1 Posets: 1.1 Extreme Elements
1 Posets: 1.1 Extreme Elements
1 Posets
Last week we defined maximal, minimal, greatest, and least elements of a poset. We will explicitly
determine the useful properties of these elements.
The proofs below will generally only present an argument for maximal and greatest elements; the
same argument can be modified very slightly to work for minimal and least elements.
Proposition 1. If (S, ) is a nonempty finite poset, then S has at least one maximal element (and
at least one minimal element).
Proof. Let x0 S. If x0 is maximal, than S clearly has a maximal element; otherwise, there
must be an x1 6= x0 such that x0 x1 . Now let us subject x1 to the same consideration; if x1 is
maximal, we are done, otherwise there is an x2 6= x1 such that x1 x2 . Iterating this process must
necessarily have one of two results: either there will be some xi which is maximal, or the sequence
x0 x1 x2 x3 is infinite. If some element is maximal, we will achieve our goal; we shall
show that the second situation is impossible.
Suppose we do have such an infinite sequence of elements of S, so that each xi xi+1 and xi 6= xi+1 .
Then, since this sequence is infinite and its elements are drawn from the finite set S, the same
element must appear twice in the sequence (a massive understatement, but all we need). Thus we
may assume there are i < j such that xi = xj . By transitivity, since xi xi+1 xj1 , it
follows that xi xj1 . However, since xi = xj and xj1 xj , it is also true that xj1 xi . This
can only be true, by antisymmetry, if xi = xj1 . However, we also know that xj1 6= xj = xi ,
leading to a contradiction.
Note that this is emphatically not true for infinite sets S; they may have maximal elements, as
([0, 1], ) does, or need not, as exhibited by (Z+ , ).
Proposition 2. If a poset (S, ) has a greatest element (or a least element) then it has exactly
one maximal (minimal) element. Furthermore, its greatest (least) element is exactly the unique
maximal (minimal) element.
Proof. First, we shall prove that a greatest element must be maximal. This is easy: suppose x
is nonmaximal: then there is some y 6= x such that x y. By antisymmetry, since x and y are
nonequal, it is clear that y x. Thus, since there is an element y which x is not greater than, x is
not a greatest element.
Now, let us show that a poset with two or more maximal elements cannot have a greatest element.
Suppose x is the greatest element of a poset S with at least two maximal elements. Since there
are two or more maximal elements, there is at least one maximal element not equal to x, which
we shall call y. Since x 6= y and y is maximal, y x, which contradicts our presumption that x is
greater than every element of S.
Corollary 1. A finite poset (S, ) has a greatest element (or a least element) if and only if it has
exactly one maximal (minimal) element. Furthermore, its greatest (least) element is exactly the
unique maximal (minimal) element.
Proof. One direction follows from the proposition above; all we need to show now is that if there
is a unique maximal element of a poset, it is in fact the greatest element. Suppose x is the unique
maximum of S. If x is not the greatest element of S, then there is some y0 such that y0 x. By
maximality of x, x y0 , so x and y0 are incomparable. Since y0 is not maximal, there is a y1 6= y0
such that y0 y1 . If y1 x, then y0 x by transitivity, so y1 x. However, by maximality
of y1 , x y1 , so x and y1 are incomparable. We may proceed with this sequence, producing a
sequence y0 y1 y2 , which cannot terminate since all the yi are distinct from x and thus
nonmaximal. We saw previously that such a sequence is impossible, so such a y0 does not exist,
and thus x is in fact the greatest element of S.
Note that the above may again be untrue when S is infinite. Consider S = Z {}, with the
ordering defined by a b when a b Z, and a b if either a = or b = . Then is
incomparable to everything and thus the sets unique maximal element, but not a greatest element.
We can draw a poset (preferably) by placing x below y if x y and drawing an edge from x to
y. This can get pretty messy if we tried to do it on a moderately large poset. For instance, if we
discussed the divisibility relation on {1, 2, 3, 4, 5, 6, 9, 35}, we get the moderately ugly diagram:
4 6 9 35
2 3 5
This has a lot of redundant lines on it. If instead we only connect a to b when a b and there is
no c not such that a c b, we get the cleaner diagram:
4 6 9 35
2 3 5
On a Hasse diagram, we can come up with easy visual interpretations for every aspect of a poset:
maximal elements are ones without upwards-facing branches; minimal elements are those without
downwards-facing branches, a chain is a collection of vertices in linear order.
Our explorations from here will be into chains and antichains: finding maximal chains and an-
tichains, and the significance thereof.