REVIEW 4 major objections 5 minor 49 references
On Conway's Numbers and Games, the Von Neumann Universe, and Pure Set Theory
T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash
Pith's one-line read This paper claims that surreal numbers are exactly the sets of ordinals with a maximal element, and that this definition yields the full Conway field inside the von Neumann universe.
desk verdict A clean, honest re-presentation of the surreal numbers as sets of ordinals with a maximum, whose order and tree structure are genuinely proved but whose field arithmetic and game equivalence are imported rather than proven. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the class $\mathrm{NO}$ of sets of ordinals with a maximal element, equipped with birthday $b(x)=\max x$. The sign-expansion $s_x$ turns each number into a sign string; the discriminant $\delta(x,y)=\min(x\triangle y)$ gives the total order by asking which side contains the discriminant; children $x^+=x\cup\{b(x)+1\}$ and $x^-=x\cup\{b(x)+1\}\setminus\{b(x)\}$ generate the binary tree. The Fundamental Existence Theorem, which gives every Conway cut $\langle L,R\rangle$ a unique number $c$ with minimal birthday and $c\preceq y$ for any $y$ between $L$ and $R$, is what lets Conway's cut formulas for addition, multiplication, and multiplicative inverses be imported, and it also identifies each stage $\mathrm{NO}_{\alpha+1}$ as the Cuesta-Dutari completion of $\mathrm{NO}_\alpha$.
What would settle it
Compute ε · ω_Co where ε = {0, ω} and ω_Co = ω + 1 is the Conway ordinal ω: if the transfer is sound, the product must be 1_Co = {0, 1}. A direct calculation from the cut formulas, or a check that the result depends on the choice of timely cut, that yields any other value would refute the claim that the new NO carries Conway arithmetic.
Extended reading notes
Core claim
The central claim is that the von Neumann universe already contains the surreal numbers as a definable subclass: NO consists of the sets of ordinals that have a maximum b(x), called the birthday. Every such set encodes a sign expansion (plus for element, minus for hole, zero after the birthday), the total order is read off the least ordinal distinguishing two numbers (the discriminant), and the tree order x ≼ y says x is an initial segment of y below b(x). Starting from nothing, the stages NO_α are the Cuesta-Dutari completions of the previous stage, and the whole tree is connected and complete. The paper argues that Conway's cut-based addition, multiplication, and inverses can be transplanted to this setting, with proofs sketched through the Fundamental Existence Theorem and deferred to detailed presentations, and that the result is the ordered Field NO containing all ordinals and a canonical copy R_Co of the reals.
Load-bearing premise
The proof depends on showing that the usual rules for adding, multiplying, and taking inverses of surreal numbers work when numbers are represented as sets of ordinals with a largest element, and match Conway's original game operations; the paper sketches this and points to other books for details.
Editorial extensions
If this is right
- If NO as defined here is a field, then surreal numbers need not be introduced through games: the whole ordered Field, with all ordinals and infinitesimals, exists inside the von Neumann hierarchy and can be taught immediately after ordinals.
- The Conway reals R_Co, short numbers plus long reals of the form X ∪ {ω}, form a field isomorphic to R, with dyadic rationals as the short numbers and rationals exactly the eventually periodic sign expansions.
- Each stage NO_α+1 is the Cuesta-Dutari completion of NO_α, so the hierarchy of numbers is generated from 0_Co = {0} by completions and limits, giving a purely set-theoretic construction of the full binary number tree.
- The equivalence with Alling's axioms, Gonshor's sign expansions, and Conway's games means that the same absolute arithmetic continuum is reached from pure sets, from sign strings, and from partizan games.
- The ordinal operations of Cantor and the Hessenberg operations both reappear inside NO, with the Conway ordinals being exactly the successor von Neumann ordinals and the field operations extending the commutative natural operations.
Reading between the lines
- A natural next test is to make the arithmetic fully combinatorial: the paper leaves open a direct sign-sequence formula for x + y and xy, and its Grothendieck-group sketch suggests that such formulas would connect surreal arithmetic to transseries and generalized power series.
- If the von Neumann-universe realization is accepted, the author's pure set theory program extends beyond numbers: the same cut and tree language could be used to build canonical copies of the surcomplex numbers and possibly p-adic-like completions, though the paper only speculates about these.
- The philosophical claim that the Conway reals are the only natural construction of R avoiding Kuratowski pairs is stronger than the mathematical equivalence claims; it would require a precise definition of naturality to become testable.
- The quantum-versus-classical framing of rank and birthday is an interpretive layer rather than a proven result, and its value would have to be judged by whether it produces new structural theorems about the von Neumann universe.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes a presentation of surreal numbers as sets of von Neumann ordinals having a maximal element, called the birthday (Definition 2.1). It develops the total order, the descendance tree, limits, and cut-theoretic completeness in Sections 2.4–2.6, and gives an algorithmic bijection between the 'Conway reals' and the usual reals in Section 2.3. The paper then states, but does not fully prove, the transfer of Conway arithmetic to this setting (Theorems 2.67–2.70) and the equivalence with Conway's original games construction (Theorem 3.30). The remaining chapters discuss nimbers, a graded von Neumann universe for partizan games, philosophical interpretations, surcomplex and 'cocomplex' numbers, and open problems.
Significance. If the arithmetic transfer and game-equivalence were fully proved, the paper would give a strikingly simple set-theoretic entry point to surreal numbers and a convincing argument that the ordered surreal tree is a natural structure of the von Neumann universe. The order-theoretic core—total order, tree completeness, and the Fundamental Existence Theorem (Theorems 2.24, 2.41, 2.45, 2.72)—is proved carefully and appears sound. The paper also deserves credit for making the Conway reals explicit through the Berlekamp-style algorithm in Theorem 2.16 and for framing impartial and partizan games in terms of 'pure set theory' and a graded universe in a way that may be pedagogically and conceptually useful. However, the central field-theoretic claims are currently delegated to [Sim], [ONAG], and [S], so the manuscript as written establishes the ordered tree and its completeness, but not the full claimed foundation of the surreal numbers as a field.
major comments (4)
- [§2.8, Theorems 2.67–2.69] Theorems 2.67 and 2.68 assert the existence, uniqueness, associativity, commutativity, distributivity, and ordered-field structure of addition and multiplication on NO, but the text gives only an 'Idea of proof' that refers to [Sim] and [ONAG]. The preceding results do not immediately supply the missing content: Theorem 2.72 guarantees a cut number for any Conway cut, yet one still has to prove, by transfinite induction, that the cut defining x+y (and xy) satisfies the required inequalities, that the result is independent of the chosen cut representation, and that the resulting operations satisfy the ring axioms and admit multiplicative inverses. Because the claimed 'firm ground of pure set theory' includes the field structure, this is a load-bearing gap rather than a cosmetic omission.
- [§3.3, Theorem 3.30] The equivalence theorem with Conway's game construction is stated without a proof. The text says 'In order to prove the theorem, there are a lot of things to check' and 'In principle, everything is contained in [S]', but no theorem-by-theorem derivation is given. In particular, the injectivity of [G], the characterization of the image by condition (3.7), and the transfer of field operations to the quotient are exactly the assertions needed to substantiate the paper's claim that NO is equivalent to Conway's original numbers-as-games. This external delegation should be stated explicitly in the theorem, or the proof should be included.
- [§2.3 and §2.8, Theorem 2.70] Theorem 2.70 claims that RCo is a subfield of NO and that the bijection of Theorem 2.16 is a field isomorphism. The proof of Theorem 2.16 only establishes an order-preserving bijection between R and RCo; compatibility of this bijection with addition and multiplication is not shown in Section 2.3 and is not covered by the 'Idea of proof' in Section 2.8. Since the paper advertises the Conway reals as a canonical copy of R inside the von Neumann universe, the field-isomorphism statement needs a proof or an explicit reference to a proved theorem.
- [§0.7 and §4.2.2] The Introduction states that the Fundamental Existence Theorem 'entails' Alling's axioms and 'establishes equivalence' with other approaches, and Section 2.9 repeats this claim. Yet Section 4.2.2 explicitly says that a purely combinatorial definition of Conway arithmetic is still a programme whose missing details are 'remote'. This discrepancy should be reconciled: either the theorems in Section 2.8 are meant as imported known results, or the introduction should describe the contribution as the order-theoretic tree plus a formal translation of previously known arithmetic rather than as a fully self-contained foundation.
minor comments (5)
- [§0.1 and §0.3] The manuscript contains several typos, including 'would should' in the abstract and 'exploses' and 'take akes' in the introductory sections; these should be corrected during revision.
- [§1.3.3 and §2.8] There are unresolved cross-reference markers such as 'Equation (??)' in the proof sketch of Theorem 2.67; these should be replaced by the intended equation numbers.
- [§2.3, Theorem 2.16] The proof of bijectivity in Theorem 2.16 is very concise: it should spell out how the 'long ends' convention removes the binary-expansion ambiguity on the negative side and how the finite/infinite distinction is preserved by the inverse map.
- [§2.6, Definition 2.50] The 'topology' of closed sets on the proper class NO is informal; since it is used mainly for motivation, the text should explicitly state that it is not a topological space in the usual sense and that no separation axioms are being claimed.
- [§4.3, Definition 4.6] Definition 4.6 and Table 4.1 are labelled 'tentative' and 'speculative', which is honest, but the surrounding text should make even clearer that the cocomplex-number construction is an outlook and not part of the paper's main theorem set.
Circularity Check
No significant circularity: the core order/tree construction is derived from Definition 2.1, and the arithmetic and equivalence results are imported from external sources rather than presupposed.
full rationale
The paper's central construction (Definition 2.1, numbers as sets of ordinals with maximum) is independent of the results it later compares with Conway's games, Gonshor's sign expansions, and Alling's axiomatization. The order, tree, truncation, limit, canonical-cut, and CD-completion theorems (Theorems 2.24, 2.26, 2.41, 2.45, 2.72, 2.77) are proved in the text from this definition and from standard ordinal facts. The load-bearing field theorems (2.67-2.69) and the game-theoretic equivalence (Theorem 3.30) are not proved in full; the paper explicitly delegates them to [Sim], [ONAG], [S], and [Go]. That is a proof-obligation gap and a conditional claim, not circularity: the cited works are external, do not use the paper's conclusions as premises, and are not self-citations. The only self-citation is the aside in Section 0.10 crediting [Be08] in Ehrlich's overview; it plays no role in the derivation. No fitted parameter is renamed as a prediction, and no definition is circularly expressed in terms of the target theorem. The paper is honest that the arithmetic transfer is imported: 'In principle, everything is contained in [S]' (Section 3.3). Thus the appropriate finding is no significant circularity; the correctness risk about unproved transfer belongs to completeness and rigor, not circularity.
Assumptions & free parameters
assumptions (5)
- standard math Zermelo-Fraenkel set theory with transfinite induction and foundation provides the von Neumann universe and the class ON.
- domain assumption Conway's cut-based definitions of addition, multiplication, and inverse on surreal numbers are valid and yield an ordered field.
- domain assumption The equivalence between the new NO and Conway's game-theoretic construction (Theorem 3.30) holds as in Siegel's book [S].
- domain assumption The graded von Neumann universe is independent of the chosen ordered-pair encoding.
- domain assumption Alling's full surreal number systems of height beta are isomorphic (Theorem 2.80).
invented entities (2)
-
Cocomplex numbers (Conway complex numbers) as sets of shuffle ordinals
-
Graded von Neumann universe with two membership relations
Cite this review
Pith. "Pith review of On Conway's Numbers and Games, the Von Neumann Universe, and Pure Set Theory." pith.science (2026). https://pith.science/paper/BD7EX3OA
@misc{pith2026250104412,
author = {Pith},
title = {Pith review of: On Conway's Numbers and Games, the Von Neumann Universe, and Pure Set Theory},
year = {2026},
howpublished = {\url{https://pith.science/paper/BD7EX3OA}},
note = {Machine review of arXiv:2501.04412}
}
read the original abstract
We take up Dedekind's question ''Was sind und was sollen die Zahlen?'' (''What are numbers, and would should they be?''), with the aim to describe the place that Conway's (Surreal) Numbers and Games take, or deserve to take, in the whole of mathematics. Rather than just reviewing the work of Conway, and subsequent one by Gonshor, Alling, Ehrlich, and others, we propose a new setting which puts the theory of surreal numbers onto the firm ground of ''pure'' set theory. This approach is closely related to Gonshor's one by ''sign expansions'', but appears to be significantly simpler and clearer, and hopefully may contribute to realizing that ''surreal'' numbers are by no means surrealistic, goofy or wacky. They could, and probably should, play a central role in mathematics. We discuss the interplay between the various approaches to surreal numbers, and analyze the link with Conway's original approach via Combinatorial Game Theory (CGT). To clarify this, we propose to call pure set theory the algebraic theory of pure sets, or in other terms, of the algebraic structures of the von Neumann universe. This topic may be interesting in its own right: it puts CGT into a broad context which has a strong ''quantum flavor'', and where Conway's numbers (as well as their analogue, the nimbers) arise naturally.
Reference graph
Works this paper leans on
-
[1]
Foundations of Analysis over Surreal Number Fields
Alling, Norman L. Foundations of Analysis over Surreal Number Fields. Mathematics Studies 141. North-Holland, Amsterdam 1987
work page 1987
-
[2]
Intermediate arithmetic operations on ordinal numbers
Altmann, H.J., "Intermediate arithmetic operations on ordinal numbers", Mathematical Logic Quaterly, 63, 3--4, (November 2017), 228--242 https://arxiv.org/abs/1501.05747v9
arXiv 2017
-
[3]
https://azinelibrary.org/other/Badiou_-_Number_and_Numbers.pdf
Badiou, A., Number and Numbers , Polity Press, Cambridge 2008 (translated from Le Nombre et les Nombres , Seuil, 1990). https://azinelibrary.org/other/Badiou_-_Number_and_Numbers.pdf
work page 2008
-
[4]
Vincent Bagayoko, Joris van der Hoeven, "Surreal substructures", https://arxiv.org/pdf/2305.02001
-
[5]
The hyperserial field of surreal numbers
Vincent Bagayoko, Joris van der Hoeven, "The hyperserial field of surreal numbers", https://arxiv.org/abs/2310.14873
-
[6]
Guy, Winning Ways , 2nd edition, Wellesley, Massachusetts: A
Berlekamp, E., Conway, J., and R. Guy, Winning Ways , 2nd edition, Wellesley, Massachusetts: A. K. Peters Ltd., 4 vols., 2001 -- 2004
work page 2001
-
[7]
Exponential fields and Conway's omega-map
Berarducci, A., and Salma Kuhlmann, Vincenzo Mantova, Mickaël Matusinski, "Exponential fields and Conway's omega-map", Proc.\ AMS, to appear, https://arxiv.org/abs/1810.03029
-
[8]
Berarducci, A., P.\ Ehrlich, and S.\ Kuhlmann, Mini-Workshop: Surreal Numbers, Surreal Analysis, Hahn Fields and Derivations , Oberwolfach Report 2016/60, https://ems.press/journals/owr/articles/15181
work page 2016
Show all 49 references
-
[9]
Surreal numbers, derivations and transseries
Berarducci, A., and V. Mantova, "Surreal numbers, derivations and transseries", J. Eur. Math. Soc. 20 (2018), 339--390 https://arxiv.org/abs/1503.00315
2018 arXiv
-
[10]
Bertram, W., Differential Geometry, Lie Groups and Symmetric Spaces over General Base Fields and Rings, Memoirs of the AMS 192, no.900 (2008) https://arxiv.org/abs/math/0502168
2008 arXiv
-
[11]
Graded sets, graded groups, and Clifford algebras
Bertram, W., "Graded sets, graded groups, and Clifford algebras", https://arxiv.org/abs/2109.00878v1, arxiv 2021
2021 arXiv
-
[12]
On group and loop spheres
Bertram, W., "On group and loop spheres", https://arxiv.org/abs/2410.17634
-
[13]
Arithmetic of ordinals with applications to the theory of ordered Abelian groups
Carruth, P.W., "Arithmetic of ordinals with applications to the theory of ordered Abelian groups", Bull. Amer. Math. Soc. 48 , (1942), 262 -- 271. Available here https://www.ams.org/journals/bull/1942-48-04/S0002-9904-1942-07649-X/S0002-9904-1942-07649-X.pdf
1942
-
[14]
Conway, J.H., On Numbers and Games (Second Edition), A K Peters Ltd., Wellesley 2001
2001
-
[15]
Conway, J.H., and R.\ Guy, The Book of Numbers , Springer, New York 1996
1996
-
[16]
Conway, J.H., and N.J.A.\ Sloane, Sphere Packings, Lattices and Groups , Springer Grundlehren 290, New York 1988
1988
-
[17]
Integration on the Surreals
Costin, O., and Ph.\ Ehrlich, "Integration on the Surreals", Adv.\ Math., to appear, https://arxiv.org/abs/2208.14331
-
[18]
See Numbers , Chapter 2, for more historical comments and references.)
Dedekind, R., Was sind und was sollen die Zahlen? Vieweg, Braunschweig 1888, http://www.opera-platonis.de/dedekind/Dedekind_Was_sind_2.pdf (This text is of historical interest, but does not contain the cut-construction of the reals. See Numbers , Chapter 2, for more historical...
-
[19]
Dehornoy, P., La th\'eorie des ensembles , Calvage et Mounet, Paris 2017
2017
-
[20]
An extended arithmetic of ordinal numbers
Doner, J, and A.\ Tarski, "An extended arithmetic of ordinal numbers", Fundamenta Mathematicae 6 (1969), 95 -- 127. http://matwbn.icm.edu.pl/ksiazki/fm/fm65/fm65110.pdf
1969
-
[21]
Ebbinghaus et al., Numbers , Springer (translated from Zahlen ), http://www.maths.ed.ac.uk/
-
[22]
Conway names, the simplicity hierarchy and the surreal number tree
Philip Ehrlich, "Conway names, the simplicity hierarchy and the surreal number tree", Journal of Logic and Analysis, vol. 3, pp. 1 -- 26 (2011)
2011
-
[23]
The absolute arithmetic continuum and the unification of all numbers great and small
Ehrlich, P., "The absolute arithmetic continuum and the unification of all numbers great and small", Bulletin of Symbolic Logic, Vol 18, no. 1, 2012
2012
-
[24]
Contemporary Infinitesimalist Theories of Continua and Their Late Nineteenth- and Early Twentieth-Century Forerunners
Ehrlich, P., "Contemporary Infinitesimalist Theories of Continua and Their Late Nineteenth- and Early Twentieth-Century Forerunners", in Stewart Shapiro, and Geoffrey Hellman (eds), The History of Continua: Philosophical and Mathematical Perspectives , Oxford, 2020; https://ar...
2020 arXiv
-
[25]
Recursive definitions on surreal numbers
Fornasiero, A., "Recursive definitions on surreal numbers", https://arxiv.org/abs/math/0612234
-
[26]
LNS 11 , Cambridge University Press, Cambridge 1986
Gonshor, H., An Introduction to the Theory of Surreal Numbers , London Math.\ Soc. LNS 11 , Cambridge University Press, Cambridge 1986
1986
-
[27]
Number Theory for the Ordinals With a New Definition for Multiplication
Gonshor, H., "Number Theory for the Ordinals With a New Definition for Multiplication", Notre Dame Journal of Formal Logic 21 (1980), 708 -- 710. link https://projecteuclid.org/journals/notre-dame-journal-of-formal-logic/volume-21/issue-4/Number-theory-for-the-ordinals-with-a-...
1980
-
[28]
Halmos, P., Naive Set Theory , Springer, New York 1974
1974
-
[29]
Numbers and Games
Hermes, H., "Numbers and Games", Chapter 13 in: Ebbinghaus et al., Numbers , see Numbers
-
[30]
Towards computable analysis on the generalised real line
Galeotti, L., and H.\ Nobrega, "Towards computable analysis on the generalised real line", https://arxiv.org/pdf/1704.02884.pdf
-
[31]
The Univalent Foundations Program: Homotopy Type Theory: Univalent Foundations of Mathematics , IAS, Princeton 2013, http://homotopytypetheory.org/book/
2013
-
[32]
The exponential-logarithmic equivalence classes of surreal numbers
Kuhlmann, S., and M.\ Matusinski, "The exponential-logarithmic equivalence classes of surreal numbers", Order 32 (1), pp 53-68 (2015), https://arxiv.org/pdf/1203.4538
2015 arXiv
-
[33]
Knuth, D., Surreal Numbers: How Two Ex-Students Turned on to Pure Mathematics and Found Total Happiness , 1974,
1974
-
[34]
Laugwitz, D., Zahlen und Kontinuum , BI Wissenschaftsverlag, Mannheim 1986
1986
-
[35]
Rethinking Set Theory
Leinster, T., "Rethinking Set Theory", The American Mathematical Monthly, 121:5, 403--415 https://arxiv.org/abs/1212.6543
-
[36]
Nim Multiplication
Lenstra, H.W., "Nim Multiplication", IHES publications, 78/211, IHES, 1978, https://scholarlypublications.universiteitleiden.nl/access/item
1978
-
[37]
Counting Sets with Surreals. Part I: Sets of Natural Numbers
Lynch, P., "Counting Sets with Surreals. Part I: Sets of Natural Numbers", https://arxiv.org/abs/2311.09951
-
[38]
Y.\ Manin, A Course in Mathematical Logic for Mathematicians , Scd.Ed., Springer GTM 53, New York 2010
2010
-
[39]
An algebraic (set) theory of surreal numbers, I
Rangel, D.R, and H.L. Mariano, "An algebraic (set) theory of surreal numbers, I", https://arxiv.org/pdf/1911.12726.pdf
1911 arXiv
-
[40]
The Ordinals as a Consummate Abstraction of Number Systems
Rea, A., "The Ordinals as a Consummate Abstraction of Number Systems", https://arxiv.org/abs/1706.08908
-
[41]
Analysis on Surreal Numbers
Rubinstein-Salzedo, S., and A. Swaminathan, "Analysis on Surreal Numbers", Journal of Logic and Analysis, Volume 6, Number 5, pp. 1--39, 2014. https://arxiv.org/abs/1307.7392
2014 arXiv
-
[42]
Ruckers, R., Infinity and the Mind , Princeton Science Library, Princeton 2005
2005
-
[43]
Schick, Moritz, Surreale Zahlen , Bachelorarbeit, Universit\"at Konstanz 2019
2019
-
[44]
An introduction to Conway's games und numbers
Schleicher, D., and M.\ Stoll, "An introduction to Conway's games und numbers", Moscow Math Journal 6 2 (2006), 359-388, https://arxiv.org/abs/math/0410026
2006 arXiv
-
[45]
Siegel, A.N., Combinatorial Game Theory , AMS Graduate Studies in Mathematics 146, AMS, Rhode Island, 2013
2013
-
[46]
Meet the Surreal Numbers
Simons, Jim, "Meet the Surreal Numbers", https://www.m-a.org.uk/resources/4H-Jim-Simons-Meet-the-surreal-numbers.pdf
-
[47]
The real numbers -- A survey of constructions
Weiss, I, "The real numbers -- A survey of constructions", Rocky Mountain J. Math. 45(3): 737 -- 762 (2015)
2015
-
[48]
Tegmark, M., ``The Mathematical Universe'', Foundations of Physics 38 2008, 101 -- 150, arxiv https://arxiv.org/abs/0704.0646v2
2008 arXiv
-
[49]
Tegmark, M., ``Our Mathematical Universe -- My Quest for the Ultimate Nature of Reality", Penguin Books, 20015
Reviewed August 10, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.