# A Infinite Basis and Supernatural Numbers

1. Sep 21, 2017

### FallenApple

I've read that it is unsatisfactory to consider infinitely many basis vectors to span an infinite dimensional space. For example, for the infinite dimensional Hilbert space, {e1,e2,e3.......} we could use this to make an arbitrary infinite tuple (a,b,c,...). If this is looked down upon, then why are Supernatural Numbers, which are the product of infinitely many primes, considered ontologically valid ?

Surely if we do a transformation, we could express those supernatural numbers as vectors, since its just converting multiplication to vector addition.

2. Sep 21, 2017

### TeethWhitener

Citation needed. For example, the fact that the infinite set of basis vectors $e^{ikx}, k\in \mathbb{Z}$ spans the space of periodic functions is the cornerstone of Fourier theory. Also, the axiom of choice is equivalent to the statement that every vector space--even if it has uncountably infinite dimensions--has a basis.

3. Sep 21, 2017

### FallenApple

http://www.math.lsa.umich.edu/~kesmith/infinite.pdf

4. Sep 21, 2017

### Staff: Mentor

For the question in your quoted text, I don't see the problem. It seems to me that (1, 1, 1, ..., ) + (2, 2, 2, ... , ) + (3, 3, 3, ... , ) = (6, 6, 6, ... , ).

IOW, component-wise vector addition, even though the vectors belong to an infinite-dimension vector space.

I wouldn't consider the sum of 3 vectors to be an infinite sum, being as there are only three of them being added. I will take a look at the link you gave to see if there is some context that is missing from what you quoted.

5. Sep 21, 2017

### Staff: Mentor

IMO, the author of the PDF from which you quoted is quite sloppy.
In Example 2 she says
Later in the same example, she says
In the sentence above she meant that v cannot be written as a finite linear combination, but neglected to write it.

Finally, from the text you quoted, she calls (1, 1, 1, ..., ) + (2, 2, 2, ..., ) + (3, 3, 3, ..., ) an infinite sum even though there are only three vectors. She asks what would be the sum of these vectors, a question she answers more generally in the first part of Example 2.

6. Sep 21, 2017

### FallenApple

Ok that makes sense. I was thinking somewhere alone those lines of totaling up the coefficients per coordinate, which is example 2. That makes sense.

In an infinite vector space, it is a bit self explanatory that infinite combinations are needed unless the resultant vector is non zero in finitely many places; of which we still get around this by placing 0 in front of the remaining basis vectors. So it's strange for her to attach significance to the fact that some vectors wouldn't be able to be written as a finite linear combo, given the context of infinite dimensions.

7. Sep 21, 2017

### Staff: Mentor

Yes. Although it's obvious that you can't generate a particular vector with a finite linear combination of basis vectors, it seemed very sloppy to me to omit "finite".
And her example didn't make any sense, namely the one with (1, 1, ...) + (2, 2, ...) + (3, 3, 3, ...) and calling that an infinite sum.

8. Sep 21, 2017

### FallenApple

Thats true. It's a tri-nary sum producing a single answer infinitely many separate times , which is not an infinite sum of a sequence.

9. Sep 25, 2017

### Erland

I think there is a typo in the text quoted by OP, so that it should be "(1, 1, 1, ...) + (2, 2, 2, ...) + (3, 3, 3, ...) + ... "

10. Sep 25, 2017

### Staff: Mentor

Maybe, but I would be embarrassed to put out something with such a glaring error.