Scribd
Upload
15 views3 pages

Projective

A brief introduction to projective spaces in Mathematics.

Uploaded by

Momin Aziz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
15 views3 pages

Projective

A brief introduction to projective spaces in Mathematics.

Uploaded by

Momin Aziz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

18.

704: Projective Space


David Glasser

2/18/2005

1 Projective Space
Projective n-space over the field F , or Pn (F ) for short, is defined to be the set
of lines through the origin in F n+1 (a more technical definition is below). Why
n + 1 and not n? Let’s look at a simple example: P1 (R), which is often referred
to as RP 1 . This is the space of all lines through the origin in the plane. Now, if
we draw the vertical line at x = 1, we see that just about every line through the
origin intersects this line in exactly one point, and in fact we can (almost!) make
a bijection between points on this 1-dimensional line and “points” (which look
like lines) in P1 (R). In fact, the point (1, m) corresponds to the line y = mx.
Since the line x = 1 is clearly 1-dimensional, and we have this very natural map
between the set of lines through the origin of R2 and that line, it makes sense to
consider this set of lines to be 1-dimensional, so we call it P1 (R). I said “almost”
above, because the vertical line through the origin (x = 0) doesn’t intersect the
line x = 1 at all. But just have a single extra point doesn’t really seem like it’ll
change the “dimension” of the space (and yes, this seems like a bad proof, but
that’s because it isn’t a proof, it’s just a justification for a definition). This line
contains all points of the form (0, a). Thus, we see that every line through the
origin (ie, every element of P1 (R)) contains exactly one point of the form (1, m)
or (0, 1).
Similarly, if we look at P2 (R), we want to consider all of the lines through
the origin in R3 . Most of these contain exactly one point of the plane x = 1.
The only ones that don’t lie entirely in the plane x = 0. But then most of them
contain exactly one point in the plane y = 1. The only line that is the single
line defined by has x = 0 and y = 0 everywhere. So we see that every element
of P2 (R) has exactly one point of the form (1, a, b) or (0, 1, c) or (0, 0, 1).
The neat thing to notice here is that in this example, we started by looking
at the lines that intersect the right-hand plane x = 1; these correspond to points
of the form (1, a, b) for all real numbers a and b, so it’s essentially a copy of the
plane R2 . And what do the “weird” lines that don’t intersect that plane look
like? Well, they’re all the lines through the origin that lie in the plane x = 0.
But if you just pretend that the plane x = 0 is the entirety of known space,
all we’re talking about is the set of lines through the origin in R2 , which is to
say P1 (R)! That is to say, the interesting intersection points we’ve noted above

1
in Pn (R) consist of a copy of Rn with “1” glued onto the front, and a copy of
Pn−1 (R) with “0” glued onto the front.

2 How It’s Actually Defined


The definition of n-dimensional projective space over an arbitrary field F is
still just “the set of n + 1-dimensional lines through the origin”, but what does
that mean? Well, consider a point (f0 , f1 , . . . , fn ) in n + 1-dimensional F -space
(so each fi ∈ F ). The points on the same line through the origin are just
(af0 , af1 , . . . , afn ) for each a ∈ F . So we can define the points of Pn (F ) to be
the equivalence classes of vectors (f0 , f1 , . . . , fn ), where two vectors g and h are
considered to be equivalent if there is a scalar a ∈ F such that g = ah. We
require both that the point is not the origin (0, 0, . . . , 0), and that the scalar
a 6= 0; otherwise the origin would end up on all of the lines, and we wouldn’t
have an equivalence relation. For example, back in P1 (R), we have that (1, 2)
and (3, 6) are equivalent and thus fall in the same “point” of P1 (R), but (1, 2)
and (2, 7) are on different “points”.
As it turns out, our decomposition argument still holds! That is, if there’s
any point (f0 , f1 , . . . , fn ) in a line through the origin with f0 6= 0, then since
F is a field we can see that f0−1 (f0 , f1 , . . . , fn ) = (1, f0−1 f1 , . . . , f0−1 fn ) is the
unique point on the line that it determines with 1 for its first coordinate. And
on the other hand, if 0 is the first coordinate for any point on a line (except
for the origin itself), then 0 is the first coordinate for every point on that line.
The lines satisfying this property are the lines through the origin in the n-
dimensional subspace defined by f0 = 0, so it’s basically just a copy of Pn−1 (F )
with a zero glued onto the front of every point.
So every line in F n+1 (ie, every point of Pn (F )) passes through exactly one
point of the form

(1, f1 , f2 , . . . , fn )

or

(0, 1, f2 , . . . , fn )

or

(0, 0, 1, . . . , fn )

or . . . or . . .

(0, 0, 0, . . . , 1)

We can use this to count the number of element of Pn (F ) where F is the finite
field with q elements (this argument works whether this q is prime and it’s one
of the nice fields Z/pZ or whether q is a prime power and it’s one of the weirder
finite fields). There are q n elements of the first type here, since each fi can be

2
any of q elements. The rest of the elements correspond to elements of Pn−1 (F ).
So we’ve shown that |Pn (Fq )| = q n + Pn−1 (Fq ) . As a base case, P0 (F ) = 1
for any field: this is essentially the scalars modulo the scalars, or the set of
lines in a line, and so there is only one. Unrolling the last formula, we get
n+1
|Pn (Fq )| = q n + q n−1 + · · · + q 0 ; this formula is equal to q q−1−1 as long as q 6= 1,
which it will be because there are no one-element fields. So this is the number
of points in Pn (Fq ). (Note that if you had insisted on plugging in 1 for q before
making the last transformation, you would have gotten n + 1; this means that
this formula gives us what is called a “q-analog” for n + 1.)

You might also like