Skip to main content

Modules and Vector Spaces as Algebraic Categories

An algebraic category is defined to be any full subcategory of Alg(Ω)\textbf{Alg}(\Omega), the category of all Ω\Omega-algebras. We can define various algebraic structures as algebraic categories. For example, in order to define Groups\textbf{Groups}, the category of Groups, we can specify Ω=e,m,i\Omega = \\{e,m,i\\} as the set consisting of three operators: one nullary, one unary, and one binary, such that, (G,e,m)(G,e,m) is a monoid and

m(i(x),x)=x=m(x,(i(x))m(i(x),x)=x=m(x,(i(x))

xG\forall x \in G, making Alg(Ω)=Groups\textbf{Alg}(\Omega) = \textbf{Groups} for Ω\Omega as specified1. However, an issue arises when considering vector spaces over a field and modules over a ring as the operators in Ω\Omega are defined to act on the elements of the underlying set, and not on elements outside of that set, say in a field or a ring. This issue arises when trying to define scalar multiplication for these structures. Consider a vector space VV over a field KK. We define a binary operator mm on VV called multiplication, then v(m):V2Vv(m):V^2\to V, where

v:ΩnNVVnv: \Omega \to \bigcup_{n \in \mathbf{N}} V^{V^{n}}

for N\mathbf{N} the natural numbers1. Now there is an issue since mm is not defined to operate on the elements of KK. What follows is an explanation as to how we can get around this issue and define vector spaces over a field KK and modules over a ring RR as algebraic categories.


We start with RR-modules. Let RR be a commutative ring with 11 and MM the underlying set of the module. Define the signature ΩM=aM,0M,iM\Omega_M = \\{a_M,0_M,i_M\\} where aMa_M is a commutative associative binary operator on MM, 0MM0_M\in M is a nullary operator such that

aM(u,0M)=u=aM(0M,u)a_M(u,0_M)=u=a_M(0_M,u)

uM\forall u\in M, iMi_M is an unary operator on MM such that

aM(iM(u),u)=0M=aM(u,iM(u))a_M(i_M(u),u)=0_M=a_M(u,i_M(u))

uM\forall u\in M. Now, in order to define scalar multiplication, we define the signature ΩR=R\Omega_R = R where each element of RR is considered as a unary operator on MM. However, we need to add some structure to ΩR\Omega_R since, in a ring, we are able to add and multiply elements of the ring to get new elements in the ring. To do this, we define a signature for an Ω\Omega-algebra on ΩR\Omega_R by ΩRring=a,0,i,m,1\Omega_{R}^{\text{ring}} = \\{a,0,i,m,1\\} with the same properties as the operators in ΩM\Omega_M above for a,0,ia,0,i and additionally, mm is an associative binary operator on ΩR\Omega_R such that

m(r,a(s,t))=a(m(r,s),m(r,t))m(r,a(s,t)) = a(m(r,s),m(r,t))m(a(r,s),t)=a(m(r,t),m(s,t))m(a(r,s),t) = a(m(r,t),m(s,t))

r,s,tΩR\forall r,s,t\in \Omega_R and 1ΩR1\in \Omega_R is a nullary operator such that

m(r,1)=r=m(1,r)m(r,1)=r=m(1,r)

rΩR\forall r\in \Omega_R. We can then make M=(M,hM)M=(M,h_M) into an ΩR\Omega_R-set (as in Example 3.10) by defining the map hM:ΩR×MMh_M:\Omega_R\times M\to M as hM(ω,p)=ωph_M(\omega,p)=\omega p, pM\forall p \in M 1. However, we require additional properties for the map hMh_M:

hM(1,u)=uh_M(1,u)=u

hM(ω,aM(u,v))=aM(hM(ω,u),hM(ω,v))h_M(\omega,a_M(u,v)) = a_M(h_M(\omega,u),h_M(\omega,v))

hM(a(ω,ω),u)=aM(hM(ω,u),hM(ω,u))h_M(a(\omega,\omega^{\prime}),u)=a_M(h_M(\omega,u),h_M(\omega^{\prime},u))

hM(m(ω,ω),u)=hM(ω,hM(ω,u))h_M(m(\omega,\omega^{\prime}),u)=h_M(\omega , h_M(\omega^{\prime},u))

ω,ωΩR\forall \omega,\omega^{\prime}\in\Omega_R u,vM\forall u,v\in M. In this way, scalar multiplication of an element of M by one of R is defined using the map hMh_M. So by letting the signature of the RR-module be Ω=ΩRΩM\Omega = \Omega_R \cup \Omega_ M, we obtain Alg(Ω)=Modules\textbf{Alg}(\Omega)=\textbf{Modules}, showing Modules\textbf{Modules} is an algebraic category.


Similarly, for a vector space VV over a field KK we can define a signature ΩV=aV,0V,iV\Omega_V = \\{a_V,0_V,i_V\\} where aVa_V is a commutative associative binary operator on VV, 0VV0_V\in V is a nullary operator such that

aV(u,0V)=u=aV(0V,u)a_V(u,0_V)=u=a_V(0_V,u)

uV\forall u\in V, iVi_V is an unary operator on VV such that

aV(iV(u),u)=0=aV(u,iV(u))a_V(i_V(u),u)=0=a_V(u,i_V(u))

uV\forall u\in V. To define scalar multiplication of elements of VV by those of KK, we define the signature ΩK=K\Omega_K = K in which each element of KK is a unary operator on VV. We add a field structure to ΩK\Omega_K by defining a signature for ΩK\Omega_K as ΩKfield=a,0,ia,m,1,im\Omega_{K}^{\text{field}}=\\{a,0,i_a,m,1,i_m\\} with the same properties as the operators in ΩV\Omega_V above for a,0,ia,0,i and additionally, mm is a commutative associative binary operator on ΩK\Omega_K such that

m(c,a(d,f))=a(m(c,d),m(c,f))m(c,a(d,f)) = a(m(c,d),m(c,f))m(a(c,d),f)=a(m(c,f),m(d,f))m(a(c,d),f) = a(m(c,f),m(d,f))

c,d,fΩK\forall c,d,f\in \Omega_K, 1ΩK1\in \Omega_K is a nullary operator such that

m(c,1)=c=m(1,c)m(c,1)=c=m(1,c)

cΩK\forall c\in \Omega_K and imi_m is an unary operator on ΩK\Omega_K such that

m(im(c),c)=0=m(c,im(c))m(i_m(c),c)=0=m(c,i_m(c))

cΩK\forall c\in \Omega_K. Next, make V=(V,hV)V=(V,h_V) into an ΩK\Omega_K-set by defining the map hV:ΩK×VVh_V:\Omega_K\times V\to V as hV(ω,u)=ωuh_V(\omega,u)=\omega u, uV\forall u \in V with the additional properties that

hV(ω,aV(u,q))=aV(hV(ω,u),hV(ω,q))h_V(\omega,a_V(u,q))=a_V(h_V(\omega ,u), h_V(\omega ,q))

hV(a(ω,ω),u)=aV(hV(ω,u),hV(ω,u))h_V(a(\omega,\omega^{\prime}),u) = a_V(h_V(\omega ,u), h_V(\omega^{\prime},u))

hV(m(ω,ω),u)=hV(ω,hV(ω,u))h_V(m(\omega,\omega^{\prime}),u) = h_V(\omega ,h_V(\omega^{\prime},u))

hV(1,u)=uh_V(1,u) = u

ω,ωΩK  u,qV\forall \omega,\omega^{\prime}\in\Omega_K\;\forall u,q\in V. So, for the signature Ω=ΩVΩK\Omega = \Omega_ V\cup\Omega_ K, we obtain Alg(Ω)=Vect\textbf{Alg}(\Omega)=\textbf{Vect}, showing that the category of vector spaces is an algebraic category.


  1. George Janelidze, Category theory: A first course, 2020. ↩︎ ↩︎ ↩︎