Tensor Product of V1 and V2 in Vector Space V: 0 Intersection Required?

Click For Summary

Discussion Overview

The discussion revolves around the conditions necessary for defining the tensor product of two subspaces V1 and V2 within a vector space V. Participants explore whether the intersection of V1 and V2 must be the zero vector for the tensor product to be defined, as well as the properties and existence of the tensor product itself.

Discussion Character

  • Debate/contested
  • Technical explanation
  • Mathematical reasoning

Main Points Raised

  • One participant questions if the intersection of V1 and V2 must be zero for their tensor product to be defined.
  • Another participant argues that the tensor product can be defined between a vector space and itself, suggesting that the intersection condition is not necessary.
  • A participant expresses confusion regarding the requirements for bilinearity in the construction of the tensor product, questioning if it is sufficient for the operation to apply to the entire Cartesian product of bases from the vector spaces.
  • Another participant asserts that the tensor product always exists, framing it as a construction rather than merely an operation, and emphasizes the bilinear property that allows for the determination of a unique tensor product.
  • There is a clarification that the tensor product includes not just ordered pairs but also linear combinations of such pairs, indicating a broader structure than initially suggested.

Areas of Agreement / Disagreement

Participants do not reach a consensus on whether the intersection of V1 and V2 must be zero for the tensor product to be defined. Multiple viewpoints regarding the existence and properties of the tensor product are presented, leading to an unresolved discussion.

Contextual Notes

Participants express varying interpretations of the conditions required for the tensor product, including the role of bilinearity and the nature of the elements in the Cartesian product of vector spaces. There is also a discussion on the dimensionality of the tensor product space versus the elements formed from the bases of the original spaces.

Werg22
Messages
1,431
Reaction score
1
If V1 and V2 are both subspaces of a vector space V, then in order for their tensor product to be defined, does the intersection of V1 and V2 have to be 0?
 
Last edited:
Physics news on Phys.org
No, infact an important example of tensor product is that between a vector space and itself.
V^2=VxV

One possible definition of tensor product is that it maps two vector spaces onto ordered pairs.
UxV contains ordered pairs (u,v) with u from U and v from V.
Edited to add. UxV also contains linear combinations of ordered pairs.
 
Last edited:
I see. But I have to say, I'm puzzled about this: according to one definition I have, the operation on pairs (u,v), u belonging to U and v belonging to V, that is used to construct the tensor product has to be bilinear. But that certainly isn't a sufficient condition, is it? One must also have that the operation in question applied to the whole of the cartesian product of a basis of U and a basis of V forms a basis for the tensor product, right? Does the tensor product always exist (i.e. is there always such a bilinear operation and vector space for any ordered pair of vector spaces)?
 
The tensor product always exists.
Existence is not a problem because tensor product is more a construction than an operation. That is the operation always endows the product with the required properties.
lets assume U and V are vector spaces over the same field F.
we want
(a1u1+a2v2,b1v1+b2v2)=a1b1(u1,v1)+a1b2(u1,v1)+a2b1(u2,v1)+a2b2(u2,v2)
so in declaring the tensor product has this property we determin the tensor product uniquely.
Since the tensor product is bilinear we can obtain a basis from the bases of the spaces used to construct it.
if
{u(i)} is a basis for U
and
{v(j)} is a basis for V
{u(i),v(j)} is a basis for UxV

it is important to notice that
X is an element of UxV does not mean X is of the form (u,v)
UxV contains all elements of the form (u,v), but also all linear combinations of such terms.
By a counting argument if
dim(U)=m
dim(V)=n
dim(UxV)=mn
dim(elements of the form (u,v))=m+n
I see I was unclear above I said
UxV contains ordered pairs (u,v) with u from U and v from V
which is true but it contains more ie sums of such.
 

Similar threads

  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 10 ·
Replies
10
Views
3K
  • · Replies 32 ·
2
Replies
32
Views
4K
  • · Replies 4 ·
Replies
4
Views
4K
  • · Replies 4 ·
Replies
4
Views
4K
  • · Replies 13 ·
Replies
13
Views
4K
  • · Replies 3 ·
Replies
3
Views
9K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 18 ·
Replies
18
Views
4K
  • · Replies 1 ·
Replies
1
Views
2K