb C^* \times G = \$
as a set and let the multiplication be defined by
:$(\lambda_1,g_1)(\lambda_2,g_2) = (\lambda_1\lambda_2\omega(g_1,g_2),g_1g_2).$
Associativity holds since is a 2-cocycle on . One has for the unit element
:$(1,e)(\lambda,g) = (\lambda\omega(e,g),g) = (\lambda,g) = (\lambda,g)(1,e),$
and for the inverse
:$(\lambda,g)^ = \left(\frac, g^\right).$

The set is an abelian subgroup of . This means that is not semisimple. The center of , includes this subgroup. The center may be larger.

At the level of Lie algebras it can be shown that the Lie algebra of is given by
:$\mathfrak_ = \mathbb CC \oplus \mathfrak g,$
as a vector space and endowed with the Lie bracket
: mu C + G_1,\nu C + G_2= _1, G_2+ \eta(G_1, G_2)C.
Here is a 2-cocycle on . This 2-cocycle can be obtained from albeit in a highly nontrivial way. The reference states the fact and that it is difficult to show. No further references are given. Expressions on a slightly different form can be found though in and .

Now by using the projective representation one may define a map by
:$\Pi_((\lambda,g)) = \lambda\Pi(g).$
It has the properties
:$\Pi_((\lambda_1,g_1))\Pi_((\lambda_2,g_2))=\lambda_1\lambda_2\Pi(g_1)\Pi(g_2)=\lambda_1\lambda_2\omega(g_1,g_2)\Pi(g_1g_2)=\Pi_(\lambda_1\lambda_2\omega(g_1,g_2),g_1g_2) = \Pi_((\lambda_1, g_1)(\lambda_2,g_2)),$
so is a bona fide representation of .

In the context of Wigner's theorem, the situation may be depicted as such (replace by ); let denote the unit sphere in Hilbert space , and let be its inner product. Let denote ray space and the ray product. Let moreover a wiggly arrow denote a group action. Then the diagram

commutes, i.e.
:$\pi_2 \circ \Pi_((\lambda, g))(\psi) = \Pi \circ \pi(g)(\pi_1(\psi)), \quad \psi \in S\mathcal H.$
Moreover, in the same way that is a symmetry of preserving , is a symmetry of preserving . The fibers of are all circles. These circles are left invariant under the action of . The action of on these fibers is transitive with no fixed point. The conclusion is that is a principal fiber bundle over with structure group .

Background material

In order to adequately discuss extensions, structure that goes beyond the defining properties of a Lie algebra is needed. Rudimentary facts about these are collected here for quick reference.

Derivations

A derivation on a Lie algebra is a map
:$\delta: \mathfrak g \rightarrow \mathfrak g$
such that the Leibniz rule
:\delta _1, G_2= delta G_1, G_2+ _1, \delta G_2/math>
holds. The set of derivations on a Lie algebra is denoted . It is itself a Lie algebra under the Lie bracket
: delta_1, \delta_2= \delta_1 \circ \delta_2 - \delta_2 \circ \delta_1.
It is the Lie algebra of the group of automorphisms of . One has to show
:\delta _1, G_1= delta G_1, G_2+ _1, \delta G_2\Leftrightarrow e^ _1,G_2= ^G_1, e^G_2 \quad \forall t \in \mathbb R.
If the rhs holds, differentiate and set implying that the lhs holds. If the lhs holds , write the rhs as
: _1,G_2; \overset\; e^ ^G_1, e^G_2
and differentiate the rhs of this expression. It is, using , identically zero. Hence the rhs of this expression is independent of and equals its value for , which is the lhs of this expression.

If , then , acting by , is a derivation. The set is the set of inner derivations on . For finite-dimensional simple Lie algebras all derivations are inner derivations.

Semidirect product (groups)

Consider two Lie groups and and , the automorphism group of . The latter is the group of isomorphisms of . If there is a Lie group homomorphism , then for each there is a with the property . Denote with the ''set'' and define multiplication by

Then is a group with identity and the inverse is given by . Using the expression for the inverse and equation it is seen that is normal in . Denote the group with this semidirect product as .

Conversely, if is a given semidirect product expression of the group , then by definition is normal in and for each where and the map is a homomorphism.

Now make use of the Lie correspondence. The maps each induce, at the level of Lie algebras, a map . This map is computed by

For instance, if and are both subgroups of a larger group and , then

and one recognizes as the adjoint action of on restricted to . Now /nowiki> if is finite-dimensional/nowiki> is a homomorphism,To see this, apply formula to , recall that is a homomorphism, and use a couple of times. and appealing once more to the Lie correspondence, there is a unique Lie algebra homomorphism .The fact that the Lie algebra of is , the set of all derivations of (itself being a Lie algebra under the obvious bracket), can be found in This map is (formally) given by

for example, if , then (formally)

where a relationship between and the adjoint action rigorously proved in here is used.

Lie algebra

The Lie algebra is, as a vector space, . This is clear since generates and . The Lie bracket is given by
: _1 + G_1,H_2 + G_2\mathfrak e = _1, H_2\mathfrak h + \psi_(H_2) -\psi_(H_1) + _1, G_2\mathfrak g.

To compute the Lie bracket, begin with a surface in parametrized by and . Elements of in are decorated with a bar, and likewise for .
:$\begin e^ &= e^e^e^=(1,e^)(e^,1)(1,e^)\\ &=(\phi_(e^), e^)(1,e^) = (\phi_(e^)\phi_(1),1)\\ &= (\phi_(e^),1) \end$

One has
:$\frac \left. e^\_ = Ad_\overline$
and
:$\frac \left. (\phi_(e^),1)\_ = (\Psi_(H), 0)$
by and thus
:$Ad_\overline = (\Psi_(H), 0).$

Now differentiate this relationship with respect to and evaluate at $:
:
\frac \left .e^\overlinee^\_ = overline, \overline
and
:$\frac \left .(\Psi_(H), 0)\_ = (\psi_G(H), 0)$
by and thus
: _1 + G_1,H_2 + G_2\mathfrak e = _1, H_2\mathfrak h + _1, H_2+ _1, G_2+ _1, G_2\mathfrak g = _1, H_2\mathfrak h + \psi_(H_2) -\psi_(H_1) + _1, G_2\mathfrak g.

Cohomology

For the present purposes, consideration of a limited portion of the theory Lie algebra cohomology suffices. The definitions are not the most general possible, or even the most common ones, but the objects they refer to are authentic instances of more the general definitions.

2-cocycles

The objects of primary interest are the 2-cocycles on , defined as bilinear alternating functions,
:$\phi:\mathfrak g \times \mathfrak g \rightarrow F,$
that are alternating,
:$\phi(G_1, G_2) = -\phi(G_2, G_1),$
and having a property resembling the Jacobi identity called the Jacobi identity for 2-cycles,
:
\phi(G_1, _2, G_3 + \phi(G_2, _3, G_1 + \phi(G_3, _1, G_2 = 0.

The set of all 2-cocycles on is denoted .

2-cocycles from 1-cochains

Some 2-cocycles can be obtained from 1-cochains. A 1-cochain on is simply a linear map,
:$f:\mathfrak g \rightarrow F$
The set of all such maps is denoted and, of course (in at least the finite-dimensional case) . Using a 1-cochain , a 2-cocycle may be defined by
:
\delta f(G_1, G_2) = f( _1, G_2.

The alternating property is immediate and the Jacobi identity for 2-cocycles is (as usual) shown by writing it out and using the definition and properties of the ingredients (here the Jacobi identity on and the linearity of ). The linear map is called the coboundary operator (here restricted to ).

The second cohomology group

Denote the image of of by . The quotient
:$H^2(\mathfrak g, \mathbb F) = Z^2(\mathfrak g, \mathbb F)/B^2(\mathfrak g, \mathbb F)$
is called the second cohomology group of . Elements of are equivalence classes of 2-cocycles and two
2-cocycles and are called equivalent cocycles if they differ by a 2-coboundary, i.e. if for some . Equivalent
2-cocycles are called cohomologous. The equivalence class of is denoted .

These notions generalize in several directions. For this, see the main articles.

Structure constants

Let be a Hamel basis for . Then each has a unique expression as
:$G = \sum_c_\alpha G_\alpha, \quad c_\alpha \in F, G_\alpha \in B$
for some indexing set of suitable size. In this expansion, only finitely many are nonzero. In the sequel it is (for simplicity) assumed that the basis is countable, and Latin letters are used for the indices and the indexing set can be taken to be . One immediately has
: _i, G_j= ^k G_k
for the basis elements, where the summation symbol has been rationalized away, the summation convention applies. The placement of the indices in the structure constants (up or down) is immaterial. The following theorem is useful:

Theorem:There is a basis such that the structure constants are antisymmetric in all indices if and only if the Lie algebra is a direct sum of simple compact Lie algebras and Lie algebras. This is the case if and only if there is a real positive definite metric on satisfying the invariance condition
:$g__=-g__.$
in any basis. This last condition is necessary on physical grounds for non-Abelian gauge theories in quantum field theory. Thus one can produce an infinite list of possible gauge theories using the Cartan catalog of simple Lie algebras on their compact form (i.e., , etc. One such gauge theory is the gauge theory of the standard model with Lie algebra .

Killing form

The Killing form is a symmetric bilinear form on defined by
:$K(G_1, G_2) = \mathrm (\mathrm_\mathrm_).$
Here is viewed as a matrix operating on the vector space . The key fact needed is that if is semisimple, then, by Cartan's criterion, is non-degenerate. In such a case may be used to identify and . If , then there is a such that
:$\langle \lambda, G \rangle = K(G_\lambda, G) \quad \forall G \in \mathfrak g.$
This resembles the Riesz representation theorem and the proof is virtually the same. The Killing form has the property
:K( _1, G_2 G_3) = K(G_1, _2, G_3,
which is referred to as associativity. By defining and expanding the inner brackets in terms of structure constants, one finds that the Killing form satisfies the invariance condition of above.

Loop algebra

A loop group is taken as a group of smooth maps from the unit circle into a Lie group with the group structure defined by the group structure on . The Lie algebra of a loop group is then a vector space of mappings from into the Lie algebra of . Any subalgebra of such a Lie algebra is referred to as a loop algebra. Attention here is focused on polynomial loop algebras of the form
:$\.$

To see this, consider elements near the identity in for in the loop group, expressed in a basis for
:$H(\lambda) = e^ = e_G + h^k(\lambda)G_k + \ldots ,$
where the are real and small and the implicit sum is over the dimension of .
Now write
:$h^k(\lambda) = \sum_^\infty \theta^k_\lambda^n$
to obtain
:$e^ = 1_G + \sum_^\infty \theta^k_\lambda^nG_k + \ldots .$
Thus the functions
:$h:S^1 \to \mathfrak g; h(\lambda) = \sum_^\infty \sum_^K\theta^k_\lambda^nG_k \equiv \sum_^\infty \lambda^nG_n$
constitute the Lie algebra.

A little thought confirms that these are loops in as goes from to . The operations are the ones defined pointwise by the operations in . This algebra is isomorphic with the algebra
:C lambda, \lambda^\otimes \mathfrak g,
where is the algebra of Laurent polynomials,
:$\sum \lambda^k G_k \leftrightarrow \sum \lambda^k \otimes G_k.$
The Lie bracket is
: (\lambda) \otimes G_1, Q(\lambda) \otimes G_2= P(\lambda)Q(\lambda) \otimes _1, G_2
In this latter view the elements can be considered as polynomials with (constant!) coefficients in . In terms of a basis and structure constants,
: lambda^m \otimes G_i, \lambda^n \otimes G_j= ^k\lambda^ \otimes G_k.

It is also common to have a different notation,
:$\lambda^m \otimes G_i \cong \lambda^mG_i \leftrightarrow T^m_i(\lambda) \equiv T^m_i,$
where the omission of should be kept in mind to avoid confusion; the elements really are functions . The Lie bracket is then

which is recognizable as one of the commutation relations in an untwisted affine Kac–Moody algebra, to be introduced later, ''without'' the central term. With , a subalgebra isomorphic to is obtained. It generates (as seen by tracing backwards in the definitions) the set of constant maps from into , which is obviously isomorphic with when is onto (which is the case when is compact. If is compact, then a basis for may be chosen such that the are skew-Hermitian. As a consequence,
:$T_i^ = (\lambda^nG_i)^ = -\lambda^G_i = -T_i^.$
Such a representation is called unitary because the representatives
:$H(\lambda) = e^ \in G$
are unitary. Here, the minus on the lower index of is conventional, the summation convention applies, and the is (by the definition) buried in the s in the right hand side.

Current algebra (physics)

Current algebras arise in quantum field theories as a consequence of global gauge symmetry. Conserved currents occur in classical field theories whenever the Lagrangian respects a continuous symmetry. This is the content of Noether's theorem. Most (perhaps all) modern quantum field theories can be formulated in terms of classical Lagrangians (prior to quantization), so Noether's theorem applies in the quantum case as well. Upon quantization, the conserved currents are promoted to position dependent operators on Hilbert space. These operators are subject to commutation relations, generally forming an infinite-dimensional Lie algebra. A model illustrating this is presented below.

To enhance the flavor of physics, factors of will appear here and there as opposed to in the mathematical conventions.Roughly, the whole Lie algebra is multiplied by , there is an occurring in the definition of the structure constants and the exponent in the exponential map (Lie theory) acquires a factor of (minus) . the main reason for this convention is that physicists like their Lie algebra elements to be Hermitian (as opposed to skew-Hermitian) in order for them to have real eigenvalues and hence be candidates for observables.

Consider a column vector of scalar fields . Let the Lagrangian density be
:$\mathcal L = \partial_\mu \phi^\dagger\partial^\mu\phi - m^2\phi^\dagger\phi.$

This Lagrangian is invariant under the transformationSince and are constant, they may be pulled out of partial derivatives. The and then combine in by unitarity.
:$\phi \mapsto e^\phi,$

where are generators of either or a closed subgroup thereof, satisfying
: _a, F_b= i^cF_c.

Noether's theorem asserts the existence of conserved currents,
:$J_a^\mu = -\pi^\mu iF_a\phi, \quad \pi^ = \frac,$

where is the momentum canonically conjugate to .
The reason these currents are said to be ''conserved'' is because
:$\partial_\mu J^\mu_a = 0,$
and consequently
:$Q_a(t) = \int J^0_a d^3x = \mathrm \equiv Q_a,$

the charge associated to the charge density is constant in time.This follows from Gauss law is based on the assumption of a sufficiently rapid fall-off of the fields at infinity. This (so far classical) theory is quantized promoting the fields and their conjugates to operators on Hilbert space and by postulating (bosonic quantization) the commutation relationsThere are alternative routes to quantization, e.g. one postulates the existence of creation and annihilation operators for all particle types with certain exchange symmetries based on which statistics, Bose–Einstein or Fermi–Dirac, the particles obey, in which case the above are derived for scalar bosonic fields using mostly Lorentz invariance and the demand for the unitarity of the S-matrix. In fact, ''all'' operators on Hilbert space can be built out of creation and annihilation operators. See e.g. , chapters 2–5.
:\begin phi_k(t, x), \pi^l(t, x)&= i\delta(x-y)\delta_k^l,\\
phi_k(t, x), \phi_l(t, x)= pi^k(t, x), \pi^l(t, x)= 0.\end
The currents accordingly become operatorsThis step is ambiguous, since the classical fields commute whereas the operators don't. Here it is pretended that this problem doesn't exist. In reality, it is never serious as long as one is consistent. They satisfy, using the above postulated relations, the definitions and integration over space, the commutation relations
:\begin _a^0(t, \mathbf x),J_b^0(t, \mathbf y)&= i\delta(\mathbf x - \mathbf y)^cJ_c^0(ct, \mathbf x)\\
_a, Q_b&= i^cQ_c\\
_a, J_b^\mu(t, \mathbf x)&= i^cJ_c^\mu(t, \mathbf x),\end
where the speed of light and the reduced Planck's constant have been set to unity. The last commutation relation does ''not'' follow from the postulated commutation relations (these are fixed only for , not for ), except for For the Lorentz transformation behavior is used to deduce the conclusion. The next commutator to consider is
: _a^0(t, \mathbf x), J_b^i(t, \mathbf y)= i^cJ_c^i(t, \mathbf x)\delta(\mathbf x - \mathbf y) + S_^\partial_j\delta(\mathbf x - \mathbf y) + ... .
The presence of the delta functions and their derivatives is explained by the requirement of microcausality that implies that the commutator vanishes when . Thus the commutator must be a distribution supported at . The first term is fixed due to the requirement that the equation should, when integrated over , reduce to the last equation before it. The following terms are the Schwinger terms. They integrate to zero, but it can be shown quite generally that they must be nonzero.

Consider a conserved current

with a generic Schwinger term
: ^0(t,\mathbf x),J^i(t,\mathbf y)= i\delta(\mathbf x - \mathbf y)J^i(t,\mathbf x) + C^i(\mathbf x, \mathbf y).
By taking the vacuum expectation value (VEV),
:\langle 0, C^i(\mathbf x, \mathbf y), 0\rangle = \langle 0, ^0(t,\mathbf x),J^i(t,\mathbf y)0\rangle,
one finds
:\begin\langle 0, \frac, 0\rangle &= \langle 0, ^0(t,\mathbf x),\frac0\rangle\\ &= -\langle 0, ^0(t,\mathbf x),\frac0\rangle = i\langle 0, ^0(t,\mathbf_x),[J^0(t,\mathbf_y),H.html" ;"title="^0(t,\mathbf_y),H.html" ;"title="^0(t,\mathbf x),[J^0(t,\mathbf y),H">^0(t,\mathbf x),[J^0(t,\mathbf y),H">0\rangle\\ &= -i\langle 0, J^0(t,\mathbf x)HJ^0(t,\mathbf y)+J^0(t,\mathbf x)HJ^0(t,\mathbf x), 0\rangle,\end
where and Heisenberg's equation of motion have been used as well as and its conjugate.

Multiply this equation by and integrate with respect to and over all space, using integration by parts, and one finds
:$-i\int\int d\mathbf x d\mathbf y\langle 0, C^i(\mathbf x, \mathbf y), 0\rangle f(\mathbf x)\fracf(\mathbf x) = 2\langle 0, FHF, \rangle, \quad F = \int J^0(\mathbf x)f(\mathbf x).$
Now insert a complete set of states,
:$\langle 0, FHF, \rangle = \sum_\langle 0, F, m\rangle\langle m, H, n\rangle\langle n, F, 0\rangle=\sum_\langle 0, F, m\rangle E_n\delta_\langle n, F, 0\rangle ) \sum_, \langle 0, F, n\rangle, ^2E_n > 0 \Rightarrow C^i(\mathbf x, \mathbf y) \ne 0.$
Here hermiticity of and the fact that not all matrix elements of between the vacuum state and the states from a complete set can be zero.

Affine Kac–Moody algebra

Let be an -dimensional complex simple Lie algebra with a dedicated suitable normalized basis such that the structure constants are antisymmetric in all indices with commutation relations
: _i,G_j= ^kG_k, \quad 1 \le i, j, N.
An untwisted affine Kac–Moody algebra is obtained by copying the basis for each (regarding the copies as distinct), setting
:$\overline = FC \oplus FD \oplus \bigoplus_ FG^i_m$
as a vector space and assigning the commutation relations

:\begin _i^m,G_j^n&= ^kG_k^ + m\delta_\delta^C,\\
,G_i^m&= 0, \quad 1 \le i, j, N,\quad m,n \in \mathbb Z\\
, G_i^m&= mG_i^m\\
,C&= 0.\end

If , then the subalgebra spanned by the is obviously identical to the polynomial loop algebra of above.

Witt algebra

The Witt algebra, named after Ernst Witt, is the complexification of the Lie algebra of smooth vector fields on the circle . In coordinates, such vector fields may be written
:$X = f(\varphi)\frac,$
and the Lie bracket is the Lie bracket of vector fields, on simply given by
: , Y= \left \frac, g\frac\right= \left(f\frac - g\frac\right)\frac.
The algebra is denoted .
A basis for is given by the set
:$\ = \left\.$
This basis satisfies

This Lie algebra has a useful central extension, the Virasoro algebra. It has dimensional subalgebras isomorphic with and . For each , the set spans a subalgebra isomorphic to .

For one has
: _0, d_= d_, \quad _0, d_= -d_,\quad _1, d_= 2d_0.
These are the commutation relations of with
:$d_0 \leftrightarrow H = \left(\begin 1 & 0\\ 0 & -1\end\right), \quad d_ \leftrightarrow X = \left(\begin 0 & 1\\ 0 & 0\end\right), \quad d_1 \leftrightarrow Y = \left(\begin 0 & 0\\ 1 & 0\end\right), \quad H, X, Y \in \mathfrak(2, \mathbb R).$
The groups and are isomorphic under the map
:$SU(1,1) = \left(\begin 1 & -i\\ 1 & i\end\right)SL(2, \mathbb R)\left(\begin 1 & -i\\ 1 & i\end\right)^,$
and the same map holds at the level of Lie algebras due to the properties of the exponential map.
A basis for is given, see classical group, by
:$U_0 = \left(\begin 0 & 1\\ 1 & 0\end\right), \quad U_1 = \left(\begin 0 & -i\\ i & 0\end\right), \quad U_2 = \left(\begin i & 0\\ 0 & -i\end\right)$
Now compute
:$\beginH_ &= \left(\begin 1 & -i\\ 1 & i\end\right)H\left(\begin 1 & -i\\ 1 & i\end\right)^ =\left(\begin 0 & 1\\ 1 & 0\end\right) = U_0,\\ X_ &= \left(\begin 1 & -i\\ 1 & i\end\right)X\left(\begin 1 & -i\\ 1 & i\end\right)^ =\frac\left(\begin i & -i\\ i & -i\end\right) = \frac(U_1+U_2),\\ Y_ &= \left(\begin 1 & -i\\ 1 & i\end\right)Y\left(\begin 1 & -i\\ 1 & i\end\right)^ =\frac\left(\begin -i & -i\\ i & i\end\right) = \frac(U_1-U_2). \end$
The map preserves brackets and there are thus Lie algebra isomorphisms between the subalgebra of spanned by with ''real'' coefficients, and . The same holds for ''any'' subalgebra spanned by , this follows from a simple rescaling of the elements (on either side of the isomorphisms).

Projective representation

If is a matrix Lie group, then elements of its Lie algebra m can be given by
:$G = \frac\left .(g(t))\_,$
where is a differentiable path in that goes through the identity element at . Commutators of elements of the Lie algebra can be computed using two paths, and the group commutator,
: _1, G_2= \frac \left .g_1(t)g_2(t)g_1(t)^g_2(t)^\_, \quad G_1 = g_1'(0), G_2 = g_2'(0).

Likewise, given a group representation , its Lie algebra is computed by
:$\begin[] [U_1, U_2] &= \frac \left .U(g_1(t))U(g_2(t))U(g_1(t))^U(g_2(t))^\_\\ &= \frac \left .U(g_1(t)g_2(t)g_1(t)^g_2(t)^)\_, \quad G_1 = g_1'(0), G_2 = g_2'(0).\end$
Then there is a Lie algebra isomorphism between and sending bases to bases, so that is a faithful representation of .

If however is a projective representation, i.e. a representation up to a phase factor, then the Lie algebra, as computed from the group representation, is ''not'' isomorphic to . In a projective representation the multiplication rule reads
:$U(g_1)U(g_2) = \omega(g_1, g_2)U(g_1g_2) = e^U(g_1g_2).$
The function ,often required to be smooth, satisfies
:$\begin\omega(g,e)&=\omega(e,g) = 1,\\ \omega(g_1, g_2g_3)\omega(g_2,g_3) &= \omega(g_1,g_2)\omega(g_1g_2,g_3)\\ \omega(g,g^)&=\omega(g^,g).\end$
It is called a 2-cocycle on .

One has
:$\begin[] [U_1, U_2] &= \frac \left .U(g_1(t))U(g_2(t))U(g_1(t))^U(g_2(t))^\_\\ &= \frac \left .e^U(g_1(t)g_2(t)g_1(t)^g_2(t)^)\_\\ &\equiv \frac \left .\Omega(g_1,g_2)U(g_1(t)g_2(t)g_1(t)^g_2(t)^)\_\\ &= \left .\frac\_ + \left .\frac\_I, \quad G_1 = g_1'(0), G_2 = g_2'(0),\end$
because both and evaluate to the identity at . For an explanation of the phase factors , see Wigner's theorem. The commutation relations in for a basis,
: _i,G_j= G_k
become in
: _i,U_j= U_k + D_I,
so in order for to be closed under the bracket (and hence have a chance of actually being a Lie algebra) a central charge must be included.

Relativistic classical string theory

A classical relativistic string traces out a world sheet in spacetime, just like a point particle traces out a world line. This world sheet can locally be parametrized using two parameters and . Points in spacetime can, in the range of the parametrization, be written . One uses a capital to denote points in spacetime actually being on the world sheet of the string. Thus the string parametrization is given by . The inverse of the parametrization provides a local coordinate system on the world sheet in the sense of manifolds.

The equations of motion of a classical relativistic string derived in the Lagrangian formalism from the Nambu–Goto action are Equation 6.53 (supported by 6.49, 6.50).
:$\frac + \frac = 0, \quad \mathcal P_\mu^\tau = -\frac\frac,\quad \mathcal P_\mu^\sigma = -\frac\frac.$
A dot ''over'' a quantity denotes differentiation with respect to and a prime differentiation with respect to . A dot ''between'' quantities denotes the relativistic inner product.

These rather formidable equations simplify considerably with a clever choice of parametrization called the light cone gauge. In this gauge, the equations of motion become
:$\ddot X^\mu - '' = 0,$
the ordinary wave equation. The price to be paid is that the light cone gauge imposes constraints,
:$\dot X^\mu \cdot ' = 0, \quad (\dot X)^2 + (X')^2 = 0,$
so that one cannot simply take arbitrary solutions of the wave equation to represent the strings. The strings considered here are open strings, i.e. they don't close up on themselves. This means that the Neumann boundary conditions have to be imposed on the endpoints. With this, the general solution of the wave equation (excluding constraints) is given by
:$X^\mu(\sigma, \tau) = x_0^\mu + 2\alpha'p_0^\mu\tau - i\sqrt\sum_\left( a_n^e^ - a_n^e^\right)\frac,$
where is the slope parameter of the string (related to the string tension). The quantities and are (roughly) string position from the initial condition and string momentum. If all the are zero, the solution represents the motion of a classical point particle.

This is rewritten, first defining
:$\alpha_0^\mu = \sqrta_ \mu,\quad \alpha_n^\mu = a_n^\mu\sqrt, \quad \alpha_^\mu = a_n^\sqrt,$
and then writing
:$X^\mu(\sigma, \tau) = x_0^\mu + \sqrt\alpha_0^\mu \tau + i\sqrt\sum_\frac\alpha_n^e^\cos n\sigma.$

In order to satisfy the constraints, one passes to light cone coordinates. For , where is the number of ''space'' dimensions, set
:$\begin X^I(\sigma, \tau) &= x_0^I + \sqrt\alpha_0^I \tau + i\sqrt\sum_\frac\alpha_n^e^\cos n\sigma,\\ X^+(\sigma, \tau) &= \sqrt\alpha_0^+ \tau,\\ X^-(\sigma, \tau) &= x_0^- + \sqrt\alpha_0^- \tau + i\sqrt\sum_\frac\alpha_n^e^\cos n\sigma. \end$

Not all are independent. Some are zero (hence missing in the equations above), and the "minus coefficients" satisfy
:$\sqrt\alpha_n^- = \frac\sum_\alpha_^I\alpha_p^I.$
The quantity on the left is given a name,
:$\sqrt\alpha_n^- \equiv \fracL_n,\quad L_n = \frac\sum_\alpha_^I\alpha_p^I,$
the transverse Virasoro mode.

When the theory is quantized, the alphas, and hence the become operators.

See also

* Group cohomology
* Group contraction (Inönu–Wigner contraction)
*Group extension
*Lie algebra cohomology
*Ring extension

Remarks

Notes

References

Books

*
*
*
*
*
*
*
*
*
*
*
*
*
*

Journals

*
*
*
* (English translation)
*
*
* This can be found i
Kac–Moody and Virasoro algebras, A reprint Volume for Physicists
* (open access)
*
*
*
*

Web

*{{cite web, last=MacTutor, year=2015, website=MacTutor History of Mathematics, access-date=2015-03-08, url=http://www-history.mcs.st-and.ac.uk/history/Biographies/Schreier.html, title=Schreier biography

Lie groups
Quantum field theory
Lie algebras
Mathematical physics
Conformal field theory
String theory