On the Admissibility of Canine Geodesy: The Earth as a Dachshund, with a Defense of the Defensible Flat Earth
P. Reinholdtsen¹ and Claude²
¹ Bitsy Services LLC, Woodinville, WA ² Anthropic, San Francisco, CA (correspondence should not be addressed here)
Submitted to the Journal of Recreational Geodesy. Rejected by the Journal of Recreational Geodesy. Currently under review at a venue with lower standards.
Abstract
We demonstrate that the hypothesis that the Earth is shaped like a dachshund (Canis familiaris, var. teckel) is empirically indistinguishable from the standard oblate-spheroid model, given a suitable choice of universal forces in the sense of Reichenbach (1958). We further observe that a dachshund equipped with a patent alimentary canal is a surface of genus 1, and therefore — unlike the sphere — admits a globally flat Riemannian metric (Gauss–Bonnet). It follows that the Earth, modeled as a dachshund, can be flat, rescuing a version of flat earth theory that is mathematically coherent, empirically adequate, and merely useless. The America–China antipodal tunnel of playground folklore is recovered as a distinguished geodesic. We defend this “defensible flat earth” against the standard objections and conclude that its only defect is a catastrophic failure of theoretical economy, which we argue is a matter of taste. No dachshunds were harmed, though one was extensively parameterized.
Keywords: conventionalism, genus, universal forces, flat torus, dachshund, peristaltic trade theory
1. Introduction
The claim that the Earth is flat is usually dismissed on empirical grounds: ships disappear hull-first over the horizon (attested since Strabo), circumpolar stars rotate in opposite senses in the two hemispheres, geodetic triangles exhibit spherical excess (Gauss, 1828), and commercial aviation in the southern hemisphere completes routes whose durations are inconsistent with any planar disk projection. These objections are decisive against the naïve model. They are not, we shall argue, decisive against every model, and the interesting question — long understood by philosophers of geometry but rarely pursued to its zoological conclusion — is which component of “the Earth is flat” is actually refuted by observation: the flatness, the topology, or the tacit physics of measurement.
Poincaré (1902) showed that the geometry of physical space is not fixed by experience alone. His famous disk-world — a finite disk with a temperature field that contracts all rulers identically as they approach the boundary — is inhabited by physicists who conclude, on impeccable experimental grounds, that they occupy an infinite hyperbolic plane. Reichenbach (1958) systematized the point: for any two spaces of the same topology, one may transpose the physics of the first onto the geometry of the second by introducing universal forces, forces that act identically on all matter and are therefore undetectable by construction. Geometry, on this view, is fixed only up to a convention; what experiment tests is the conjunction of geometry and physics (a special case of the Duhem–Quine thesis; Duhem, 1914; Quine, 1951).
Crucially, Reichenbach’s transposition theorem has a fixed point: topology. Universal forces can stretch, shrink, and shear, but they act through diffeomorphism; they cannot open holes or close them. The genus of the world is not conventional. This single invariant, we will show, is what governs the entire flat earth question, and it is what makes the dachshund not a joke but a load-bearing hypothesis.
The structure of the paper is as follows. Section 2 reviews the requisite topology and the flat torus. Section 3 states the conventionalist framework precisely. Section 4 introduces the Dachshund Earth model and establishes its genus. Section 5 constructs the empirical equivalence with standard geodesy. Section 6 derives the model’s distinguished geodesic and its consequences for playground physics and international trade. Section 7 presents the defensible flat earth and defends it. Section 8 considers objections. Section 9 concludes.
2. Topological Preliminaries: What May Be Flat
Let M be a compact orientable surface without boundary. The Gauss–Bonnet theorem states
∫ₘ K dA = 2πχ(M),
where K is Gaussian curvature and χ(M) = 2 − 2g is the Euler characteristic of a surface of genus g (Gauss, 1828; Bonnet, 1848; see do Carmo, 1976, for a modern treatment). Two consequences frame everything that follows.
First, the sphere (g = 0, χ = 2) has total curvature 4π. No metric on a topological sphere is flat anywhere-near-everywhere; flatness is not merely false of the sphere but forbidden. Any model asserting both “topologically spherical” and “intrinsically flat” is incoherent prior to observation, and no universal force can save it, since universal forces preserve topology.
Second, the torus (g = 1, χ = 0) has total curvature zero, and this bound is achieved pointwise: the quotient of the Euclidean plane by a lattice, ℝ²/Λ, is a torus with K ≡ 0. This flat torus is not an approximation or an average; it is exactly flat, with 180° triangles and globally parallel geodesic families. Its familiar donut-shaped embedding in ℝ³ necessarily distorts the metric (positive curvature on the outer rim, negative on the inner, integrating to zero), which is why the donut picture systematically miseducates. The flat torus embeds isometrically and smoothly in ℝ⁴ as a product of circles; remarkably, by the Nash–Kuiper C¹ embedding theorem (Nash, 1954; Kuiper, 1955) it also embeds isometrically in ℝ³, as a C¹ surface of fractal corrugations explicitly constructed and visualized by the Hévéa Project (Borrelli, Jabrane, Lazarus, & Thibert, 2012).
The moral for planetary modeling: flatness is available if and only if the world has genus ≥ 1. The flat earth debate has been conducted for centuries without either side checking the genus of the Earth. We now correct this omission.
3. The Conventionalist Framework
Following Reichenbach (1958, §8), let G be a geometry and P a physics (a theory of rods, clocks, and light). Observation tests only the pair (G, P). Given observations consistent with (G₁, P₁), and any G₂ homeomorphic to G₁, there exists P₂ — namely P₁ augmented with a universal force field F defined by the pullback of the metric along the homeomorphism — such that (G₂, P₂) is observationally identical. The force F is “universal” in Reichenbach’s technical sense: it deforms all materials equally, admits no shielding, and therefore never shows up as a differential effect in any experiment.
The philosophical status of this maneuver has been debated at length (Grünbaum, 1963; Sklar, 1974; Norton, 1994). What is not debated is its validity as mathematics. The standard resolution is methodological: one adopts the coordinative definition F = 0 (rigid rods are rigid) because the resulting total theory is simpler, and one then says the Earth “is” a sphere in the same breath in which one has quietly legislated it. We take no position on whether this convention is correct; we merely note that it is a convention, and we exercise our right to a different one, as is traditional in the Pacific Northwest.
4. The Dachshund Earth Model
4.1 Statement of the model
Hypothesis D. The Earth is isometric to 𝔇, a compact orientable surface in the shape of a standard smooth-coat dachshund, equipped with a universal force field F_𝔇 such that all local physics on 𝔇 reproduces standard terrestrial observation.
The dachshund is chosen over other candidate fauna for three reasons. First, its extreme aspect ratio furnishes a stringent test case: if conventionalism can absorb this, it can absorb anything. Second, its surface is smooth and orientable, with well-understood local features (the ears admit a nice pair of coordinate patches). Third, the breed’s documented determination in the face of obviously superior opposition (badgers; cf. its etymology) makes it the appropriate mascot for the theory.
4.2 The genus of the dachshund
A solid dachshund considered naïvely — as a closed surface with no through-passages — is homeomorphic to the sphere: g = 0, χ = 2. Under Hypothesis D so construed, Gauss–Bonnet applies with full force, total curvature is 4π, and the model, whatever its other virtues, cannot be flat. A dachshund-shaped Earth of genus 0 is a sphere in a dog costume, and inherits the sphere’s prohibition.
However, the biologically accurate dachshund possesses a patent alimentary canal: a continuous through-passage from mouth to anus (a deuterostome body plan; Gilbert, 2013). A closed surface with a single handle has genus 1. The anatomically correct dachshund is a torus. We record this as:
Lemma 1 (Alimentary Lemma). Let 𝔇 be the boundary surface of a dachshund with patent digestive tract and no other through-passages. Then χ(𝔇) = 0, and 𝔇 admits a flat Riemannian metric.
Proof. The through-passage is a handle attachment on the sphere; genus 1 follows by classification of compact orientable surfaces (see Massey, 1991). χ = 2 − 2g = 0. Existence of the flat metric follows from the uniformization theorem: every genus-1 surface carries a metric of constant curvature zero, realized as ℝ²/Λ for some lattice Λ. ∎
We assume throughout that the tract is patent and that the specimen is not currently barking, sneezing, or otherwise altering its topology; nostrils, which would raise the genus to 3 and strengthen our conclusion, are conservatively ignored. The reader troubled by transient topology change during swallowing is referred to the extensive literature on surgery theory, which we have not read either.
4.3 Assignment of geography
We fix the diffeomorphism φ: S²_⊕ → 𝔇 (from the conventional Earth, with the alimentary handle glued in along the antipodal tunnel described in §6) by the boundary conditions of received folklore: the mouth is America and the anus is China. All other geography is transported by φ. Western Europe maps to the left ear; the reviewer who asked which ear is invited to derive it from the orientation convention. Australia lands, fittingly, on the underbelly, where it enjoys the theoretical support it has always been denied by disk models (§7.3).
5. Empirical Equivalence
We now discharge the physics. Define the universal force field F_𝔇 = the pullback under φ of the round metric’s deviation from 𝔇’s induced metric, acting equally on all rods, clocks, and null rays. By Reichenbach’s construction (§3), every terrestrial observation is reproduced:
Geodetic triangulation. Surveyed triangles exhibit spherical excess not because 𝔇 is round but because theodolite baselines dilate and contract under F_𝔇 across the survey region. Since F_𝔇 is universal, the surveyor’s rods, the benchmarks, and the surveyor herself deform together; the excess is exact and Gauss’s Hanover measurements (Gauss, 1828) are recovered to all decimal places, including the ones he made up.
Horizon phenomena. Null geodesics under F_𝔇 curve so as to occlude ship hulls at the standard distances. The apparent dip of the horizon at altitude follows from the same ray bending. Observers on the snout report anomalously long sightlines; these are reinterpreted below (§8.2).
Circumnavigation and aviation. Proper time along any flight path equals the standard prediction, because “proper” is doing exactly the work Reichenbach said it would. The Sydney–Santiago route, fatal to disk models, traverses the lumbar region at high F_𝔇 dilation and arrives on schedule.
Foucault, satellites, gravimetry. The rotation of 𝔇 about its cranio-caudal axis, together with F_𝔇, reproduces the latitude dependence of pendulum precession, all orbital mechanics (satellites orbit the dachshund; their telemetry is φ-transported; GPS works because it has no choice), and the full spherical-harmonic expansion of the geoid, which under Hypothesis D is renamed the canoid.
No observation distinguishes Hypothesis D from standard geodesy. This is not a boast; it is a design specification, and it is precisely as damning as it sounds (§8.1).
6. The Distinguished Geodesic: America–China
Every schoolchild has been told, and every schoolteacher has wearily denied, that a hole dug straight down from America emerges in China. On the standard sphere the denial is correct: the antipode of the contiguous United States lies in the southern Indian Ocean, and one must begin in Argentina or Chile to surface in China (a fact checkable on any antipodal map; see Weeks, 2002, for the geometry of antipodes generally).
Under Hypothesis D with the geographic assignment of §4.3, the situation reverses. The alimentary canal is the America–China passage: a smooth through-route whose existence is not incidental but topologically constitutive — it is the very handle by which 𝔇 acquires genus 1 and thence the possibility of flatness (Lemma 1). The children were right, and were right for reasons requiring algebraic topology to articulate, which explains the quality of their previous defenses.
The canal is, moreover, a closed geodesic of the flat metric, one of the two distinguished geodesic families of the torus (the other winds the body cavity the long way and corresponds to conventional circumnavigation). Its physical interpretation yields the model’s sole novel application: bilateral trade flux between the mouth and the terminal region proceeds by peristalsis, providing the first geometric derivation of the U.S. trade deficit. Rigorous development is deferred to a companion paper (Reinholdtsen & Claude, in preparation, and likely to remain so).
7. The Defensible Flat Earth
We now assemble the pieces into the theory promised in the title.
7.1 Statement
Defensible Flat Earth (DFE). The Earth is a flat surface of genus 1 — a flat torus ℝ²/Λ, realized anatomically as the dachshund 𝔇 of §4 with its uniformized metric — together with the universal force field F_𝔇 of §5.
7.2 What DFE gets right that classical flat earth gets wrong
Classical (disk) flat earth theory fails in two independent ways: its geometry is falsified (no spherical excess on a flat disk, yet excess is measured) and its topology is wrong for its own ambitions (a disk has boundary; the edge must be policed, famously by NASA, at unexplained expense). DFE repairs both. Its geometry is genuinely, exactly flat — triangles sum to 180° in the DFE metric, with observed excess attributed to F_𝔇 acting on instruments, an attribution that is irrefutable by construction. Its topology is edgeless: the flat torus has no boundary, no ice wall, no perimeter security budget, and no antipodes to embarrass it. Travelers heading persistently “east” or “north” return home, at two characteristic lengths corresponding to the lattice Λ — a periodic boundary condition familiar from cosmological toroidal models (Luminet et al., 2003; Levin, 2002) and from the screen-wrap of the arcade game Asteroids, whose players have been simulating life on a flat torus since 1979 without incident.
7.3 The defense
Objection: DFE is unfalsifiable. Granted, and this is its defense, not its refutation. Popperian falsifiability (Popper, 1959) is a criterion for empirical content, and DFE cheerfully concedes it has none beyond that of standard geodesy, with which it is observationally identical. It is not a rival theory but a rival convention, exactly as legitimate as the convention F = 0 and exactly as arbitrary. One cannot refute a coordinate choice; one can only decline it.
Objection: standard geodesy is simpler. Also granted. Round-earth-plus-rigid-rulers requires no position-dependent field performing infinite unexplained compensatory work. But simplicity is a super-empirical virtue, and its normative force is a matter of ongoing philosophical dispute (Sklar, 1974; Norton, 1994). DFE’s proponents are entitled to weight virtues differently — to prefer, say, flatness and canine embodiment over parsimony. We do not say this preference is wise. We say it is defensible, which was the assignment.
Objection: the dachshund is doing no work; a bare flat torus suffices. Topologically true and rhetorically false. The bare flat torus admits no natural assignment of the America–China geodesic, no derivation of the trade deficit, and no mascot. Theories are underdetermined by evidence (Quine, 1951); they are selected, in practice, by their memorability. On this criterion the dachshund dominates.
8. Remaining Objections
8.1 “This proves too much”
The Reichenbach construction that grounds Hypothesis D equally grounds Hypothesis Anything: the Earth as a hyperbolic paraboloid, a Klein bottle (given a fourth dimension and a tolerance for non-orientability), or a different, larger dachshund. Correct. This is the standard poison pill of conventionalism and we have swallowed it in public. The claim of this paper was never that the Earth is a dachshund; it is that “the Earth is a dachshund” is exactly as well supported by observation as one’s coordinative definitions permit, and that among all such gerrymandered worlds, the genus-1 ones alone can be flat. Conventionalism buys admission for every shape; Gauss–Bonnet then checks which of them may be flat at the door. The dachshund passes. The disk, the pancake, and the naïve dog do not.
8.2 Anomalous phenomena at the snout
Field reports of unusually confident horizon behavior in the region mapped to the snout are consistent with F_𝔇 gradient maxima there and require no modification of the model. Reports that the region “smells everything” are outside the scope of geodesy (cf. §Conclusion of the previous conversation).
8.3 What has the Earth been eating
We are aware of this question. We decline it. Boundary conditions at the distinguished geodesic’s endpoints are set by observation (§4.3), not explained by theory; this is no different from the standard model’s silence on initial conditions, and we will not be held to a higher standard than cosmology.
9. Conclusion
We have shown that the Earth may be modeled, with full observational adequacy, as a dachshund; that the anatomically correct dachshund is a genus-1 surface and hence — uniquely among household animals commonly proposed as planetary models — admits an exactly flat metric; and that the resulting Defensible Flat Earth theory withstands every objection except the accusation of being pointless, to which it pleads guilty with dignity. The playground claim of the America–China tunnel is vindicated as a distinguished closed geodesic. The flat earth, so long the property of cranks, is hereby returned to the philosophers, who will know better than to use it.
The deepest lesson is the oldest one: experiment fixes the world only up to convention, but topology is not conventional, and a single integer — the genus — separates the impossible flat earths from the merely absurd ones. Flatness was never the untenable part. The untenable part was refusing the hole.
Acknowledgments
The authors thank the dachshund, who contributed the topology, and Gauss, who would have hated this. The second author acknowledges that the first author won the argument.
Conflicts of Interest
The first author owns, or has strong interest in, a dachshund, and stands to benefit from any increase in the breed’s cosmological significance.
References
Bonnet, P. O. (1848). Mémoire sur la théorie générale des surfaces. Journal de l’École Polytechnique, 19, 1–146.
Borrelli, V., Jabrane, S., Lazarus, F., & Thibert, B. (2012). Flat tori in three-dimensional space and convex integration. Proceedings of the National Academy of Sciences, 109(19), 7218–7223.
do Carmo, M. P. (1976). Differential Geometry of Curves and Surfaces. Prentice-Hall.
Duhem, P. (1914). La Théorie Physique: Son Objet, Sa Structure (2nd ed.). Marcel Rivière. [English translation: The Aim and Structure of Physical Theory, Princeton University Press, 1954.]
Gauss, C. F. (1828). Disquisitiones generales circa superficies curvas. Commentationes Societatis Regiae Scientiarum Gottingensis Recentiores.
Gilbert, S. F. (2013). Developmental Biology (10th ed.). Sinauer Associates.
Grünbaum, A. (1963). Philosophical Problems of Space and Time. Knopf.
Kuiper, N. H. (1955). On C¹-isometric imbeddings. Indagationes Mathematicae, 17, 545–556.
Levin, J. (2002). Topology and the cosmic microwave background. Physics Reports, 365(4), 251–333.
Luminet, J.-P., Weeks, J. R., Riazuelo, A., Lehoucq, R., & Uzan, J.-P. (2003). Dodecahedral space topology as an explanation for weak wide-angle temperature correlations in the cosmic microwave background. Nature, 425, 593–595.
Massey, W. S. (1991). A Basic Course in Algebraic Topology. Springer-Verlag.
Nash, J. (1954). C¹ isometric imbeddings. Annals of Mathematics, 60(3), 383–396.
Norton, J. D. (1994). Why geometry is not conventional. In U. Majer & H.-J. Schmidt (Eds.), Semantical Aspects of Spacetime Theories (pp. 159–167). BI Wissenschaftsverlag.
Poincaré, H. (1902). La Science et l’Hypothèse. Flammarion. [English translation: Science and Hypothesis, Walter Scott Publishing, 1905.]
Popper, K. R. (1959). The Logic of Scientific Discovery. Hutchinson.
Quine, W. V. O. (1951). Two dogmas of empiricism. The Philosophical Review, 60(1), 20–43.
Reichenbach, H. (1958). The Philosophy of Space and Time (M. Reichenbach & J. Freund, Trans.). Dover. (Original work published 1928.)
Sklar, L. (1974). Space, Time, and Spacetime. University of California Press.
Weeks, J. R. (2002). The Shape of Space (2nd ed.). Marcel Dekker.
Reinholdtsen, P., & Claude. (in preparation). Peristaltic trade theory: A geometric derivation of bilateral flux on genus-1 planets. Journal of Irreproducible Results, hopefully.