Pegasi Ascendant: Ranking Constellation Genitives on their Aesthetic Merit

Pranav Nagarajan · 2026

Read Pegasi Ascendant: Ranking Constellation Genitives on their Aesthetic Merit by Pranav Nagarajan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-00144.txt.

Introduction

Nagarajan ranks the genitive forms of all 88 constellations recognized by the International Astronomical Union, by beauty, from an anonymous survey of 74 astronomers. Each new pair comparison is picked by active sampling to gain the most information, and scores update under Bayesian scoring. Pegasi wins overall, just ahead of Centauri and Andromedae, while Orionis, the early favorite, lands seventh. Gruis comes last, and the author adds notes on the sound and cultural weight of the top and bottom five, with hints of taste gaps by age group.

Catalog record

CREATORPranav Nagarajan
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20897
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.00144

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Pegasi Ascendant: Ranking Constellation Genitives on their Aesthetic Merit Pranav Nagarajan Abstract Despite their ubiquity in the astronomical literature, there is no consensus tier list of the genitive forms of the 88 constellations officially recognized by the International Astronomical Union. To address this pressing open question, I conduct an anonymous pair comparison survey of 74 professional astronomers to rank these constellation genitives on their aesthetic merit. After each survey response, I use active sampling to select a new set of pair comparisons that maximizes expected information gain, and update overall scores based on a fully Bayesian framework. I find that Pegasi is the most aesthetically pleasing constellation genitive overall, narrowly edging out Centauri and Andromedae. While most astronomers self-report Orionis to be their top choice before taking the survey, this well-recognized constellation genitive only places seventh in the final ranking. Gruis, meanwhile, receives the dubious honor of last place. When breaking down the ranking by career stage, I find tentative evidence for generational differences in aesthetic taste. A larger sample of faculty members is needed to confirm this result. Finally, I offer unsolicited commentary on the phonetic appeal and cultural significance of the genitives ranked in the top and bottom five. show][email protected] I Introduction There are 88 constellations recognized by the International Astronomical Union (IAU) ( 2 ) . While it is unclear how many of these constellations would be recognized by professional astronomers on a game show, pride requires that that number is more than 12 (i.e., the zodiac constellations). 1 1 1 Otherwise, we would be astrologers. Together, these 88 constellations demarcate regions of the sky in right ascension and declination that (when taken all together) cover the entire celestial sphere ( 2 ) . In astronomy, stars are named after their parent constellations, typically using either the Bayer ( 1 ) or Flamsteed ( 3 ) designations. Furthermore, exoplanets are identified by adding lowercase letters to the names of their host stars, and meteor showers are also named after their constellation of origin. Importantly, the genitive forms of the constellations are used in these naming conventions, and constellation abbreviations are also based on genitive forms. Constellation genitives are thus important across a wide range of astrophysical disciplines. To date, the author is not aware of a definitive tier list of the 88 official constellations. I propose that any subjective ranking should be based on how aesthetically pleasing the genitive forms of these constellations are — in other words, which genitives sound the best? When spoken aloud, the names of astronomical objects should move the soul, and it is a noble cause to determine which constellation genitive(s) are best at achieving this goal. Indeed, auditory impact is of paramount importance in science outreach efforts. In this paper, I survey professional astrophysicists 2 2 2 I use the term “professional” loosely to include graduate students, who have arguably suffered enough to qualify. to determine a ranking of all 88 constellation genitives based on their aesthetic merit. I describe my survey methodology in Section II . I reveal the final ranking in Section III , and speculate on the reasoning of the respondents in Section IV . Finally, I provide directions for future investigation in Section V . II Methods To conduct an anonymous survey of professional astronomers, I vibe coded a web application. 3 3 3 The author thanks ChatGPT (version GPT-5.2) for (usually helpful) guidance during this iterative learning process. I used Flask, a lightweight Python-based framework, to handle the server-side logic. The responses were stored using SQLite, a file-based relational database engine. The user interface was implemented using HTML, CSS, and JavaScript. The application was hosted on PythonAnywhere and was accessible via a custom domain. I promoted the survey to my research group at Caltech, scientists who attended the “Stellar-Mass Black Holes at the Nexus of Optical, X-ray, and Gravitational Waves” program at the Kavli Institute of Theoretical Physics at UC Santa Barbara, and early career researchers on a Discord server dedicated to binaries. I intentionally targeted stellar astronomers, who are most likely to come across constellation genitives in their everyday work. The contacted astronomers were mostly located in North America and Europe. In the end, I received a total of 74 responses from 47 graduate students, 17 postdoctoral researchers, and 10 faculty members. The \approx 5 minute survey was designed as a series of 87 pair comparisons. To start, the survey collected information about the career stage of the participant, and optionally allowed the participant to report their favorite constellation genitive (in case they already had strong feelings about the issue). Each survey question then displayed two constellation genitives, requesting that the participant compare them head-to-head and choose the option that sounded more aesthetically pleasing. The responses were stored as directed win/loss records associated with each anonymous respondent. Figure 1: Scores of top-ranked constellation genitives, broken down by career stage. Each bar shows the score and 1\sigma uncertainty for genitives that appeared in at least one of the top 10 lists for students, postdocs, or faculty members. Before any data was collected, the pair comparisons were initialized as a uniformly random spanning tree over the 88 options. In detail, the 88 constellations can be thought of as representing the nodes of a graph with 88\times 87/2=3,828 possible edges. The algorithm to initialize the survey shuffled these edges and performed a union-find traversal over them, adding an edge if and only if its endpoints corresponded to two different connected components, and terminating once 87 pairs had been selected. In this way, it was guaranteed that every genitive appeared in at least one comparison in the initial survey. Once the initial survey response was recorded, selection of pair comparisons for each subsequent survey was performed using the computationally efficient Active Sampling for Pairwise Comparisons (ASAP) algorithm ( 6 ) . Given a global 88\times 88 pairwise comparison matrix M , where M_{ij} records the number of times option i was selected over option j so far, the ASAP algorithm chooses 87 pairs that maximize the expected information gain. Specifically, the algorithm selects a batch of comparisons that forms a minimum spanning tree of the inverse of information gain. After each survey response, the ASAP algorithm was also used to compute and update the ranking. In detail, the algorithm assumes that each score is a normal random variable, and utilizes a fully Bayesian framework to estimate the score and 1\sigma uncertainty of each constellation genitive. The current “Top 10” list was displayed on the “thank you” page following each survey. III Results As described in the previous section, I used the ASAP algorithm to extract estimated scores (and corresponding uncertainties) for each constellation genitive from the final, reconstructed pair comparison matrix. As telegraphed in the title of this paper, the winning constellation genitive is Pegasi. The top 10 constellation genitives are listed in Table 1 , and the bottom 10 constellation genitives are listed in Table 2 . The full ranking is provided in Table 3 in the Appendix. Table 1: Top 10 constellation genitives overall, ranked on aesthetic merit. Constellation Genitive Score Pegasi 1.032\pm 0.124 Centauri 0.983\pm 0.125 Andromedae 0.944\pm 0.126 Delphini 0.864\pm 0.124 Carinae 0.811\pm 0.124 Sagittarii

Read the complete text