# Alexandria around 250 BCE: historical basis and reconstruction limits

Research date: 1 September 2026. Companion data: `src/evidence.js`.

## What the reconstruction can honestly claim

This is a source-led **possible setting for the royal Library and Mouseion**, not a recovered architectural plan. The distinction is essential to accuracy: it is not a disclaimer that licenses arbitrary architecture. Surviving evidence constrains the date, institutional setting, activities and material culture; analogy constrains architectural forms; the exact spatial arrangement remains a design decision.

“Around 250 BCE” places the visit in the reign of Ptolemy II. It selects a flourishing early phase rather than asserting that scholarship has identified one uniquely measurable peak year. Later ancient witnesses can inform the modern evidence panel without becoming books, people or practices present in the simulated year.

The audit should therefore ask both **is this plausible for the period?** and **does the interface tell the visitor why it is here?** Merely placing a source link below an overconfident claim is insufficient.

## Three evidence levels

| Label | Meaning | What it does not mean |
| --- | --- | --- |
| Attested | The particular feature, text or testimony is directly documented. | Every detail of the associated model is confirmed, or a late witness is automatically reliable. |
| Inferred | A reasoned extension across date, site or context. | There is a contemporary description of this exact room. |
| Conjectural | The evidence does not determine the design decision. | A fantasy setting exempt from period constraints. |

All dimensions, room arrangements, furnishings and sightlines are conjectural. An attested papyrus book format does not authenticate the cupboard containing it. An attested description of a dining place centuries later does not authenticate a dining room in 250 BCE.

## The principal witnesses, and how far they take us

### Strabo: the place and the community

In *Geography* 17.1.8, Strabo says: “The Museum is also a part of the royal palaces.” He goes on to name a walk, a seated exedra and a shared dining place. His text belongs to the late first century BCE/early first century CE, roughly two centuries after this visit. He does **not** provide a plan, measurements, a reading-room description or a catalogue. His adjacent coastal description is useful for the relationship between the royal district, harbours and Pharos, not for a precise view from a Library window. [Primary text, H. L. Jones translation](https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Strabo/17A1*.html).

A common error is to combine everything ever said about ancient libraries into “Strabo’s description of the Library.” This project must not do that. The terms *Mouseion* and Library are related, but their physical relationship is not securely recoverable.

### Timon: an earlier trace of supported scholarly life

Athenaeus, writing around 200 CE, preserves third-century BCE verse by Timon mocking scholars fed in the “Muses’ bird-cage.” This is useful earlier evidence for a maintained scholarly community. It is satire, survives through later transmission, and does not establish a roster, building form, number of scholars or daily timetable. [Athenaeus 1.22d](https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Athenaeus/1B*.html).

Together, Timon and Strabo justify giving communal support a place in the experience. The couches, food, table arrangement and occupants of the reconstructed dining room still require conjectural labels.

### Pseudo-Aristeas and Galen: famous acquisition stories

The *Letter to Philocrates*, usually dated to the second century BCE, presents a story about the translation of Jewish law under Ptolemy II. Sections 9–11 report enormous book totals; 28–32 describe the supposed translation proposal. It is evidence for a later conception of royal collecting, not a surviving administrative file. [Pseudo-Aristeas, translation](https://www.attalus.org/translate/aristeas1.html).

Galen’s *Commentary on Epidemics III* (Kühn XVIIa, 606–608) reports books “from the ships” and a king retaining Athenian tragic texts while returning copies. He wrote in the second century CE. The passages are important but do not prove that every arriving ship was searched in 250 BCE, or that a particular present-day edition reproduces a confiscated roll. A compulsory port-confiscation game mechanic would turn a disputed late anecdote into routine fact. [Galen, translated extracts](https://www.attalus.org/translate/extracts.html).

The date of the Greek Pentateuch and the historicity of a royal-library translation commission are separate questions. A complete modern Septuagint is especially unsafe as a single contemporary shelf object: its constituent translations have different dates.

### Callimachus: bibliography without an extant catalogue

The tenth-century *Suda* κ 227 lists Callimachus’ works and reports the *Pinakes* in 120 books. Its critical notes also flag faulty chronology in the biographical entry. This late witness can establish a received title and tradition; it is not the *Pinakes* themselves. [Suda On Line, κ 227](https://cs.uky.edu/~raphael/sol/sol-cgi-bin/search.cgi?db=REAL&enlogin=&field=adlerhw_gr&login=&searchstr=kappa%2C227).

Susan Stephens describes the *Pinakes* as a classified account of earlier Greek literature with information about authors and the beginnings of works. Her introduction also warns how much of Callimachus’ biography is inference. Do not identify Callimachus confidently as the Library’s head, or reproduce a modern eleven-category list as if it were an excavated catalogue. [Stephens, “Callimachus of Cyrene”](https://dcc.dickinson.edu/callimachus-aetia/callimachus).

Max Leventhal discusses surviving evidence for literary classification and counting lines. The searchable collection can evoke bibliographical thinking, but it is our interface, not a restored ancient information system. Scholars differ in how directly they describe the *Pinakes* as a Library inventory; the visitor-facing text preserves that distinction. [Leventhal, *Poetry and Number in Graeco-Roman Antiquity*, chapter 1](https://www.cambridge.org/core/books/poetry-and-number-in-graecoroman-antiquity/callimachus-and-his-legacy/CBB1BC31F2E1CE3E9D359A1525DBE1C9).

### Modern historical criticism

Roger Bagnall critically examines foundation traditions, numerical claims, proposed equipment identifications and destruction narratives. In particular, his pp. 353–355 reject treating the inscribed “Dioskourides” granite block as securely identified Library roll storage; pp. 356–360 explain the uncertainty surrounding destruction and continued copying. Do not replace the popular confident narrative with an equally confident alternative destruction date. [Bagnall, “Alexandria: Library of Dreams,” 2002](https://archive.nyu.edu/handle/2451/28263).

Andrew Erskine situates the institutions in Ptolemaic royal competition and Greek cultural identity in Egypt. This guards against portraying a neutral modern university or universal public lending service. His interpretation of the politics should be described as scholarship, not as a recovered ancient policy statement. [Erskine, “Culture and Power in Ptolemaic Egypt,” 1995](https://www.pure.ed.ac.uk/ws/portalfiles/portal/11871537/Culture_and_Power_in_Ptolemaic_Egypt_the_Library_and_Museum_at_Alexandria.pdf).

## Claim-to-model register

| Evidence ID | Basis | Boundary that must remain visible |
| --- | --- | --- |
| `royal-quarter` | Strabo’s location; historical analysis of royal patronage. | The precise Library site and boundaries are unknown. |
| `courtyard` | Hellenistic palatial peristyles as comparisons; Strabo’s walk. | Neither this court nor its garden, basin or proportions is recovered. |
| `book-room` | Royal collecting in literary testimony; papyrus book culture. | No known plan, cupboard design, capacity or exact shelf arrangement. |
| `exedra` | Named in Strabo’s later description. | Presence in 250 BCE is an inference; its form and placement are conjecture. |
| `dining` | Later Strabo plus Timon’s earlier satire preserved by Athenaeus. | Furniture, menu, staffing and schedule are not documented. |
| `copying` | Handwritten transmission and writing materials; late Galen. | A separate copying department and universal ship-search procedure are unproven. |
| `harbour` | Ancient coastal description and Pharos research. | The terrace, distance compression, skyline and precise sightline are invented. |
| `pinakes` | Titles and fragments preserved by later authors; modern philology. | Not an intact catalogue, and not securely dated to this exact year. |
| `rolls` | Egyptian papyri and book history. | One roll is not necessarily one complete work; no Library equipment is reproduced as proven. |
| `holdings` | Individual work histories, editorial associations and plausibility. | A selected reading collection is not a recovered inventory. |
| `losses` | Quotations, summaries, papyri and ancient title lists. | “Lost” does not mean “destroyed here,” “no surviving fragment” or “only copy.” |

### Architecture and objects

Andreas Scholl’s contribution to the Met’s 2016 *Pergamon and the Hellenistic Kingdoms of the Ancient World*, p. 50, discusses peristyles in Hellenistic palace architecture. That is a comparative constraint, not permission to import a second-century Pergamene building unchanged into third-century Alexandria. A planted court needs an additional conjectural label: a peristyle is not necessarily a garden. [Met scholarly catalogue, PDF](https://resources.metmuseum.org/resources/metpublications/pdf/Pergamon_and_the_Hellenistic_Kingdoms_of_the_Ancient_World.pdf).

The University of Michigan collection exhibition supports papyrus, ink, reed writing implements, ostraca and wax tablets. Its displayed material spans Graeco-Roman Egypt; do not imply that every individual exhibited object dates to 250 BCE. [“Media for Writing in Graeco-Roman Egypt”](https://apps.lib.umich.edu/files/collections/papyrus/exhibits/writing/media.html).

Janice Kamrin’s Met essay explains papyrus manufacture and long use in Egypt. It is not evidence for the Library’s storage system. Its discussion of changing Egyptian writing implements also should not be indiscriminately substituted for the history of Greek bookhands. [“Papyrus in Ancient Egypt”](https://www.metmuseum.org/essays/papyrus-in-ancient-egypt).

Adam Bülow-Jacobsen’s chapter in the 2009 *Oxford Handbook of Papyrology* describes ordinary roll heights of about 20–30 cm within a wider range. This prompted a material-scale correction to the initial model: ordinary stored rolls should not all have the unusually large heights of 50–60 cm. The final dimensions are illustrative, not measurements of Alexandrian finds. [“Writing Materials in the Ancient World,” 3–29, Duke-hosted chapter](https://people.duke.edu/~wj25/Greek_586_Literary_Papyrology/Assignments_files/WJ%20Papyrology%20CH%201.pdf).

The writing texture on open 3D papyrus repeats the opening of the *Iliad* in unaccented Greek capitals. It is decorative typography, not an image of a recovered manuscript or a claim about a particular Alexandrian editor’s text. The identified reading editions are available separately through the catalogue.

Period checks: no modern book spines or printed pages in the 3D world; no quill-pen shorthand for antiquity; no modern glazed roof or baroque multilevel reading hall; no later scientific instrument simply because its inventor was Alexandrian. Modest oil lamps are plausible objects but do not establish an ancient lighting scheme. Avoid treating cinematic dust, glare and acoustic effects as reconstructed evidence.

### Interior decoration and the living scene

The interiors use painted plaster fields, palmette and rosette ornament, narrow architectural mouldings, and floral and geometric pebble floors. These are original designs constrained by Classical and early Hellenistic comparisons, not reproductions of excavated Library decoration. The Hellenic Ministry of Culture's guide to Pella describes coloured plaster and mosaic floors in reception and banquet rooms. That establishes a comparative decorative vocabulary, not a claim that these Macedonian rooms, motifs or colours occurred in Alexandria in exactly 250 BCE. [Pella archaeological site guide](https://culturalegnatia.culture.gov.gr/en/route-stops/pella/).

Decoration is differentiated by room: the book room has a painted upper frieze and cupboard-height fields, the copying room is quieter, and the exedra and dining room have stronger ornamental surrounds. The choices do not reproduce later Pompeian narrative frescoes, a named portrait gallery, a known cult statue or an excavated Muses cycle. The small floor designs are not an assertion of wall-to-wall mosaic luxury. Doors and working aisles remain clear.

Residents are anonymous illustrative people. Clothing, skin tone, hairstyle and voice do not identify ethnicity or legal status. Walking, carrying, writing and conversation make activities legible, but no pose or conversation reconstructs an attested individual. The people are stylized 3D characters, not documentary portraits. Audio voices, environmental sounds and every everyday exchange are synthetic dramatizations; see [Voices in Alexandria](audio-notes.md). Narration can be selected independently for the current room or heard in an optional tour. It is not required to visit, read or explore.

### Pharos and the separate Serapeum

The CEAlex Pharos project reports completion under Ptolemy II, using the much later chronology of Eusebius. This supports a lighthouse in the period. It does not authenticate every detail of a modern three-stage silhouette, its precise height or the view from a hypothetical Library terrace. [CEAlex Pharos Project](https://pharos.cealex.org/the-lighthouse-of-alexandria/).

McKenzie, Gibson and Reyes analyze the distinct Ptolemaic and Roman architectural phases of the Serapeum. Its archaeological remains cannot be relabelled as the royal Library. The age and form of the Serapeum’s book collection require separate evidence. This project does not merge the sanctuaries or import the late-antique Serapeum narrative into 250 BCE. [“Reconstructing the Serapeum in Alexandria from the Archaeological Evidence,” 2004](https://www.cambridge.org/core/journals/journal-of-roman-studies/article/abs/reconstructing-the-serapeum-in-alexandria-from-the-archaeological-evidence/2B41066E2FD232545126880FC6D0E5F3).

## Reading collection: rules for historical accuracy

For every work, distinguish four propositions:

1. The ancient work existed, with this authorship or traditional attribution.
2. It plausibly existed by the chosen date.
3. There is a stated strength of association with Alexandrian scholarship or collecting.
4. The offered text is an identified surviving edition, translation, excerpt or fragment.

These are not interchangeable. A surviving manuscript is not evidence of a copy on these shelves. A work’s study by an Alexandrian editor does not prove a surviving modern text reproduces that editor’s precise version. A later author quoting an early text belongs in the source apparatus, not among the contemporary shelf objects.

Modern translation prose, punctuation, paragraphing and section numbering are reading aids. A work offered as complete must be complete **in the specified edition and surviving state**, not necessarily as originally composed. Thucydides is the surviving unfinished history; a single Theocritean idyll can be a complete poem without implying a complete authentic Theocritus corpus. Later introductions and anthology neighbours must not be silently included as ancient contemporary content.

### Chronological watchlist

| Item | Required treatment |
| --- | --- |
| Callimachus, *Aetia* I–II | Date uncertain; commonly placed earlier than III–IV. Label the assumption. |
| *Aetia* III–IV and *Lock of Berenice* | Post-246 BCE context; not a contemporary 250 BCE reading. |
| A collected Callimachus or Theocritus anthology | Extract only the specified work; do not include later pieces or pseudepigrapha silently. |
| Sappho’s nine-book or Alcaeus’ ten-book arrangement | Archaic songs predate the visit; the standardized later editorial arrangement need not. |
| Aristotle’s transmitted collections | Date and authorship of appended material require scrutiny; *History of Animals* X cannot be assumed authentic. |
| Euclid | Ancient mathematical text; later testimony for the familiar Alexandrian biography should remain identifiable as later. |
| Archimedes’ mature treatises and Eratosthenes’ mature geography | Their later-third-century fame does not make those works available by 250 BCE. |
| Apollonius of Perga, Hipparchus, Hero, Claudius Ptolemy, Hypatia | Later figures, not occupants or authors of contemporary shelf texts. |
| The entire Septuagint | Multiple translation dates; do not treat it as a single completed 250 BCE project. |
| The Greek Pentateuch | An explicitly borderline selection: completion of all five Greek books by 250 BCE is uncertain. |
| Theophrastus, *Enquiry into Plants* | Nine surviving books, not the separate *Plant Explanations*. Identify the minor works appended in Hort’s second modern volume separately. |

For the first two rows, Stephens explicitly explains that III–IV belong early in Ptolemy III’s reign, while I–II are not securely dated. She also identifies the papyrus and quotation witnesses to the incomplete *Aetia*. [Date and transmission discussion](https://dcc.dickinson.edu/callimachus-aetia/introduction).

The added Greek Pentateuch is not a claim that the entire Septuagint existed in the reign of Ptolemy II. Greenspoon and Ross’s introduction places the five-book translation in Ptolemaic Egypt in the middle or later third century BCE, probably around Alexandria. Its inclusion in this approximately 250 BCE reading collection is therefore explicitly borderline. James K. Aitken proposes a scribal setting in Egypt, particularly the Fayum, and a date in the middle or second half of the century; that is a scholarly argument, not proof of either a Fayum or a royal Library commission. Neither source makes the later *Letter of Aristeas* a contemporary acquisition record. The Brenton reading copy descends from later Greek textual transmission and includes Genesis through Deuteronomy only. [Greenspoon and Ross, “Septuagint,” introduction, version of 2024](https://academic.oup.com/reference/62341/reference-article-abstract/561720572); [Aitken, “The Ptolemaic Setting for the Translation of the Greek Pentateuch,” 2020, accepted version](https://www.repository.cam.ac.uk/items/b34fdcd3-1095-483f-8460-af288e6a626c).

Theophrastus died around 287 BCE. His botanical writing belongs before the visit, but a later printed edition is not an unchanged authorial manuscript: the *Enquiry* was revised and is associated with teaching material. Its nine surviving books are distinct from the six surviving books of *Plant Explanations*. Hort’s 1916 second volume also prints *On Odours* and *On Weather Signs*; the catalogue identifies those as separate works in a modern binding, not additional books of the *Enquiry*. [Katerina Ierodiakonou, “Theophrastus,” especially sections 1 and 5](https://plato.stanford.edu/entries/theophrastus/).

### Lost and fragmentary works: useful witness groups

The corpus should contain specific witnesses, not just generalized claims that “ancient sources mention it.” Useful routes include:

- Callimachus: *Suda* κ 227 for titles; Stephens’ apparatus for the *Aetia* and *Hecale*; later quotations do not restore the whole texts.
- Epic Cycle: summaries attributed to Proclus, transmitted through later manuscripts and Photius; the identity and date of this Proclus should not be treated as certain. [Durham Living Poets, primary excerpt](https://livingpoets.dur.ac.uk/w/index.php/Proclus%2C_Chrestomathy).
- Democritus: Diogenes Laertius 9.46–49 reports a catalogue associated with Thrasyllus, much later than Democritus. The title list is not an Alexandrian shelf list. [Primary text](https://www.perseus.tufts.edu/hopper/text?doc=Perseus%3Atext%3A1999.01.0258%3Abook%3D9%3Achapter%3D7).
- Manetho and Berossus: Josephus’ *Against Apion* I preserves material in a later polemical context: modern sections 73–105 and 128–160 respectively. The surviving words are mediated testimony, not complete surviving histories. [Primary text, Whiston translation](https://penelope.uchicago.edu/josephus/apion-1.html).
- Medical writings: distinguish an attested treatise title from a modern collective label such as “anatomical writings.” Herophilus and Erasistratus require witness-specific records rather than a claim that the Library possessed all their writings.
- Reconstructed philosophical works: a modern assemblage of fragments is not an ancient complete text and must never be labelled as such.

No record should claim that a title disappeared in a named Alexandrian fire without evidence specific to that title. “No complete text survives” is generally the appropriate statement, with surviving fragments or quotations separately explained.

## Source-access and review record

The central ancient passages above and the relevant Bagnall, Erskine, Stephens and Leventhal discussions were inspected. The Met comparative peristyle passage was identified in the indexed PDF at printed p. 50. The McKenzie article’s publisher abstract and bibliographic record were inspected; this memo does not claim a full reading of its paywalled text. CEAlex’s relevant indexed text was available, while subsequent direct page requests intermittently returned a challenge or timeout; no extra architectural claim depends solely on inaccessible page detail.

This document records an agent’s source checks and reconstruction recommendations. It is not approval by a human academic expert. Build checks, rendered-world inspection and an independent historical pass on the finished reading records are separate acceptance steps.

The concrete review of the implemented world, catalogue metadata and selected downloaded passages is recorded in `qa/historical-review.md`. That report distinguishes a source check, a sampling check and a complete page-by-page comparison; those are not equivalent levels of verification.

## Historical acceptance checklist

- Every named zone links to the correct evidence card; room geometry never inherits false certainty from an attested activity.
- The first encounter explains that the layout is hypothetical and the holdings selection editorial.
- No specific shelf total is presented as the Library’s proven collection size.
- No lost text is represented by invented ancient prose.
- Every readable asset has a real source, an edition/translation identity and a statement of completeness.
- Work dates, later anthology content and pseudepigrapha are checked in the actual downloaded text, not only its catalogue metadata.
- Shelf association is kept separate from the work’s survival and its date.
- The complete catalogue does not imply all famous Alexandrians were contemporaries.
- The Serapeum is not conflated with the royal-quarter Library.
- Final labels, tour narration and overlays are checked for confident claims stronger than the sources.