Measurable Numbers in Bolzano's Theory

• Measurable numbers are defined via rational approximations, where each number lies between adjacent rational intervals p/q and (p+1)/q for every positive integer q.
• Bolzano introduced measurable numbers to extend arithmetic and provide a complete ordered field, predating modern constructions of the reals.
• The framework addresses equality issues through infinitesimals and underpins convergence proofs, linking historical analysis with modern mathematical ideas.

Searching arXiv for the cited papers to ground the article. Measurable numbers are a number concept introduced by Bernard Bolzano in the final section of his Reine Zahlenlehre during the early 1830s in order to provide the real-number domain needed for convergence arguments, especially the criterion later known as the Cauchy criterion. In Bolzano’s framework, a number is measurable when it can be trapped, for every positive integer $q$, between two adjacent rational numbers of the form $p/q$ and $(p+1)/q$. The surviving manuscript, partially published in 1962 and more fully published in 1976, has been interpreted as an early construction of the real numbers, while also generating a substantial historiographical controversy about equality, infinitesimals, and textual incompleteness (Russ et al., 2018).

1. Historical origin and analytical purpose

Bolzano introduced measurable numbers in the final section of his Reine Zahlenlehre as part of a broader attempt to extend arithmetic beyond ordinary finite rational expressions. The motivating problem was foundational rather than merely computational: Bolzano had already stated in 1817 a version of the convergence criterion later associated with Cauchy, but a proof of the existence of limits required a sufficiently rich number domain. Measurable numbers were intended to supply that missing domain (Russ et al., 2018).

The manuscript’s point of departure is the insufficiency of rational arithmetic for irrational limits. Up to finite combinations of addition, subtraction, multiplication, and division, one remains within the rational numbers. To accommodate limits such as $\sqrt 2$, Bolzano introduced what he called “infinite quantity concepts” or “infinite number expressions.” The terminology is unstable in the manuscript, but the functional role is consistent: these expressions are meant to designate numbers arising from endless processes of approximation rather than from finite symbolic composition alone (Russ et al., 2018).

The paper on Bolzano emphasizes that this project pre-dates the better-known constructions of Dedekind, Cantor, Weierstrass, and Méray by several decades. In modern terms, the intended program was to start with rational arithmetic, define a richer domain of measurable numbers, and show that every sequence satisfying the Bolzano–Cauchy criterion converges to a measurable number. A plausible implication is that Bolzano was aiming at an ordered field with a completeness property adequate for analysis (Russ et al., 2018).

2. Definition by rational approximation

Bolzano’s definition is stated in terms of measurement by rational approximations. For a number expression $S$, one can “determine it by approximation, or measure it,” if for every positive integer $q$ there exists an integer $p$ such that

$$S = \frac{p}{q} + P_1 \qquad\text{and}\qquad S = \frac{p+1}{q} - P_2,$$

where $P_1$ and $P_2$ are strictly positive number expressions. Equivalently,

$p/q$0

The fraction $p/q$1 is the “measuring fraction” (Russ et al., 2018).

This definition associates a measurable number with an infinite family of rational approximation intervals

$p/q$2

one for each denominator $p/q$3. The paper describes these intervals as not necessarily nested in the strict modern sense, but “dually directed”: as $p/q$4 increases, they become finer and their common intersection identifies the number. This interval-based reading places Bolzano’s construction close in spirit to later approaches to the real numbers based on approximation or cuts (Russ et al., 2018).

A geometric interpretation is also given. If one imagines a line segment of length $p/q$5 measured with a ruler divided into $p/q$6 equal parts, then $p/q$7 either lands exactly on a mark or lies between two adjacent marks. In either case it is measurable. This interpretation makes explicit that measurability concerns arbitrarily fine rational localization rather than symbolic expansion alone (Russ et al., 2018).

3. Infinite number expressions, equality, and infinitesimals

Bolzano’s measurable numbers are built from “infinite number expressions,” expressions involving infinitely many arithmetic operations. The examples cited in the manuscript include $p/q$8 in inf. and $p/q$9 in inf. The manuscript makes clear that “infinite” is not meant literally as a requirement of surveying infinitely many constituents; rather, the expression designates a single number concept generated by an endless process (Russ et al., 2018).

One of the most delicate technical issues is equality. Bolzano’s original definition treated two measurable expressions as equal if they always had the same measuring fraction $(p+1)/q$0 for each $(p+1)/q$1. He then recognized that this criterion was too rigid when quantities differ only by something infinitesimal. The manuscript gives examples such as $(p+1)/q$2 and $(p+1)/q$3, which can have different measuring fractions despite differing only infinitesimally (Russ et al., 2018).

Bolzano therefore revised the definition. Under the revised version, if the difference $(p+1)/q$4 always has measuring fraction $(p+1)/q$5, then $(p+1)/q$6; if the difference is positive, then $(p+1)/q$7; if negative, then $(p+1)/q$8. The paper judges this revision to be much closer to a genuine equivalence relation and notes that Bolzano seems to have been moving toward the modern idea that numbers are equivalence classes under a suitable relation. At the same time, the manuscript was not fully rewritten after the revision, and this textual fact is one source of later interpretive difficulty (Russ et al., 2018).

The manuscript also contains a notable treatment of infinitesimals. Bolzano considers

$(p+1)/q$9

and shows that it is measurable with measuring fraction $\sqrt 2$0, however large $\sqrt 2$1 is, yet should not be identified with $\sqrt 2$2. He calls it an “infinitely small positive number.” The paper emphasizes that this represents a major shift from Bolzano’s earlier 1816 criticism of infinitesimals as “self-contradictory,” and that the manuscript contains evidence of a developing infinitesimal theory within the measurable-number framework (Russ et al., 2018).

4. Ordered-field structure and completeness

The paper summarizes Bolzano’s results as showing that measurable numbers behave like a linearly ordered field. Among the order properties Bolzano proves are transitivity, trichotomy or linearity, density, the Archimedean property, and compatibility of order with addition. On the algebraic side he proves closure under addition and multiplication, together with identities and laws such as $\sqrt 2$3, associativity and commutativity of multiplication, distributivity, and rules for fractions (Russ et al., 2018).

These results are central to the modern assessment of the theory. Taken together, they are exactly the structural properties expected of the real numbers. The paper’s authors therefore treat Bolzano’s theory not as an isolated doctrine of approximation, but as a candidate construction of a complete ordered field. This suggests that the measurable-number framework was intended to supply the continuum required by analysis rather than merely a collection of admissible limits (Russ et al., 2018).

Bolzano’s key theorem concerns convergence. In modern restatement, if a sequence $\sqrt 2$4 is such that for every positive fraction $\sqrt 2$5, the differences $\sqrt 2$6 eventually become smaller than $\sqrt 2$7, then there exists one and only one measurable number $\sqrt 2$8 to which the sequence converges. His proof is organized by cases: non-decreasing sequences, non-increasing sequences, and alternating sequences. For monotone sequences, the measuring fractions determine a unique limit; for alternating sequences, the proof passes to monotone subsequences and uses their limits (Russ et al., 2018).

The paper judges that, despite some roughness, the logical structure of this completeness argument is basically sound. In modern terminology, it reads as a proof that the measurable numbers form a complete ordered field adequate for the Bolzano–Cauchy convergence criterion (Russ et al., 2018).

5. Interpretation and historiographical controversy

The central interpretive question is whether Bolzano’s measurable numbers are genuinely real numbers. The paper’s affirmative view is that they can indeed be regarded as real numbers in a modern sense because they form a complete linearly ordered field, are defined by rational approximation intervals, support a proof of the Bolzano–Cauchy convergence criterion, and provide a foundation for irrational numbers without presupposing them (Russ et al., 2018).

The cautionary view arises from textual and conceptual difficulties. The manuscript is incomplete; the exact meaning of infinite number expressions is not fully settled; the equality relation was revised but not systematically integrated into the whole text; and some alternating-series examples are technically delicate. Later historians therefore disagreed strongly over whether the theory is consistent and whether it presupposes what it seeks to construct (Russ et al., 2018).

The paper identifies several major positions in this debate. van Rootselaar criticized the theory as incorrect and inconsistent and argued that Bolzano still presupposed the real numbers. Laugwitz replied that small repairs to the definitions would render the theory consistent and real-number-like, and after the fuller 1976 publication he took a more strongly positive position because Bolzano had already moved toward the needed repair. Rusnock, Sebestik, and others offered more balanced analyses, with Rusnock concluding that Bolzano was “almost entirely successful” in characterizing the reals (Russ et al., 2018).

The resulting consensus in the paper is deliberately nuanced: historically, Bolzano’s measurable numbers are an early pre-Dedekind and pre-Cantor construction of the reals; conceptually, they are intended to be real numbers or something closely equivalent; textually, the identification is reconstructive rather than explicit; mathematically, on the authors’ reading, the system is isomorphic to the real numbers as now understood (Russ et al., 2018).

6. Later and distinct uses of “measurable numbers”

In later literature, the phrase “measurable numbers” also appears in contexts quite different from Bolzano’s theory. One such context is numerosity theory, which develops elementary numerosity as a generalized counting function

$\sqrt 2$9

into a non-Archimedean ordered field $S$0. Given a positive unit $S$1, the associated real-valued finitely additive measure is defined by

$S$2

The paper explicitly states that its “measurable numbers” are the non-Archimedean values obtained from numerosity after normalization and taking standard part,

$S$3

and presents these as a bridge between classical measure theory, numerosity theory, and standard-part projection (Benci et al., 2014).

A different modern usage appears in work based on Sergeyev’s grossone methodology. There the objective is to “measure” infinite sets more finely than Cantorian cardinality permits. For the set $S$4 of algebraic numbers, the paper gives lower and upper estimates for the grossone-based “number of elements”: $S$5 The methodological claim is that grossone-based numerals provide a numerical size for certain infinite sets rather than merely a cardinal type (Sergeyev, 2022).

These later uses are conceptually distinct from Bolzano’s measurable numbers. In Bolzano, measurability is a criterion of rational approximation for individual numbers and a basis for completeness. In numerosity and grossone settings, measurability concerns the assignment of numerical sizes to sets, often within non-Archimedean frameworks. A plausible implication is that the phrase has acquired a broader semantic range in modern foundational work, even though its classical historical referent remains Bolzano’s theory of rationally approximable numbers (Benci et al., 2014).