Difference between properties of 3D unit vectors and 4D unit vectors(in the context of quaternions)

114 Views Asked by At

In the book I'm reading about quaternions, the properties of operations with imaginary values $i,j,k$ are compared to properties of the cross product of Cartesian vectors $\mathbf{i}, \mathbf{j}, \mathbf{k}$. E.g. $$ij=k\\\mathbf{i} \times \mathbf{j}=\mathbf{k}.$$ So then we have $$\mathbf{i}\times\mathbf{j}=\mathbf{k}\\ \mathbf{i}\times\mathbf{k}=\mathbf{-j}.$$ And then the author assumes that these vectors obey distributive and assocaitive axioms to prove their imaginary properties: $$\mathbf{iij}=\mathbf{ik}=\mathbf{-j} \\ \mathbf{ii}=\mathbf{i}^2=-1$$ But what I don't understand is why these vectors would obey associative axiom, since $\mathbf{i}\mathbf{i}\mathbf{j}$ are still cross products, or at least I assume so. And what about $\mathbf{i}\times\mathbf{i}=0$? Does this mean that these are not actually Cartesian vectors? I suppose, that I am missing some key points in this explanation...

2

There are 2 best solutions below

0
On BEST ANSWER

You are confusing cross product with quaternion product. The notation $i \times j = k$ is referring to cross product, while the notation $i j = k$ is referring to (pure) quaternion product. Let me explain the later:

Multiplication of two pure quaternions $A$ and $B$ can be written in vector form as:

$$A B = - A \cdot B + A \times B$$

Where a dot product is involved as well as cross prpduct. You can check that by calculanting $A B$ using the quaternion multiplication rules and grouping terms.

Using that product you can check that $i i = -1$ and that $i k = -j$ and also that it is associative is associative like $i i j = -j$

2
On

Quaternions have only 4 letters to write (1,i,j,k), so you can check by hand that multiplication is assiociative.

However, multiplication of quaternions is not commutative.

Problem with 3 dimensions, 5 dimensions, 7 dimensions etc is that there is no way to define a field structure on 3/5/7/9...etc dimensional real space, even if we don't require commutavitity. On 1 dimension we have natural field structure, on 2 dimensions the only possible field structure is complex numbers. On 4/6/8/10...etc dimensions, we can define structure of field but only noncommutative. When dim=4, these are quaternions.

It is not true that i*i=0, I don't understand what do you mean.