Basic Concepts and Notation

Gabriel Robins

"When I use a word," Humpty Dumpty said, in a rather scornful tone,

"it means just what I choose it to mean -- neither more nor less."

A set is formally an undefined term, but intuitively it is a (possibly empty) collection of

arbitrary objects. A set is usually denoted by curly braces and some (optional) restrictions.
Examples of sets are {1,2,3}, {hi, there}, and {k | k is a perfect square}. The symbol ∈ denotes

set membership, while the symbol ∉ denotes set non-membership; for example, 7∈{p | p
prime} states that 7 is a prime number, while q∉{0,2,4,6,...} states that q is not an even number.

Some common sets are denoted by special notation:

The natural numbers: ˆ = {1,2,3,...}

The integers: 8 = {...,-3,-2,-1,0,1,2,3,...}

The rational numbers: 9 = { ab | a,b ∈ 8, b≠0}

The real numbers: B = {x | x is a real number}

The empty set: Ø = {}

When only the positive elements of a numerical set are sought, a superscript "+" may be used to
denote this. For example, 8+ = ˆ denotes the positive integers (i.e., the natural numbers), B+

denotes all the positive reals, and more generally, S+ = {s∈S | s>0}.

The logical symbol | (pronounced "such that", and sometimes also denoted as ∋) denotes a
conditional (which usually follows this symbol). The logical symbol ∀ (pronounced "for all")
denotes universal quantification. For example, the formula "∀x∈B x ≤ x2 +1" reads "for all

x a member of the real numbers, x is less than or equal to x-squared plus one" (i.e., no real
number is greater than one more than its own square). The logical symbol ∃ (pronounced "there

1
exists") denotes existential quantification. For example, the formula "∃x∈8 | x2 =5x" states

that there exists an integer whose square is equal to 5 times itself (i.e., x is either 5 or 0) . These
connectives may be composed in more complicated formulae, as in the following example: "∀x∈8

∃y∈8 | y>x" which states that there is no largest integer.

The logical connective ¤ (pronounced "and") is a boolean-valued function that yields true if

and only if both of its two arguments are true. The logical connective ⁄ (pronounced " or") is a

boolean-valued function that yields true if and only if one or more of its two arguments are true.
The symbol ⇒ (pronounced "implies") denotes logical implication; that is, A⇒B means that B is
true whenever A is true; for example, "1 < x < y ⇒ x3 < y3 ". The symbol ⇔ (pronounced "if and

only if", and sometimes written as "iff") denotes logical equivalence; that is, A⇔B means that
B is true if and only if A is true. More formally, A⇔B means A⇒B ¤ B⇒A; an example is

"min(x,y)=max(x,y) ⇔ x=y". It is easily shown that A⇒B implies ¬ B ⇒ ¬ A, where ¬ denotes

logical negation.

A set S is a subset of a set T (denoted S@T) if every element that is a member of S is also a

member of T. More formally, S@T ⇔ (x∈S ⇒ x∈T). A set S is a proper subset of a set T

(denoted S2T) if S is a subset of T, but S and T are not equal. More formally, S2T ⇔ (S@T ¤

S≠T). Clearly every set has the empty set and the set itself as subsets (i.e., ∀S Ø@S ¤ S @ S).

Two sets are equal if and only if each is a subset of the other, i.e., S=T ⇔ (T@S ¤ S@T).

The union of two sets S and T (denoted S∪T) is the (duplicate-free) "merger" of the two
sets. More formally, S∪T={x | x∈S ⁄ x∈T}. The intersection of two sets S and T (denoted
S∩T) is their greatest common subset. More formally, S∩T={x | x∈S ¤ x∈T}. Two sets are

said to be disjoint if their intersection is empty (i.e., S and T are disjoint ⇔ S∩T=Ø).

2
 The union and intersection operators are commutative (S∪T=T∪S, and S∩T=T∩S),

associative S∪(T∪V) = (S∪T)∪V, and S∩(T∩V) = (S∩T)∩V, and distribute over each other

S∪(T∩V)=(S∪T)∩(S∪V), and S∩(T∪V) = (S∩T)∪(S∩V). Absorption occurs as follows:

S∪(S∩T)=S, and S∩(S∪T)=S. The complement of a set S (with respect to some universe set)

is the collection of all elements (in the universe set) that are not in S , and is denoted S ' (or by S
with a horizontal bar over it). More formally, S' = {x | x∉S}.

A set is said to be closed under a given operation if the operation preserves membership in
the set. Formally, S is said to be closed under an operation ◊ iff x◊y∈S ∀ x,y∈S. For

example, the set of integers 8 is closed under addition (+), since the sum of any two integers is
also an integer; on the other hand, 8 is not closed under division.

A relation over a domain D is a set of ordered pairs, or more generally, a set of ordered k-
tuples. For example, the relation ♥ defined as {(a,1), (b,2), (b,3)} means that "a" is related to 1,

and "b" is related to both 2 and 3; this may also be written as a♥1, b♥2, and b♥3. A more

familiar relation (over 8) is the "less than" relation, often denoted as < , which actually consists of

an infinite set of ordered pairs such that the first element is less than the second; that is, the <
relation is formally defined to be the set {(x,y) | x,y∈8, y>x}.

A relation is said to be reflexive if every element in the relation domain is also related to
itself; i.e., ♥ is reflexive iff x♥x ∀x∈D. A relation is said to be symmetric if it commutes; i.e.,
♥ is symmetric iff x♥y ⇒ y♥x. A relation is transitive if x♥y ¤ y♥z ⇒ x♥z. For example,
the subset operator is reflexive (S@S), and transitive (S@T ¤ T@V ⇒ S@V), but not

symmetric. The transitive closure of a relation is the extension of that relation to all pairs that
are related by transitivity; i.e., the transitive closure of ♥ contains all pairs of ♥, as well as all pairs
(x,y) such that for some finite set of elements d1, d2, d3, . . . , dk in ♥'s domain, all of x♥d1,
d1 ♥d2 , d2 ♥d3 , ..., dk-1♥dk , dk ♥y hold. Put another way, the transitive closure ♠ of a relation ♥

3
is the smallest relation containing ♥ but which is still closed under transitivity (i.e., satisfying x♠y
¤ y♠z ⇒ x♠z). For example, the predecessor relation ‡ may be defined as {(x,x-1) | x∈ 8 }, and

the transitive closure of ‡ is the > relation. Similarly, the symmetric closure of a relation is the
smallest containing relation that is closed under symmetry, etc.

"I don't understand you," said Alice. "Its dreadfully confusing!"

A relation that is reflexive, symmetric, and transitive is called an equivalence relation; an

example of this is the familiar equality relation = . It is easy to show that an equivalence relation

partitions its domain into mutually disjoint subsets, called equivalence classes. A special kind

of relation is called a graph, where the domain elements are called nodes and the relation pairs are

referred to as edges. For example, one simple graph may be {(a,b), (a,c), (b,d)}. Graphs are

often drawn using ovals to represent the nodes and arcs to represent the edges. A graph is said to

be undirected when the relation that it represents is symmetric, and directed otherwise. The

transitive closure of an undirected graph is an equivalence relation where the equivalence classes

correspond to the connected components of the graph.

"You'll get used to it in time," said the Caterpillar;

An important property of set operations is the analogue of DeMorgan's Law: (S∪T)' =

S'∩T' and (S∩T)' = S'∪T'. These equalities follow from DeMorgan's law for classical logic:
if X and Y are boolean variables, then (X¤Y)'=X'⁄Y' and (X ⁄ Y)'=X' ¤ Y' always hold. This

is an artifact of the elegant duality between the operators ¤ and ⁄ in the prepositional calculus: if

one starts with a true theorem (logical preposition) and simultaneously replaces all the ¤ 's with

⁄'s, and all the ⁄'s with ¤'s, the result is also a true theorem.

The difference between two sets S and T is the set containing all elements that are in S but
not in T. More formally, S-T = {s | s∈S ¤ s∉T} = S∩T'. The symmetric difference between

two sets S and T is defined as S∪T - S∩T. The cross-product of two sets S and T, denoted by

S×T, is the set of all ordered pairs whose first element comes from S and whose second element

4
comes from T. More formally, S×T = {(s,t) | s∈S, t∈T}. For example, {1, 2, 3} × {a,b} =

{(1,a),(1,b),(2,a),(2,b),(3,a),(3,b)}. A set may be crossed with itself a number of times: Si =

S×Si-1 where S1 = S.

The cardinality (or size) of a finite set is defined to be the number of elements in it, and is

denoted by vertical bars placed around the set. For example, |{a,b,c}|=3, |{p | p a prime less than

20}|=8, and |Ø|=0. The powerset of a set S (denoted 2S ) is the collection of all subsets of S;
more formally, 2S ={T | T@S}. If S is finite, the cardinality of its powerset is precisely 2 raised to

the cardinality of S (i.e., |2S |=2|S|); this is true because each subset of S can be represented by a
unique sequence of |S| binary digits (where 1 represents membership of the corresponding element

in the subset, and 0 represents non-membership). Since there are 2 |S| such sequences, and each

corresponds to a unique subset of S, there must be 2|S| subsets of S.

A function ƒ which maps a set S to a set T (denoted ƒ:S→T) is said to be one-to-one (or

injective) if any two distinct elements in S are mapped to distinct elements in T. More formally,
ƒ is injective iff a,b∈S ¤ a≠b ⇒ ƒ(a)≠ƒ(b). In this context S is said to be the domain of ƒ,

while T is said to be the range of ƒ. Intuitively, a function is one-to-one if no two distinct

elements in its domain are mapped to the same element in its range. For example, ƒ:8→8 defined

as f(x)=2x is one-to-one, while g(x)=x2 is not, since g maps both -2 and 2 to 4.

The rate of growth of numerical functions is often described asymptotically. This is

especially useful when discussing the time or space complexities of algorithms, since it enables

implementation- / hardware-independent comparisons of the relative merits of specific algorithms.

A function ƒ(x) is said to be O(g(x)) (pronounced "big-oh of g(x)") if for some positive constant

c we have c • ƒ(x) < g(x) for all but a finite number of values of x. If other words, g(x) is an

upper bound for ƒ(x) in the limit, modulo a multiplicative constant. More formally, this may be
expressed as ƒ ( x ) = O(g(x)) ⇔ ∃ c∈B+ ∃ x'∈ B+ ∋ ƒ(x) < c • g(x) ∀ x > x ' . Similarly, a

5
function ƒ(x) is said to be Ω(g(x)) (pronounced "omega of g(x)") if for some positive constant c

we have c • ƒ(x) > g(x) for all but a finite number of values of x. If other words, g(x) is a lower

bound for ƒ(x) in the limit, modulo a multiplicative constant. More formally, this may be
expressed as ƒ(x) = Ω(g(x)) ⇔ ∃ c∈B+ ∃ x'∈B+ ∋ ƒ(x) > c • g(x) ∀ x > x'.

Finally, a function ƒ(x) is said to be Θ(g(x)) (pronounced " theta of g(x)") if both the

relations ƒ(x) = Ω(g(x)) and ƒ(x) = O(g(x)) hold; in other words, ƒ(x) and g(x) have the same

asymptotic growth rate, modulo a multiplicative constant, so that each of ƒ(x) and g(x) gives a

tight bound (or exact bound) for the other. For example, ƒ(n) = n is both O(n) and also O(n3 ).
Similarly, g(n) = 8 • n log n is Ω(n) and O(n1.5), but not Ω(n2 ). Finally, the constant function h(n)

= 100100 is O(1), as is any constant, no matter how large. Note that care must be taken when

considering asymptotic notation; for example, h(x) = O(1) does not imply that h is a constant

function, since non-constant yet bounded functions such as h(x) = sin(x) are also O(1) by the
above definitions. Both O and Ω are reflexive and transitive relations, but are not commutative.

On the other hand, is Θ is commutative as well.

ƒ:S→T is said to be onto (or surjective) if for any element t in T, there exists an element s
in S such that ƒ(s)=t. More formally, ƒ is onto iff ∀ t ∈T ∃ s∈S ∋ ƒ(s)=t. Intuitively, a function
is onto if its entire range is "covered" by its domain. For example, ƒ:8→8 defined as f(x)=13-x

is onto (and coincidentally one-to-one as well), while g(x)=x2 is not, since some elements of g's

range do not have a corresponding element x in g's domain (i.e., there is no integer k such that

g(k)=3).

A function that is both injective and surjective is called bijective, and is said to be (or to
constitute) a one-to-one-correspondence between its domain and range. Intuitively, a
bijection (denoted ↔) is a perfect pairwise matching between two sets, with each element in each

set participating in exactly one match with an element of the other set. For example, the identity

6
function on an arbitrary domain D is always a bijection (i.e. ƒ:D ↔D ∋ f(x)=x). Another example
x-1 -x
of a bijection is h:ˆ→8 defined as h(x)= 2 if x is odd, 2 if x is even. The last example

illustrates the fact that an infinite set can be put into one-to-one correspondence with a proper

subset of itself! (which is of course never possible for a finite set).

The cardinality of a set S is said to be at least as large as the cardinality of a set T, if there
exists an onto function from S to T. More formally, |S|≥|T| ⇔ ∃ ƒ:S→T, f is onto. Note how this

definition generalizes the notion of cardinality comparisons to infinite sets. For example, the onto
function r:B→8 defined as "r(x)=integer closest to x" is witness to the fact that the reals have a

cardinality at least as large as the integers.

If |S|≥|T| and a bijection between S and T exists, the cardinality of S is said to be the same

as the cardinality of T. If |S|≥|T| but no bijection between S and T exists, the cardinality of S is

said to be strictly larger than the cardinality of T, denoted |S|>|T|. The bijection h defined earlier

proves that the natural numbers have the same cardinality as do the integers, even though the

former is a proper subset of the latter!

It turns out that the cardinality of the reals is strictly larger than the cardinality of the natural
numbers (formally |B|>|ˆ|). This can be proved using a diagonalization argument: we already

know that |B|≥|ˆ|, since y:B→ˆ defined as "y(x)=abs(truncate(x))" is onto. Now assume that

there exists an arbitrary bijection f:ˆ↔B. Now consider the real number Ω defined so that Ω's

k th digit (to the right of the decimal point) is equal to [f(k)'s k th digit] + 1 (modulo 10), for
k=1,2,3,... Clearly Ω is a well-defined real number, but is not in the range of f by construction.

It follows that f therefore cannot be a bijection as claimed, and since f was arbitrary, no bijection
between B and ˆ can possibly exist. Diagonalization is a powerful proof method which is often

employed to establish non-existence results.

7
 Bijections may be composed to form new bijections, so that if we have two bijections a:S→T

and b:T→V, then we can form a new bijection c:S→V, defined as c(x)=b(a(x)). As an example of
an application of this composition principle, we note that no bijection between B and 8 can

possibly exist: h (as defined earlier) is a bijection between ˆ and 8, and we already know that no

bijection between B and ˆ can possibly exist (by our earlier diagonalization proof). Therefore a

bijection between B and 8 would automatically yield (using our composition principle) a bijection

between B and ˆ, a contradiction.

"Oh!" said Alice. She was too much puzzled to make any other remark.

An infinite set is a set than can be put into one-to-one correspondence with a proper subset
of itself (or intuitively, a set with a cardinality greater than any integer k∈8). Any set that is finite,

or else that can be put into one-to-one correspondence with the integers is said to be countable (or

countably infinite). Any infinite set that can not be put into one-to-one correspondence with the
integers is said to be uncountable (or uncountably infinite). For example, ˆ, 8×8, 9, and {p |

p prime} are all countable sets, while B, {x | x∈B, 0≤x≤1}, and 2ˆ are all uncountable sets.

An alphabet is a finite set of symbols (e.g., Σ = {a,b,c}). A string is a finite sequence of

symbols chosen from a particular alphabet (e.g., w = abcaabbcc). The length of a string is the

numbers of symbols it is composed of (e.g., |bca| = 3). A language is a set of strings over some
alphabet. For example, for the alphabet Σ={a,b}, aaabbbabab is a string of length 10 over Σ , and

{anbn | n>0} is an infinite language over Σ . The unique string of length 0 is the empty string,
and is denoted by ε or ^. The concatenation of two strings x and y (denoted xy) is obtained by
following the symbols of x with the symbols of y, in order. More formally, if x=x1x2x3 ...xn and
y=y1y2y3 ...ym , where xi ∈Σ for 1≤i≤n and yj ∈Σ for 1≤j≤m, then xy=x1x2x3 ...xny1y2y3 ...ym . It

follows that for all strings w over some alphabet, wε=εw=w. For example, the string "hi"

concatenated to the string "there" yields the string "hithere".

8
 The concatenation operator may be extended to languages L1 and L2 as follows: L1L2 ={xy |
x∈L1 and y∈L2 } . LL may be denoted by L2; more generally, Lk =LLk-1, where L0 ={ε}. The

Kleene closure of a language L (denoted by L*) is defined as the infinite union

L0 ∪L1 ∪L2 ∪L3 ∪..., while L+ is defined as the infinite union L1 ∪L2 ∪L3 ∪... It follows that

L+ =LL* (note that this "superscript plus" notation is distinguished from the "superscript plus" used
earlier to denote the positive elements of a numerical set, e.g., 8+ = ˆ, and usually the context

may be consulted to avoid confusion).

For example, the language {a, b} concatenated to the language {1, 2, 3} yields the language

{a1, a2, a3, b1, b2, b3}, while {a,b}* denotes the set of all finite strings over the two symbols a
and b; more generally, Σ* denotes the set of all finite strings over the alphabet Σ . It turns out that
(L*)* =L* , and that unless L is the trivial language (i.e., {ε}) or the empty language (i.e., Ø)

then L* is countably infinite. Note that the trivial language {ε} is not the same as the empty

language Ø: the former contains one exactly string (i.e., the empty string) but the latter contains

none.

Any language L over a finite alphabet Σ is composed of some collection of finite strings.
More formally, L@Σ*. Clearly Σ * is countable (simply arrange the finite strings in Σ * by

increasing length, and within length by lexicographic dictionary order). Similarly, the set of all
finite descriptions is countable (simply arrange the description by increasing lengths and
*
lexicographically within the same length). On the other hand, the set of all languages 2Σ is

uncountable. This immediately implies that some languages are not finitely describable! Put

differently, the set of all possible finite algorithms (or descriptions) is countable (sort the finite

computer programs by size and lexicographically), while the set of problems (languages) is

uncountable; this means that any way of matching solutions to problems must leave out some

problems unmatched (i.e., unsolved), and therefore some problems have absolutely no solutions,

even in theory! Exhibiting an actual "finitely undescribable" set requires a little more work, but is

not altogether difficult; this is what Alan Turing did in his 1936 dissertation.

9
 "Curiouser and curiouser!" cried Alice.

The infinity corresponding to the cardinality of the integers is denoted by ℵ0

(pronounced "aleph null"). The infinity corresponding to the cardinality of the reals is
denoted as ℵ1. Our previous discussion established that ℵ0 < ℵ1, and formally we have ℵ1 =

2ℵ0 . For many years mathematicians have tried to find some infinity Ω such that ℵ0 < Ω < ℵ1 , or

prove that none exists. This question of whether there exists some infinity strictly larger than that

of the integers, yet strictly smaller than that of the reals, came to be known as the " continuum
hypothesis," and was finally settled by Cohen in 1966, who showed that to be independent o f

the axioms of set theory. That is, the consistency of set theory would not be changed if one

chooses to assume as an axiom either this hypothesis, or its negation! Several other well-known

mathematical statements enjoy this unique status of being independent of the axioms, and these

include the parallel postulate, as well as the axiom of choice (shown to be independent of

the other axioms by Godel in 1938).

More generally, we can obtain a whole hierarchy of infinities, each one strictly greater
than its predecessor; in particular, we have ℵi+1 = 2ℵi , where ℵi < ℵi+1. But when the indexes

of the alephs keep growing, nothing prevents them from soon becoming alephs themselves! In

other words, our "number-line" now looks like:

0, 1, 2,..., k, k+1,..., ℵ0 , ℵ1 , ℵ2 ,..., ℵk , ℵk+1 ,..., ℵℵ , ℵℵ ,..., ℵℵ , ℵℵ , . . .

0 1 k k+1

where the subscripts soon acquire subscripts which are themselves alephs, giving rise to an infinite

hierarchy of infinities! Does there exist any infinity "bigger" then any of these unimaginably large
cardinalities? It turns out that there is! The next "jump" in this sequence is denoted by ω

(pronounced "omega") and is bigger than any of the alephs "below" it. It is sometimes referred to

as the "first inaccessible infinity" because there is no way to "reach" it via any composition,

10
exponentiation, or subscript-nesting of alephs, etc., very much like there is no way to reach the

first aleph via any finite sequence of arithmetic operations on the ordinary integers.

The Red Queen shook her head. "you may call it 'nonsense' if you
like," she said, "but I've heard nonsense, compared with which that
would be as sensible as a dictionary!"

Interestingly, this fascinating progression of ever-increasing infinities does not stop; using
certain logical constructs is it possible to exhibit a vast hierarchy of inaccessible infinities past ω!

Logicians have even "found" infinities "larger" than any of the inaccessible ones, by stretching the

power of their axiomatic proof systems to the limit. Note that finding a new families of infinities

requires new and novel proof techniques, since the "jump" from one "level" of infinities to the next

"level" is as fundamental and conceptually difficult as the initial jump from the integers to the first
level at ℵ0 , or the jump from the alephs to ω! Currently only about six more fundamental "jumps"

in conceptualization are known to logicians, enjoying names such as the hyper-Mahlo cardinals,

the weakly compact cardinals, and the ineffable cardinals. It is not clear (even in theory) what

exotic mathematical constructs, if any, lay beyond that.

References

Behnke, H., Bachmann, Fladt, K., and Suss, W., (eds.), Fundamentals of Mathematics, Volume
I , MIT Press, Cambridge, Mass, 1973, pp. 50-61.

Davis, P., Hersh, R., The Mathematical Experience, Birkhauser, Boston, 1980, pp.152-157, 223-
236.

Harel, D., Algorithmics: the Spirit of Computing, 2nd Ed., Addison-Wesley, 1992.
Hilbert, D., On the Infinite, in Benacerrat, P., and Putnam, H., (eds.), Philosophy of
Mathematics: Selected Readings, Englewood Cliffs: Prentice-Hall, 1964, pp. 134-151.

Hopcroft, J., and Ullman, J., Introduction to Automata Theory, Languages, and Computation,
Addison-Wesley, Reading, Massachusetts, 1979.

Rucker, R., Infinity and the Mind: the Science and Philosophy of the Infinite, Harvester Press,
1982.

Zippin, L., Uses of Infinity, Mathematical Association of America, Washington, D.C., 1962.

11