This paper examines a fundamental but rarely articulated asymmetry in stellar astrophysics: while binarity can be positively established through observation, stellar singleness cannot be proven. The article argues that, for any individual star, it is in principle impossible to demonstrate the absence of all possible companions using finite observational means. The analysis shows that this limitation is not merely instrumental or technological, but epistemic. All detection methods—direct imaging, spectroscopy, astrometry, transits, eclipses, gravitational waves, and gravitational lensing—operate within constrained sensitivity domains and leave vast regions of binary parameter space unexplored. The absence of detected companions therefore reflects observational limits rather than physical isolation. Crucially, the paper distinguishes between epistemic undecidability in principle and operational adequacy in practice. Tight observational constraints—such as those achieved by CARMENES for GJ 486—can render remaining admissible companions dynamically or astrophysically negligible for specific modeling purposes. The category "single star" thus retains genuine physical utility within well-defined observational domains, even as it remains provisional and non-absolute at the epistemic level. The argument concerns the logic of classification and revision, not the promotion of hidden binarity as an unfalsifiable explanatory device. Detection limits depend critically on two factors: the mass of the primary star and its distance from the Sun. For nearby solar-type stars, the binary parameter space (period P, mass ratio q) has been probed quite deeply by combinations of methods. The epistemic argument is therefore strongest for distant stars, low-mass primaries, extreme mass ratios, and specific "blind corners" of parameter space that even multi-method surveys cannot reach. By examining the structure of binary parameter space, star formation theory, multiplicity statistics, and the role of compact objects such as black holes, the paper demonstrates that "single star" is not a physical category but a provisional observational status. The epistemic asymmetry between detection and non-detection leads to a one-directional logic of revision: stars may be reclassified from apparently single to binary, but never from binary to proven single. The argument is extended to population-level inference, showing that stellar catalogs represent a survivorship-biased census dominated by long-lived, low-mass stars, while massive stars rapidly disappear into dark remnants and undetectable binaries. As a result, true stellar population statistics are fundamentally underdetermined. Population statistics can constrain what is possible, but cannot serve as decisive evidence for star formation mechanisms. Sections on black holes, dark companions, and missing mass are presented as illustrative boundary cases demonstrating epistemic opacity, not as settled empirical claims. The paper further discusses anthropic bias, the concealment of dark stellar mass, the contribution of compact binaries to galactic mass budgets, and the limits of gravitational lensing as a corrective to electromagnetic bias. Together, these considerations support a methodological shift: binarity should be treated as a standing possibility for any star, while singleness should be regarded as an unprovable hypothesis rather than a default assumption. This work is intended as a methodological and conceptual contribution to stellar astrophysics and the philosophy of astronomical inference, clarifying the scope and limits of what observations can legitimately establish about stellar multiplicity. --- ACKNOWLEDGEMENTS: The author thanks José A. Caballero, Carlos Cifuentes San Román (Centro de Astrobiología, Madrid), Frédéric Arenou (Observatoire de Paris), and Andrei Tokovinin (NOIRLab/CTIO) for valuable correspondence that substantially improved this manuscript. Caballero drew attention to CARMENES results on stellar "singlicity" constraints, particularly the work on GJ 486 representing the tightest observational limits on companion exclusion to date. Cifuentes provided constructive criticism that helped clarify the crucial distinction between epistemic undecidability and operational adequacy, warning against the logical gap between "cannot be excluded in principle" and "should be treated as a standing alternative explanation"—a clarification now explicitly integrated into the argument. Arenou (Observatoire de Paris) confirmed from Gaia experience that "il y a de la marge partout" (there is room everywhere for hidden companions) and contributed the striking example of equal-mass twins ("jumelles parfaites") as a structural blind spot undetectable even by precision astrometry. Tokovinin emphasized that detection limits depend critically on primary mass and distance, noting that for nearby solar-type stars the (P, q) parameter space has been probed quite deeply by combinations of methods; he also cautioned against overstating the apparent decline of binary fraction with stellar age, which may reflect detection biases rather than true dynamical dissolution. --- KEYWORDS: stellar multiplicity, binary stars, single stars, observational astronomy, epistemology, detection limits, stellar statistics, star formation, astrometry, CARMENES, Gaia, black holes, dark matter, philosophy of science --- VERSION HISTORY: v1 (2026-01-08): Initial submissionv2 (2026-01-09): Integrated CARMENES references and acknowledgement of J.A. Caballerov3 (2026-01-10): Substantially revised based on expert feedback from C. Cifuentes and F. Arenou; clarified distinction between epistemic undecidability and operational adequacy; added methodological constraints section; reframed speculative extensions as illustrative boundary casesv4 (2026-01-11): Added discussion of mass- and distance-dependence of detection limits based on feedback from A. Tokovinin; clarified that epistemic argument is strongest for distant stars, extreme mass ratios, and specific blind corners of parameter space; added caution about interpreting apparent decline of binary fraction with age