The system of hyperreal numbers represents a rigorous method of treating the infinite and infinitesimal quantities. The hyperreals, or nonstandard reals, *R, are an extension of the real numbers R that contains numbers greater than anything of the form
 1 + 1 + \cdots + 1. \,
Such a number is infinite, and its inverse is infinitesimal. The term "hyperreal" was introduced by Edwin Hewitt in 1948.^{[1]}
The hyperreal numbers satisfy the transfer principle, a rigorous version of Leibniz's heuristic Law of Continuity. The transfer principle states that true first order statements about R are also valid in *R. For example, the commutative law of addition, x + y = y + x, holds for the hyperreals just as it does for the reals; since R is a real closed field, so is *R. Since \sin{\pi n}=0 for all integers n, one also has \sin{\pi H}=0 for all hyperintegers H. The transfer principle for ultrapowers is a consequence of o ' theorem of 1955.
Concerns about the soundness of arguments involving infinitesimals date back to ancient Greek mathematics, with Euclid replacing such proofs with ones using other techniques such as the method of exhaustion.^{[2]} In the 1960s, Abraham Robinson proved that the hyperreals were logically consistent if and only if the reals were. This put to rest the fear that any proof involving infinitesimals might be unsound, provided that they were manipulated according to the logical rules which Robinson delineated.
The application of hyperreal numbers and in particular the transfer principle to problems of analysis is called nonstandard analysis. One immediate application is the definition of the basic concepts of analysis such as derivative and integral in a direct fashion, without passing via logical complications of multiple quantifiers. Thus, the derivative of f(x) becomes f'(x) = {\rm st}\left( \frac{f(x+\Delta x)f(x)}{\Delta x} \right) for an infinitesimal \Delta x, where st(·) denotes the standard part function, which associates to every finite hyperreal the unique real infinitely close to it.
The transfer principle
The idea of the hyperreal system is to extend the real numbers R to form a system *R that includes infinitesimal and infinite numbers, but without changing any of the elementary axioms of algebra. Any statement of the form "for any number x..." that is true for the reals is also true for the hyperreals. For example, the axiom that states "for any number x, x + 0 = x" still applies. The same is true for quantification over several numbers, e.g., "for any numbers x and y, xy = yx." This ability to carry over statements from the reals to the hyperreals is called the transfer principle. However, statements of the form "for any set of numbers S ..." may not carry over. The only properties that differ between the reals and the hyperreals are those that rely on quantification over sets, or other higherlevel structures such as functions and relations, which are typically constructed out of sets. Each real set, function, and relation has its natural hyperreal extension, satisfying the same firstorder properties. The kinds of logical sentences that obey this restriction on quantification are referred to as statements in firstorder logic.
The transfer principle, however, doesn't mean that R and *R have identical behavior. For instance, in *R there exists an element ω such that
 1<\omega, \quad 1+1<\omega, \quad 1+1+1<\omega, \quad 1+1+1+1<\omega, \ldots.
but there is no such number in R. This is possible because the existence of such a number cannot be expressed as a first order statement.
Use in analysis
Calculus with algebraic functions
Informal notations for nonreal quantities have historically appeared in calculus in two contexts: as infinitesimals like dx and as the symbol , used, for example, in limits of integration of improper integrals.
As an example of the transfer principle, the statement that for any nonzero number x, 2x x, is true for the real numbers, and it is in the form required by the transfer principle, so it is also true for the hyperreal numbers. This shows that it is not possible to use a generic symbol such as for all the infinite quantities in the hyperreal system; infinite quantities differ in magnitude from other infinite quantities, and infinitesimals from other infinitesimals.
Similarly, the casual use of 1/0 = is invalid, since the transfer principle applies to the statement that division by zero is undefined. The rigorous counterpart of such a calculation would be that if is infinitesimal, then 1/ is infinite.
For any finite hyperreal number x, its standard part, st x, is defined as the unique real number that differs from it only infinitesimally. The derivative of a function y(x) is defined not as dy/dx but as the standard part of dy/dx.
For example, to find the derivative f′(x) of the function f(x) = x^{2}, let dx be an infinitesimal. Then,
 f'(x)\, =\operatorname{st}\left(\frac{f(x + dx)  f(x)}{dx}\right)   =\operatorname{st}\left(\frac{x^2 + 2x \cdot dx + dx^2 x^2}{dx}\right)   =\operatorname{st}\left(\frac{2x \cdot dx + dx^2}{dx}\right)   =\operatorname{st}\left(2x + dx\right)   =2x\, }
The use of the standard part in the definition of the derivative is a rigorous alternative to the traditional practice of neglecting the square of an infinitesimal quantity. After the third line of the differentiation above, the typical method from Newton through the 19th century would have been simply to discard the dx^{2} term. In the hyperreal system, dx^{2} 0, since dx is nonzero, and the transfer principle can be applied to the statement that the square of any nonzero number is nonzero. However, the quantity dx^{2} is infinitesimally small compared to dx; that is, the hyperreal system contains a hierarchy of infinitesimal quantities.
Integration
One way of defining a definite integral in the hyperreal system is as the standard part of an infinite sum on a hyperfinite lattice defined as a, a + dx, a + 2dx, ... a + ndx, where dx is infinitesimal, n is an infinite hypernatural, and the lower and upper bounds of integration are a and b = a + n dx.^{[3]}
Properties
The hyperreals *R form an ordered field containing the reals R as a subfield. Unlike the reals, the hyperreals do not form a standard metric space, but by virtue of their order they carry an order topology.
The use of the definite article the in the phrase the hyperreal numbers is somewhat misleading in that there is not a unique ordered field that is referred to in most treatments. However, a 2003 paper by Kanovei and Shelah^{[4]} shows that there is a definable, countably saturated (meaning saturated, but not of course countable) elementary extension of the reals, which therefore has a good claim to the title of the hyperreal numbers. Furthermore, the field obtained by the ultrapower construction from the space of all real sequences, is unique up to isomorphism if one assumes the continuum hypothesis.
The condition of being a hyperreal field is a stronger one than that of being a real closed field strictly containing R. It is also stronger than that of being a superreal field in the sense of Dales and Woodin.^{[5]}
Development
The hyperreals can be developed either axiomatically or by more constructively oriented methods. The essence of the axiomatic approach is to assert (1) the existence of at least one infinitesimal number, and (2) the validity of the transfer principle. In the following subsection we give a detailed outline of a more constructive approach. This method allows one to construct the hyperreals if given a settheoretic object called an ultrafilter, but the ultrafilter itself cannot be explicitly constructed. (Kanovei and Shelah^{[4]} have found a method that gives an explicit construction, at the cost of a significantly more complicated treatment.)
From Newton to Robinson
When Newton and (more explicitly) Leibniz introduced differentials, they used infinitesimals and these were still regarded as useful by later mathematicians such as Euler and Cauchy. Nonetheless these concepts were from the beginning seen as suspect, notably by George Berkeley. Berkeley's criticism centered on a perceived shift in hypothesis in the definition of the derivative in terms of infinitesimals (or fluxions), where dx is assumed to be nonzero at the beginning of the calculation, and to vanish at its conclusion (see Ghosts of departed quantities for details). When in the 1800s calculus was put on a firm footing through the development of the ( , )definition of limit by Bolzano, Cauchy, Weierstrass, and others, infinitesimals were largely abandoned, though research in nonArchimedean fields continued (Ehrlich 2006).
However, in the 1960s Abraham Robinson showed how infinitely large and infinitesimal numbers can be rigorously defined and used to develop the field of nonstandard analysis.^{[6]} Robinson developed his theory nonconstructively, using model theory; however it is possible to proceed using only algebra and topology, and proving the transfer principle as a consequence of the definitions. In other words hyperreal numbers per se, aside from their use in nonstandard analysis, have no necessary relationship to model theory or first order logic, although they were discovered by the application of model theoretic techniques from logic. Hyperreal fields were in fact originally introduced by Hewitt (1948) by purely algebraic techniques, using an ultrapower construction.
The ultrapower construction
We are going to construct a hyperreal field via sequences of reals.^{[7]} In fact we can add and multiply sequences componentwise; for example:
 (a_0, a_1, a_2, \ldots) + (b_0, b_1, b_2, \ldots) = (a_0 +b_0, a_1+b_1, a_2+b_2, \ldots)
and analogously for multiplication. This turns the set of such sequences into a commutative ring, which is in fact a real algebra A. We have a natural embedding of R in A by identifying the real number r with the sequence (r, r, r, ...) and this identification preserves the corresponding algebraic operations of the reals. The intuitive motivation is, for example, to represent an infinitesimal number using a sequence that approaches zero. The inverse of such a sequence would represent an infinite number. As we will see below, the difficulties arise because of the need to define rules for comparing such sequences in a manner that, although inevitably somewhat arbitrary, must be selfconsistent and well defined. For example, we may have two sequences that differ in their first n members, but are equal after that; such sequences should clearly be considered as representing the same hyperreal number. Similarly, most sequences oscillate randomly forever, and we must find some way of taking such a sequence and interpreting it as, say, 7+\epsilon, where \epsilon is a certain infinitesimal number.
Comparing sequences is thus a delicate matter. We could, for example, try to define a relation between sequences in a componentwise fashion:
 (a_0, a_1, a_2, \ldots) \leq (b_0, b_1, b_2, \ldots) \iff a_0 \leq b_0 \wedge a_1 \leq b_1 \wedge a_2 \leq b_2 \ldots
but here we run into trouble, since some entries of the first sequence may be bigger than the corresponding entries of the second sequence, and some others may be smaller. It follows that the relation defined in this way is only a partial order. To get around this, we have to specify which positions matter. Since there are infinitely many indices, we don't want finite sets of indices to matter. A consistent choice of index sets that matter is given by any free ultrafilter U on the natural numbers; these can be characterized as ultrafilters which do not contain any finite sets. (The good news is that Zorn's lemma guarantees the existence of many such U; the bad news is that they cannot be explicitly constructed.) We think of U as singling out those sets of indices that "matter": We write (a_{0}, a_{1}, a_{2}, ...) (b_{0}, b_{1}, b_{2}, ...) if and only if the set of natural numbers { n : a_{n} b_{n} } is in U.
This is a total preorder and it turns into a total order if we agree not to distinguish between two sequences a and b if a b and b a. With this identification, the ordered field *R of hyperreals is constructed. From an algebraic point of view, U allows us to define a corresponding maximal ideal I in the commutative ring A (namely, the set of the sequences that vanish in some element of U), and then to define *R as A/I; as the quotient of a commutative ring by a maximal ideal, *R is a field. This is also notated A/U, directly in terms of the free ultrafilter U; the two are equivalent. The maximality of I follows from the possibility of, given a sequence a, construct a sequence b inverting its nonnull elements and not altering its null entries. The product ab is in this case identified with the number 1, and any ideal containing 1 must be A. In the resulting field, these a and b are inverses.
The field A/U is an ultrapower of R. Since this field contains R it has cardinality at least that of the continuum. Since A has cardinality
 (2^{\aleph_0})^{\aleph_0} = 2^{\aleph_0^2} =2^{\aleph_0},\,
it is also no larger than 2^{\aleph_0}, and hence has the same cardinality as R.
One question we might ask is whether, if we had chosen a different free ultrafilter V, the quotient field A/U would be isomorphic as an ordered field to A/V. This question turns out to be equivalent to the continuum hypothesis; in ZFC with the continuum hypothesis we can prove this field is unique up to order isomorphism, and in ZFC with the negation of continuum hypothesis we can prove that there are nonorderisomorphic pairs of fields which are both countably indexed ultrapowers of the reals.
For more information about this method of construction, see ultraproduct.
An intuitive approach to the ultrapower construction
The following is an intuitive way of understanding the hyperreal numbers. The approach taken here is very close to the one in the book by Goldblatt.^{[8]} Recall that the sequences converging to zero are sometimes called infinitely small. These are almost the infinitesimals in a sense; the true infinitesimals include the classes of sequences that contain a sequence converging to zero. However, there may be infinitesimals not represented by null sequences; see Ppoint.
Let us see where these classes come from. Consider first the sequences of real numbers. They form a ring, that is, one can multiply add and subtract them, but not always divide by nonzero. The real numbers are considered as the constant sequences, the sequence is zero if it is identically zero, that is, a_{n} = 0 for all n.
In our ring of sequences one can get ab = 0 with neither a = 0 nor b = 0. Thus, if for two sequences a, b\quad one has ab = 0, at least one of them should be declared zero. Surprisingly enough, there is a consistent way to do it. As a result, the equivalence classes of sequences that differ by some sequence declared zero will form a field which is called a hyperreal field. It will contain the infinitesimals in addition to the ordinary real numbers, as well as infinitely large numbers (the reciprocals of infinitesimals, including those represented by sequences diverging to infinity). Also every hyperreal which is not infinitely large will be infinitely close to an ordinary real, in other words, it will be the sum of an ordinary real and an infinitesimal.
This construction is parallel to the construction of the reals from the rationals given by Cantor. He started with the ring of the Cauchy sequences of rationals and declared all the sequences that converge to zero to be zero. The result is the reals. To continue the construction of hyperreals, let us consider the zero sets of our sequences, that is, the z(a)=\{i: a_i=0\}\quad, that is, z(a)\quad is the set of indexes i\quad for which a_i=0\quad. It is clear that if ab=0\quad, then the union of z(a)\quad and z(b)\quad is N (the set of all natural numbers), so:
 One of the sequences that vanish on 2 complementary sets should be declared zero
 If a\quad is declared zero, ab\quad should be declared zero too, no matter what b\quad is.
 If both a\quad and b\quad are declared zero, then a^2+b^2\quad should also be declared zero.
Now the idea is to single out a bunch U of subsets X of N and to declare that a=0\quad if and only if z(a)\quad belongs to U. From the above conditions one can see that:
 From 2 complementary sets one belongs to U
 Any set containing a set that belongs to U, also belongs to U.
 An intersection of any 2 sets belonging to U belongs to U.
 Finally, we do not want an empty set to belong to U because then everything becomes zero, as every set contains an empty set.
Any family of sets that satisfies (2)(4) is called a filter (an example: the complements to the finite sets, it is called the Fr chet filter and it is used in the usual limit theory). If (1) also holds, U is called an ultrafilter (because you can add no more sets to it without breaking it). The only explicitly known example of an ultrafilter is the family of sets containing a given element (in our case, say, the number 10). Such ultrafilters are called trivial, and if we use it in our construction, we come back to the ordinary real numbers. Any ultrafilter containing a finite set is trivial. It is known that any filter can be extended to an ultrafilter, but the proof uses the axiom of choice. The existence of a nontrivial ultrafilter (the ultrafilter lemma) can be added as an extra axiom, as it is weaker than the axiom of choice.
Now if we take a nontrivial ultrafilter (which is an extension of the Fr chet filter) and do our construction, we get the hyperreal numbers as a result.
If f\quad is a real function of a real variable x\quad then f\quad naturally extends to a hyperreal function of a hyperreal variable by composition:
 f(\{x_n\})=\{f(x_n)\}\,
where \{ \dots\} means "the equivalence class of the sequence \dots relative to our ultrafilter", two sequences being in the same class if and only if the zero set of their difference belongs to our ultrafilter.
All the arithmetical expressions and formulas make sense for hyperreals and hold true if they are true for the ordinary reals. One can prove that any finite (that is, such that x < a\quad for some ordinary real a\quad) hyperreal x\quad will be of the form y+d\quad where y\quad is an ordinary (called standard) real and d\quad is an infinitesimal.
Now one can see that f\quad is continuous means that f(a)f(x)\quad is infinitely small whenever xa\quad is, and f\quad is differentiable means that
 (f(x)f(a))/(xa)f'(a)\quad
is infinitely small whenever xa\quad is. Remarkably, if one allows a\quad to be hyperreal, the derivative will be automatically continuous (because, f\quad being differentiable at x\quad,
 f'(x)(f(x)f(a))/(xa)=f'(x)(f(a)f(x))/(ax)\quad
is infinitely small when xa\quad is, therefore f'(x)f'(a)\quad is also infinitely small when xa\quad is).
Properties of infinitesimal and infinite numbers
The finite elements F of *R form a local ring, and in fact a valuation ring, with the unique maximal ideal S being the infinitesimals; the quotient F/S is isomorphic to the reals. Hence we have a homomorphic mapping, st(x), from F to R whose kernel consists of the infinitesimals and which sends every element x of F to a unique real number whose difference from x is in S; which is to say, is infinitesimal. Put another way, every finite nonstandard real number is "very close" to a unique real number, in the sense that if x is a finite nonstandard real, then there exists one and only one real number st(x) such that x – st(x) is infinitesimal. This number st(x) is called the standard part of x, conceptually the same as x to the nearest real number. This operation is an orderpreserving homomorphism and hence is wellbehaved both algebraically and order theoretically. It is orderpreserving though not isotonic, i.e. x \le y implies \operatorname{st}(x) \le \operatorname{st}(y), but x < y does not imply \operatorname{st}(x) < \operatorname{st}(y).
 We have, if both x and y are finite,

 \operatorname{st}(x + y) = \operatorname{st}(x) + \operatorname{st}(y)

 \operatorname{st}(x y) = \operatorname{st}(x) \operatorname{st}(y)
 If x is finite and not infinitesimal.

 \operatorname{st}(1/x) = 1 / \operatorname{st}(x)

 \operatorname{st}(x) = x
The map st is continuous with respect to the order topology on the finite hyperreals; in fact it is locally constant.
Hyperreal fields
Suppose X is a Tychonoff space, also called a T_{3.5} space, and C(X) is the algebra of continuous realvalued functions on X. Suppose M is a maximal ideal in C(X). Then the factor algebra A = C(X)/M is a totally ordered field F containing the reals. If F strictly contains R then M is called a hyperreal ideal (terminology due to Hewitt (1948)) and F a hyperreal field. Note that no assumption is being made that the cardinality of F is greater than R; it can in fact have the same cardinality.
An important special case is where the topology on X is the discrete topology; in this case X can be identified with a cardinal number and C(X) with the real algebra \Bbb{R}^\kappa of functions from to R. The hyperreal fields we obtain in this case are called ultrapowers of R and are identical to the ultrapowers constructed via free ultrafilters in model theory.
See also
References
Further reading

 Hatcher, William S. (1982) "Calculus is Algebra", American Mathematical Monthly 89: 362–370.
 Hewitt, Edwin (1948) Rings of realvalued continuous functions. I. Trans. Amer. Math. Soc. 64, 45 99.

 Keisler, H. Jerome (1994) The hyperreal line. Real numbers, generalizations of the reals, and theories of continua, 207 237, Synthese Lib., 242, Kluwer Acad. Publ., Dordrecht.

External links
ar: ca:Nombre hiperreal cs:Hyperre ln slo de:Hyperreelle Zahl es:N mero hiperreal eo:Hiperreela nombro fr:Nombre hyperr el ko: it:Numero iperreale nn:Hyperreelt tal pt:N mero hiperreal ro:Num r hiperreal ru: sl:Hiperrealno tevilo fi:Hyperreaaliluku sv:Hyperreella tal zhclassical: zh: ( )
