The Student Room Group

Vector spaces span

Can someone explain what span actually is. Ive looked at definitions but still dint get it.

Posted from TSR Mobile

Scroll to see replies

Reply 1
Original post by cooldudeman
Can someone explain what span actually is. Ive looked at definitions but still dint get it.

The span of vectors (1,0,0) and (0,1,0) is the subspace consisting of vectors of the form (a,b,0). More generally, the span of {v1, v2, …, vn} is the space of vectors {a1 v1 + a2 v2 + + an vn}.
Reply 2
Original post by Smaug123
The span of vectors (1,0,0) and (0,1,0) is the subspace consisting of vectors of the form (a,b,0). More generally, the span of {v1, v2, …, vn} is the space of vectors {a1 v1 + a2 v2 + + an vn}.


So its just the constants being multiplied to each vector in the set?
Could you help me on this question please. I dont know what an ordered basis is.

The eg question

Posted from TSR Mobile
Reply 3
Original post by cooldudeman
So its just the constants being multiplied to each vector in the set?
Could you help me on this question please. I dont know what an ordered basis is.

The eg question

Posted from TSR Mobile

Your "proof" that the span is R^3 isn't really enough - you'd have to show that for every a,b,c it is possible to find the λ1,λ2,λ3\lambda_1, \lambda_2,\lambda_3. However, because we're in a finite-dimensional vector space, all we need to do is show one of spanning or linearly independent; then because the space is of dimension 3, and we have three spanning/LI vectors, they must necessarily be LI/spanning.

The basis is ordered if it has an ordering. That is, you just impose an order on a basis and it becomes an ordered basis. That's all there is to it.

You may also be interested to know that you draw your lambdas left-to-right reversed from the conventional way.
Reply 4
Original post by Smaug123
Your "proof" that the span is R^3 isn't really enough - you'd have to show that for every a,b,c it is possible to find the λ1,λ2,λ3\lambda_1, \lambda_2,\lambda_3. However, because we're in a finite-dimensional vector space, all we need to do is show one of spanning or linearly independent; then because the space is of dimension 3, and we have three spanning/LI vectors, they must necessarily be LI/spanning.

The basis is ordered if it has an ordering. That is, you just impose an order on a basis and it becomes an ordered basis. That's all there is to it.

You may also be interested to know that you draw your lambdas left-to-right reversed from the conventional way.


I'm sorry but im not understanding what ypu are saying. I did show that it is LI and you said 'all we need to do is show one of spanning or linearly independent'.

And yeah its just a habit.

Posted from TSR Mobile
Reply 5
Original post by cooldudeman
I'm sorry but im not understanding what ypu are saying. I did show that it is LI and you said 'all we need to do is show one of spanning or linearly independent'.

And yeah its just a habit.

Posted from TSR Mobile

Yes, you showed it was LI - but before that, you "showed" that it was spanning, and that "showing" was very much incomplete. If you'd just left out that "spanning" bit, I wouldn't have complained!
Reply 6
If you're wondering what the point of ordering a basis is, it's because when you move onto change of basis matrices, they are only well-defined if the basis is ordered. Essentially, if you don't fix the basis when dealing with change of basis matrices, permuting the elements of a basis would permute the rows (or columns) of the change of basis matrix. So when a change of basis matrix is given, it has to be with respect to the bases ordered.
Reply 7
Original post by Smaug123
Yes, you showed it was LI - but before that, you "showed" that it was spanning, and that "showing" was very much incomplete. If you'd just left out that "spanning" bit, I wouldn't have complained!


Oh I get you. Say if I were to just show that it spans R3. What would I do? It kinda made sense to me - what I did. Ive never really seen and eg

Posted from TSR Mobile
Reply 8
Original post by cooldudeman
Oh I get you. Say if I were to just show that it spans R3. What would I do? It kinda made sense to me - what I did. Ive never really seen and eg

To show that it spans R^3, you need to pick an arbitrary vector (a,b,c) in R^3, and show that it can be made as a linear combination of the spanning vectors. That is, you need to find the λi\lambda_i such that λ1v1+λ2v2+λ3v3=(a,b,c)\lambda_1 v_1 + \lambda_2 v_2 + \lambda_3 v_3 = (a,b,c).
Reply 9
Original post by Smaug123
To show that it spans R^3, you need to pick an arbitrary vector (a,b,c) in R^3, and show that it can be made as a linear combination of the spanning vectors. That is, you need to find the λi\lambda_i such that λ1v1+λ2v2+λ3v3=(a,b,c)\lambda_1 v_1 + \lambda_2 v_2 + \lambda_3 v_3 = (a,b,c).

Like that?


Posted from TSR Mobile
Reply 10
Original post by cooldudeman
Like that?

Yep, exactly. (I'm assuming your arithmetic is correct.) As you can see, linear independence is often easier than spanning to prove.
Reply 11
Original post by Smaug123
Yep, exactly. (I'm assuming your arithmetic is correct.) As you can see, linear independence is often easier than spanning to prove.


Thanks. Lastly how do I impose an order on the basis? What does that mean?

Posted from TSR Mobile
Reply 12
Original post by cooldudeman
Thanks. Lastly how do I impose an order on the basis? What does that mean?

Posted from TSR Mobile

You just announce that "the first vector in the basis is <this one>, the second is <this one> and the third is <this one>". That's it.
Reply 13
Original post by Smaug123
You just announce that "the first vector in the basis is <this one>, the second is <this one> and the third is <this one>". That's it.


Thank you

Posted from TSR Mobile
Reply 14
Original post by Smaug123
You just announce that "the first vector in the basis is <this one>, the second is <this one> and the third is <this one>". That's it.


Hi sorry to bother again but what does it mean by extend? The span of those two vectors give already equals R^4 doesnt it.

Posted from TSR Mobile
Reply 15
Original post by cooldudeman
Hi sorry to bother again but what does it mean by extend? The span of those two vectors give already equals R^4 doesnt it.

Posted from TSR Mobile


Certainly not. R4\mathbb{R}^4 is four dimensional, so any set of vectors that spans it must contain at least four vectors.
Reply 16
Original post by BlueSam3
Certainly not. R4\mathbb{R}^4 is four dimensional, so any set of vectors that spans it must contain at least four vectors.


Totally forgot that. So if I extend it by adding for eg (1,1,1,1) and (2,2,2,2) to the given set, would that work?

Posted from TSR Mobile
(edited 10 years ago)
Reply 17
Original post by cooldudeman
Totally forgot that. So if I extend it by adding for eg (1,1,1,1) and (2,2,2,2) to the given set, would that work?

Posted from TSR Mobile

No. The resulting set will not be linearly independent, and hence it won't span. (Can you see why?)
You can make your job a lot easier, because you already know of a collection of four basis vectors. It's likely that two of them will be independent of the two already given. (It's not certain, I believe, but it's likely.)
Reply 18
Original post by Smaug123
No. The resulting set will not be linearly independent, and hence it won't span. (Can you see why?)
You can make your job a lot easier, because you already know of a collection of four basis vectors. It's likely that two of them will be independent of the two already given. (It's not certain, I believe, but it's likely.)


Oh damn. Because 2 (1111) is 2222. I'm guessing (1110) and (0111) would work after considering if they can be turned into the ones given. Pretty sure they cant because of those zeros.
Reply 19
Original post by cooldudeman
Oh damn. Because 2 (1111) is 2222. I'm guessing (1110) and (0111) would work after considering if they can be turned into the ones given. Pretty sure they cant because of those zeros.

They do work, yes. You should prove it, though - which involves proving either that any linear combination which is 0 is the zero linear combination, or by showing that they span (which often is easiest by showing that you can make each of (1,0,0,0), (0,1,0,0),… out of them).

Quick Reply

Latest