Sundaland Research Program: Critias 115a–b & 118e: The Provisioning Complex of Staple and Companion

A Rice-Inclusive Base and Legumes under a Wet/Dry Calendar — A Semiotic–Philological Reading with Consilience-Based Reconstruction

A research by Dhani Irwanto, 13 September 2025

Abstract

This article advances a disciplined, text-first reconstruction of food provisioning in Plato’s Critias by treating 115a–b and 118e as a single, coupled textual object—the “Provisioning Complex.” The first element (115a–b) sets the consumption grammar in the present tense of the priest’s address to Solon: a property-defined base (“the dry sort that is our sustenance”) paired with a companion domain introduced by προσχρᾶσθαι (“use in addition”) and named in Greek as ὄσπρια (ospria; pulses/legumes). The second element (118e) supplies the operating calendar that makes the provisioning system intelligible: two harvests per year, in winter by the rains of heaven and in summer by water from canals. We tag all of 115a–b as B (audience gloss) and carry 118e alongside as a textual seasonality constraint from Order-1 onward. Under a context-clue hypothesis, the Greek wording intentionally supplies categories rather than foreign species names, preserving intelligibility while remaining non-diagnostic among cereals and species-neutral for the companion.

Methodologically, Orders-1/-2 (text and pragmatics) enforce lexical discipline (property vs. class), structural tests, and timeline hygiene, and license two legitimate parses to be carried in parallel: Dual-Basket (DB: a staple cereal umbrella with a distinct legume companion) and Single-Umbrella (SU: one dry-seed staple umbrella in which ospria functions as an illustrative example). Order-3 (reconstruction) proceeds by a consilience-based Puzzle Model that scores independent properties for Internal Coherence (IC, −2…+2) and External Consistency (EC, A/B/C). Seasonality (PP3) is treated as a text-driven hard constraint; base–complement pairing (PP17) registers the relationship-level interlock captured by προσχρᾶσθαι without forcing a species-level identity for the companion. The companion is deliberately kept species-agnostic—legume-class—in Order-3, because in provisioning practice it “follows” the base (co-located or logistically tethered nodes for procurement, processing, and storage).

Applied to the southern Kalimantan case (Sundaland), the pattern that emerges—perimeter/inland/transverse canals with plausible tidal modulation; a wet/dry agricultural calendar consistent with 118e; drying floors and granaries; standardized logistics (container metrology, canal-node wharfage); and regional plausibility for a rice-inclusive staple umbrella—generates strong puzzle interlock. Using the 17-piece catalogue (including PP17), the integrated results yield Σ(IC) DB = 31 (mean ≈ 1.82) and Σ(IC) SU = 29 (mean ≈ 1.71). These exceed the adoption threshold of no negative IC and mean IC ≥ +1 across the hydraulic–storage–cereal core and favor DB where provisioning contexts show cereal + legume co-signals; SU remains a viable fallback where cereal dominance persists and legume traces are sparse.

We conclude that the phrase “the dry sort that is our sustenance” is best read as a dry, storable staple umbrella into which rice plausibly fits by processing logic (cultivation → harvest → drying → storage) and provisioning infrastructure, while ὄσπρια marks the companion class. The analytical division of labor is deliberate: the text supplies grammar; the archaeology supplies species. Decisive next steps include dated co-films of cereal + legume residues on provisioning containers (starch morphometry, paleoproteomics, targeted biomarkers), stratified double-harvest signatures aligned to hydraulic phases, and granary microstratigraphy that records drying → storage sequences. Positive legume co-signals in provisioning contexts will raise PP17 and prefer DB; sustained cereal-only provisioning would strengthen SU. In either case, the Provisioning Complex remains the governing constraint that ties meal grammar to agronomic seasonality from the first line of analysis.

Keywords: Critias 115a–b; Critias 118e; ospria; dried staple; rice; double harvest; canals; anjir; Sundaland; semiotics; consilience; PP3; PP17; Dual-Basket; Single-Umbrella.

1. Problem Definition

1.1 Aim & Scope

This section defines the exact problem the article tackles and the textual constraints we will carry forward. At Critias 115a–b, the priest’s speech to Solon presents a triad: “the cultivated produce; the dry sort that is our sustenance; and the things we use in addition for the sake of the staple—we call all its kinds ‘ospria’.” We treat this triad as a deliberate context clue crafted for a Classical Greek audience.

Instead of naming species from the source region, the speaker uses a property label for the base food (“dry … sustenance”) and a Greek class-name for the companion (ospria, ‘pulses/legumes’). This rhetorical economy implies lexical unfamiliarity: the specific base cereal and the specific companion items in the remote region were not available—or not secure—in recognizable Greek crop-names. The audience receives intelligible categories, not species.

1.2 Textual Anchors

We anchor our reading in two clauses:

Critias 115a–b (consumption grammar): “… τὸν ἥμερον καρπόν, τὸν τε ξηρόν, ὃς ἡμῖν τῆς τροφῆς ἕνεκα ἐστιν, καὶ ὅσοις χάριν τοῦ σίτου προσχρώμεθα — καλοῦμεν δὲ αὐτοῦ τὰ μέρη σύμπαντα ὄσπρια;” “… the cultivated produce, and the dry [kind] which is for us for the sake of nourishment, and the things that we use in addition for the sake of the staple — we call all its kinds ‘ospria’ (pulses/legumes).” The present forms (ἐστιν, προσχρώμεθα, καλοῦμεν) mark an audience-directed gloss in the priest’s present (Solon’s time).
Critias 118e (operating calendar): “δύο γὰρ τοῦ ἐνιαυτοῦ θερισμοί — τὸν μὲν ἐν χειμῶνι ὑπὸ τοῦ οὐρανοῦ ὑετοῦ, τὸν δ᾽ ἐν θέρει ὑπὸ τῶν ἐκ τῶν ὀχετῶν ὑδάτων;” “twice in the year they gathered the fruits—in winter by the rains of heaven and in summer by water from canals.”

1.3 Key Lexemes

Several Greek words in these passages are decisive for interpretation:

A. Base (staple) vs. Companion (supplement)

σῖτος (sîtos) — staple grain/bread‐stuff domain. In our reading: base (provisioning staple), class-level, species-neutral.
ὄσπρια (óspria) — pulses/legumes (class name). In our reading: companion domain; species left open.
προσχρῶμαι (proscrômai) — “use in addition, employ besides.” Signals the supplemental function of ὄσπρια relative to σῖτος.
τροφή (trophḗ) — sustenance, nourishment. Anchors “staple” in the provisioning sense (quantity, storability, logistics).

B. Property label for the base (the “dry sort”)

ξηρός (xērós) — dry. Implies post-harvest drying → storage.
τὸν ἥμερον καρπόν (ton hḗmeron karpón) — “the cultivated produce/crop.” Grounds agricultural (not wild) production.

C. Naming & categorization (how the text frames classes)

καλέω (kaléō) — to call, name. (“We call… ὄσπρια.”) Signals Greek class-labeling rather than foreign species names.
μέρος/τὰ μέρη σύμπαντα (méros/tà mérē sýmpanta) — part(s), kinds (collectively). Marks a category set rather than a singular item.
χάριν (τοῦ σίτου) (chárin [tou sîtou]) — “for the sake of (the staple).” Purpose/ advantage construction tying companions to the base.
ὅσος/ὅσοις (hósos/hósois) — “those [things] which.” Introduces the companion set used alongside σῖτος.

D. Seasonality & hydraulics (118e anchors)

θερισμός/θερισμοί (therismós/therismoí) — harvest/harvests. Basis for double-harvest reading.
χειμών (cheimṓn) — winter (rainy half). Part of the wet season anchor.
θέρος (théros) — summer (dry half). Part of the dry season anchor.
ὑετός (hyetós) — rain. Source of winter-season water.
ὀχετός (ochetós) — channel/ditch/canal. Points to managed water in the dry season.
ὕδωρ/ὕδατα (hýdōr/hýdata) — water (pl.). The medium carried by canals in summer.
ἐνιαυτός (eniautós) — (agricultural) year. Frame for the “two harvests in the year.”

E. Deictics & perspective (audience gloss)

ἡμῖν (hēmîn) — “for us.” Dative of advantage; reinforces audience-directed present-tense framing (Timeline-B).
ἐστίν (estin) — “is”; plus present forms throughout (ἐστίν/προσχρώμεθα/ καλοῦμεν) — mark present-tense gloss to Solon.

1.4 Context-Clue Hypothesis and Unfamiliarity Claim

The deliberate use of a functional triad rather than a name implies a communicative act designed to overcome unfamiliarity. The Egyptian priest, aware that Solon would not recognize the base food and the companion by name, supplied its uses as context clues. These clues were pedagogical in nature: they bridged the cultural gap between an Egyptian knowledge of exotic products and a Greek listener unacquainted with them. For Plato’s audience, however, the effect was one of marvel and exoticism, reinforcing Atlantis as a land of abundance and strangeness. This unfamiliarity claim is central to understanding why the description survives not as a loanword but as a triadic inventory of functions.

1.5 Timeline Policy

We apply a conservative all-B tagging to 115a–b: the present-tense forms are read as a Solon-era gloss for a Greek audience. 118e is carried alongside as a textual seasonality constraint from Order-1 onward. We do not presuppose that these categories persisted unchanged over the ~9,000-year interval; any carry-back into the remote epoch is treated as an Order-3 hypothesis subject to independent corroboration (hydraulics, storage/drying, residues, botanical remains, chronology).

1.6 Research Questions (What Must Be Solved)

RQ1 — Identify the unfamiliar foods. Which specific base staple and companion best satisfy the Provisioning Complex (115a–b + 118e) when tested materially?
RQ2 — Choose the better parse. Under what evidence patterns should we prefer DB (paired cereal + legumes) or SU (single dry-seed umbrella), and what findings would flip that preference?
RQ3 — Specify decisive evidence. Which hydraulic features, storage/drying infrastructures, residue profiles, and micro/macro-botanical signatures would settle species identification and parse preference while respecting the Order-1 neutrality of the wording?

2. Methods

2.1 Overview & Design Goals

This section specifies the analytic toolkit used to interpret Critias 115a–b together with 118e and to carry that reading from text to reconstruction. We adopt a conservative, text-first discipline: 115a–b is treated as an all-B audience gloss (present-tense: ἐστιν, προσχρώμεθα, καλοῦμεν), and 118e contributes the seasonal operating schedule (“twice yearly”: winter by rains; summer by canal water). From Order-1 onward these two clauses are handled as a single Provisioning Complex that constrains all subsequent analysis. We carry two legitimate parses in parallel—Dual-Basket (DB) and Single-Umbrella (SU)—and evaluate reconstructions by consilience, not assertion.

2.2 Semiotics

Purpose. This subsection specifies how semiotic theory is operationalized in the article. We use (i) Saussure to discipline lexical claims at the clause level, (ii) Peirce to trace reference from words to world (and to classify evidence types), and (iii) Barthes to keep the three “orders” of meaning distinct (Order-1 text, Order-2 audience/pragmatics, Order-3 reconstruction).

2.2.1 Saussure’s Dyadic Model (signifier ↔ signified)

Unit of analysis. The signifier is the Greek wording; the signified is the concept invoked by that wording, not a species label.

Signifiers in 115a–b:

τὸν ἥμερον καρπόν (the cultivated produce)
τὸν τε ξηρόν … ὃς ἡμῖν τῆς τροφῆς ἕνεκα ἐστιν (“the dry sort that is our sustenance”)
σῖτος (the staple grain domain)
ὄσπρια (the legume class)

Guardrail. Do not collapse a signified (e.g., “dry, storable staple domain”) into a modern species name at Order-1. Species labels, if any, are earned later by material tests (Order-3).

Axes.

Syntagmatic (in-phrase fit): the triad forms a coordinated schema (base + in-addition companion), not a head-term with glosses.
Paradigmatic (contrast set): ὄσπρια (legumes) must not duplicate the staple domain (σῖτος); the contrast underwrites the “two baskets” reading (while still allowing a single-umbrella parse).
Commutation tests. Replace candidate terms with near-synonyms and check if provisioning sense breaks. If swapping out ὄσπρια with a non-legume “companion” destroys the provisioning logic, the legume class is semantically load-bearing.

Outcome. Saussure constrains our lexicon: “dry sort = property-defined staple domain; ὄσπρια = named companion class.” No species is asserted at the textual order.

2.2.2 Peirce’s Triadic Model (Representamen — Object — Interpretant)

Mapping.

Representamen (R): the clauses themselves (115a–b; 118e).
Object (O): the provisioning reality at Solon’s horizon (Timeline B), including cultivation, drying/storage, and the wet/dry calendar.
Interpretant (I): the understanding intended for a Classical Greek audience (context-clue function).

Immediate vs dynamic object.

Immediate O: the text-internal categories (staple domain; legume class).
Dynamic O: the empirically testable agrarian-hydraulic system (canals, double harvests, granaries).

Types of signs (for evidence triage).

Symbols (conventional): σῖτος, ὄσπρια.
Indices (causal/contiguous): 118e’s “winter by rain/summer by canal water” is indexical of a specific climatic–hydraulic regime; granary floors, drying yards, and canal sills are likewise indexical traces.
Icons (resemblance): rectilinear canal plans or maps that mimic hydrological geometry.

Operational rule. In adjudicating Order-3, indexical evidence (seasonality signatures, residue films, stratified storage) is privileged over purely symbolic analogies. Icons help, but indices decide.

Outcome. Peirce gives us a chain of reference and an evidence hierarchy: lexical symbols set categories; indices anchor them in the world.

2.2.3 Barthes’ Orders of Signification (denotation → connotation → myth)

Order-1 (denotation). Literal clause sense: a property-defined staple (“dry, storable, for sustenance”) plus a named companion class (ὄσπρια), all in the present of the priest’s address (B-tag).
Order-2 (connotation/audience pragmatics). The phrase operates as a context clue: where foreign species names would fail, the speaker supplies categories intelligible to Greeks. “Two harvests per year” (118e) is carried from Order-1 onward as the neutral seasonal frame.
Order-3 (myth/second-order system). Here “myth” is non-pejorative: a higher-order, culturally legible story we reconstruct by consilience (Puzzle Model). Order-3 may propose a rice-inclusive base and a legume companion, but only as a testable synthesis anchored by indices.

Outcome. Barthes supplies the three “floors” for our method: keep textual denotation, audience connotation, and reconstructed narrative strictly separated, with explicit hand-offs.

2.2.4 Putting It Together: An Operational Protocol

Name only what the text names (Saussure, Order-1). Treat “dry sort” as a property label for the staple domain; treat ὄσπρια as a class name for the companion.

Trace reference to the world (Peirce). Seek indices of the provisioning system specified by 118e: double-harvest signatures, canal-water phasing, drying/storage micro-stratigraphy, residue co-films.

Keep the orders distinct (Barthes).

Order-1: quote/parse; no species claims.
Order-2: justify the context-clue function and the all-B policy.
Order-3: run the Puzzle Model; prefer indexical corroboration; report IC/EC; keep the companion species-agnostic unless residues decide.

Parse discipline. Carry Dual-Basket (staple cereal + legume companion) and Single-Umbrella (one dry-seed staple umbrella with ospria illustrative) in parallel until Order-3 evidence picks a winner.

Fail-safe. If an Order-3 claim conflicts with Order-1 denotation (e.g., proposes a companion outside the legume class), the claim is rejected or re-scoped.

Net effect. Saussure protects the wording, Peirce binds words to world with an evidence hierarchy, and Barthes polices order-boundaries. Together they ensure that any species-level proposal (e.g., a rice-inclusive base) is a consequence of converging indices—not an assumption smuggled in at the textual order.

2.3 Text-Side Discipline & Translation Guardrails

σῖτος denotes the staple grain domain (‘bread-stuff’); ὄσπρια denotes the legume class (pulses). The phrase τὸν τε ξηρόν, ὃς ἡμῖν τῆς τροφῆς ἕνεκα ἐστιν functions as a property label for a dry, storable staple. The construction καὶ ὅσοις χάριν τοῦ σίτου προσχρώμεθα … καλοῦμεν … ὄσπρια assigns the companion role and names that domain in Greek. The result is a functional pairing—base (σῖτος) and complement (ὄσπρια)—rather than a mere list.

We read exactly what the words can bear. At 115a–b, the staple is signaled by a property label—τὸν τε ξηρόν, ὃς ἡμῖν τῆς τροφῆς ἕνεκα ἐστιν (‘the dry sort … for our sustenance’)—and the companion domain is defined by a Greek class-name—ὄσπρια (‘pulses/legumes’)—introduced with προσχρᾶσθαι (‘use in addition’). Guardrails: avoid translating σῖτος as “corn” (modern readers may take this as maize); retain ὄσπρια as “pulses/legumes”; render τὸν ξηρόν to preserve the dry-storage implication.

2.4 Two Parses Carried in Parallel

To avoid premature commitment, we carry two legitimate parses into later sections:

Dual-Basket (DB): “dry … sustenance” = staple cereal umbrella (σῖτος domain); ὄσπρια = distinct legume complement used “in addition.” This aligns closely with προσχρᾶσθαι and Classical meal grammar (base + with-food).
Single-Umbrella (SU): “dry … sustenance” = one dry-seed staple umbrella spanning cultivation → harvest → drying → storage; ospria = Greek-familiar example inside that umbrella. This maximizes enumerative economy and transmission robustness while flattening the cereal–pulse contrast.

2.5 Linguistics

We distinguish property labels (functional traits intelligible across contexts) from named classes (lexemes that anchor Greek taxonomy). In 115a–b, the staple is described by property (“dry … for sustenance”), while the supplement is named (ὄσπρια). This asymmetry is deliberate: it preserves audience intelligibility without forcing a species identification into the staple umbrella.

We treat the two clauses as one case. The complex constrains reconstruction along two axes: (i) consumption grammar—a dry, storable base with a companion domain used “in addition”; and (ii) agronomic seasonality—a wet/dry schedule yielding two annual harvests, winter by rainfall and summer by canal water. Any candidate that honors only one axis fails at the threshold.

2.6 Language Analysis

We apply four micro-tests:

Syntagmatic: the triad is balanced coordination—not head + gloss—supporting a base + companion schema.
Paradigmatic: the companion must not duplicate the staple umbrella; non-redundancy favors legumes.
Commutation: alternative companions must preserve provisioning sense; dry, storable ‘with-foods’ fit best.
Pragmatics: property + Greek class serves audience intelligibility.

A Cross-Clause Coherence test requires that any parse honoring the base–companion grammar also accommodate the 118e wet/dry operating schedule.

2.7 Philology & Transmission Controls

Purpose. This subsection sets the textual guardrails that keep our reading anchored to what the Greek can actually bear while minimizing anachronism from the Sais → Sonchis → Solon → Athenian oral circulation → Plato chain. Philology here is not about forcing species identifications out of a sparse lexicon; it is about respecting the form of Plato’s language—its categories, aspect, and coordinations—so that Order-3 reconstruction starts from a clean linguistic floor.

2.7.1 Base text, scope, and stance

Base text. We work from the standard OCT/Teubner text for Critias and adopt a conservative posture: no conjectural emendations and no special punctuation that would collapse coordinated items into apposition.

Unit of analysis. The “Provisioning Complex” comprises Critias 115a–b (consumption grammar) and 118e (seasonality). These are treated as a single textual case from Order-1 onward.

All-B tagging. The present-tense forms (e.g., ἐστιν, προσχρώμεθα, καλοῦμεν) function as an audience-facing gloss in Solon’s time. We therefore tag 115a–b as Timeline B (present frame), and we carry 118e as a textual seasonality constraint.

2.7.2 Lexical ranges and translation guardrails

σῖτος (sitos): the staple grain domain (“bread-stuff,” “cereal staple”), not a species label. Avoid rendering as “corn” to prevent maize confusions; “staple grain(s)” is acceptable where a gloss is needed.
ὄσπρια (ospria): the legume/pulse class (plural). This is a Greek category name, not a particular bean. We keep it species-agnostic by design.
τὸν ἥμερον καρπόν: “the cultivated produce/crop,” marking human-managed production as the field of discourse.
τὸν τε ξηρόν … ὃς ἡμῖν τῆς τροφῆς ἕνεκα ἐστιν: literally “and the dry kind, which is for us for sustenance.” The adjective ξηρός qualifies the state of use (dry, storable), not a botanical genus. The clause supplies a property-based label for the staple (dry, storage-grade).
καὶ ὅσοις χάριν τοῦ σίτου προσχρώμεθα: “and the things we use in addition for the sake of the staple [σῖτος].” The verb προσχρᾶσθαι carries the sense “to make additional use of,” encoding a functional complement to the staple.
καλοῦμεν δὲ αὐτοῦ τὰ μέρη σύμπαντα ὄσπρια: “and we call all its kinds ospria.” μέρη (“parts/kinds”) plus σύμπαντα (“the whole set”) yields a classificatory sweep, not a list of species.

Rendering policy. Keep σῖτος ~ “staple grain(s)”; ὄσπρια ~ “pulses/legumes”; preserve “dry sort” to carry storage semantics. Do not smuggle species names (e.g., “rice,” “wheat”) into Order-1 translations.

2.7.3 Syntax and information structure: the triad as coordination, not apposition

The sequence at 115a–b (“cultivated produce” · “the dry sort that is our sustenance” · “the things used in addition, called ospria”) is a coordinated triad, not a head-term with appositive clarifications. Consequences:

The second member (“the dry sort … sustenance”) is not a synonym of the first; it narrows by property to the provisioning base.
The third member introduces a second domain by function (supplement to σῖτος) and names it by Greek class (ὄσπρια).
This coordination structurally licenses two parses that we carry forward in parallel:
Dual-Basket (DB): property-defined staple (dry, storable cereal) plus a distinct legume companion (ὄσπρια) used in addition.
Single-Umbrella (SU): one umbrella of dry-seed staple; ὄσπρια functions as the illustrative Greek-named subset within that umbrella.

The Greek does not decide between DB and SU; it permits both. Selection becomes an empirical matter in Order-3.

2.7.4 Aspect, deixis, and timeline hygiene

Present-tense deixis (ἐστιν/προσχρώμεθα/καλοῦμεν) situates the gloss in the audience’s present, justifying our all-B treatment of 115a–b.

Historical projection back to the remote epoch is a different operation (Order-3) and cannot be inferred from the grammar. We prevent “timeline bleed” by quarantining such projections for material tests (hydraulics, residues, macro/micro-botanicals, chronology).

2.7.5 Transmission and “legendization” risks (what philology will and won’t do)

Economy of enumeration. Across long oral transmission, categories endure better than species lists. The pairing property-term (staple) + Greek class (companion) is exactly the sort of audience-durable packaging one expects to survive.

What we will not infer. We will not retroject Greek taxonomies onto the source culture; we will not treat ὄσπρια as a single species; we will not read ξηρός as a crop identity.

What we will infer. The text encodes a consumption grammar (staple + companion) and a calendar (118e) that any reconstruction must honor.

2.7.6 Philology-to-method handoff

Order-1 deliverables from philology: (i) a property-defined staple (dry, storage-grade) in the σῖτος domain; (ii) a named companion class (ὄσπρια) used in addition; (iii) 118e seasonality as a hard textual constraint; (iv) all-B tagging for 115a–b.

Order-2 consequence: treat 115a–b as a context clue—categories supplied where foreign species names are absent/unsafe.

Order-3 mandate: test DB vs SU with independent evidence (hydraulic regime, storage infrastructure, double-harvest signatures, residues). Philology permits both parses; the archaeology adjudicates.

2.7.7 Mini-glossary (working senses)

σῖτος: staple grain domain; provisioning base; not a species label.
ὄσπρια: pulses/legumes as a class; species-neutral.
ξηρός: “dry” as state of use (storage-ready), not botanical identity.
προσχρᾶσθαι: “to use in addition,” encoding a functional supplement to the staple.
μέρη σύμπαντα: “all the kinds/parts,” i.e., class coverage, not an item list.

Bottom line: Philology secures a disciplined reading—property-based staple + class-named companion under a wet/dry calendar—and explicitly limits what the Greek can decide. That discipline is what makes the subsequent consilience tests meaningful.

2.8 Context-Clue Hypothesis and Unfamiliarity Claim

We adopt the context-clue hypothesis: the speaker substitutes property terms and familiar Greek classes because the specific foods in the source region—the dominant base staple and its companion—were unfamiliar or lexically unavailable in precise Greek names. The speaker used (i) a property term for the staple and (ii) a Greek class-name for the companion so the audience would grasp the provisioning logic. Thus ospria operates as a category pointer here, not a species label for the remote region. This principle precludes extracting a species identity from wording alone and motivates carrying DB and SU in parallel. This stance is conservative and testable: it resists over-reading the text while predicting that species-level identities will be resolved via material evidence, not wording alone.

2.9 Timeline Discipline

To avoid assuming continuity across ~9,000 years, we tag all of 115a–b as Timeline-B (present-tense audience address) and carry 118e as a textual seasonality constraint from Order-1 forward. Any projection of these categories into the remote epoch is treated as an Order-3 hypothesis to be tested by independent evidence (hydraulics, storage, residues, botanical remains, chronology).

2.10 The Puzzle Model — Definition and Use

Reconstruction proceeds via a Puzzle Model. Each puzzle piece is an independently testable property of the target system (environmental, hydraulic, logistical, botanical, textual). We judge placements by interlock—how pieces fit without contradiction—and track two metrics: Internal Coherence (IC) scored from −2 (contradiction) to +2 (tight interlock), and External Consistency (EC) rated A (direct/dated), B (indirect), C (plausible). Adoption into synthesis requires no negative IC and mean IC ≥ +1 across the hydraulic–storage–cereal core. We report Σ(IC) and mean IC separately for DB and SU.

2.11 Evidence Classes for Order-3

We use six evidence classes:

Hydraulics & landscape (perimeter/inland/transverse canals; inflows/outlets; bunds/levees; drying floors; flood management);
Storage & logistics (granaries; ventilation; raised floors; standardized containers; nodes/wharfage; tallies/seals);
Plant remains (micro-botanical—phytoliths, starch, weed spectra—and macro-botanical—charred grains/chaff; chaff temper);
Residues on artefacts (starch/protein/lipid films on vessels/tools);
Chronology (AMS on plant remains; OSL on canal fills/bunds/drying surfaces);
Comparative texts (supporting context only, not primary for identification).

2.12 Puzzle Piece Catalogue (17 Items)

Tropical belt placement.
Wet-cultivation capacity.
Twice-yearly harvest seasonality (118e).
Coconut/palm co-occurrence.
Alluvial/deltaic lowlands (broad plain).
Monsoon reliability.
Tidal modulation.
Waterways & irrigation (perimeter/inland/transverse canals).
Hydraulic competence (sluices/gates; maintained levels).
Storage infrastructure (granaries/drying floors).
Logistics standardization (container metrology; wharfage; tallies/seals).
Rice-origins proximity (biogeographic plausibility within SEA/South/East Asia).
Indian Ocean connectivity (maritime access and trade context).
Faunal correlates (e.g., elephants) aligned with provisioning landscapes.
Aromatics/commodities (archaeochemical/textual correlates).
Staple provisioning dominance (capacity/standardization governs logistics).
Base–complement pairing (σῖτος + ὄσπρια): operationalized by co-occurrence or dominance patterns in residues/storage—higher IC for DB where co-signals are robust; neutral to low for SU unless cereals dominate.

2.13 Sampling & Laboratory Protocols (Priority Contexts)

Prioritize: (i) stratified granary floors and drying yards; (ii) canal margins and bund crowns for weed/diatom assemblages; (iii) interior surfaces of cooking vessels and tools; (iv) use-wear zones on harvesting/processing implements; (v) foundation trenches of sluices/gates. Every sample receives full provenience, context description, and a pre-registered hypothesis (e.g., DB co-occurrence of cereal + legume signals vs SU cereal dominance).

2.14 Orders of Signification — Workflow & Gates

Order-1 (text-only): establish permissible readings; bar crop identities and geography. Order-2 (audience pragmatics): apply context-clue principle, structural tests, translation guardrails; still no geography. Order-3 (reconstruction): introduce geography and material culture; test cereals inside the staple umbrella and legume signals for DB; score via IC/EC.

2.15 Parse Gate (DB vs SU)

Prefer the parse that improves IC without contradiction. Choose DB when legume + cereal indicators co-occur in provisioning contexts (granaries, storage vessels, residue co-films) and when logistical organization indicates deliberate pairing (raises PP17). Choose SU when cereal signals dominate and legume traces are sparse/secondary. If undecided, carry both with parallel justifications.

2.16 Scoring & Acceptance Rule (IC/EC)

Score each piece for IC (−2…+2) and EC (A/B/C). Adopt a reconstruction into synthesis only when there is no negative IC and mean IC ≥ +1 across the hydraulic–storage–cereal core (PP5, PP8–PP11, PP16, with PP3 seasonality satisfied). Report Σ(IC) and mean IC separately for DB and SU.

2.17 Risks, Confounds, and Falsifiers

Equifinality: natural levees and paleo-channels can mimic canals—mitigate by converging geometry + gates + fills + logistics nodes.
Temporal palimpsests: later re-cuts overlay earlier networks—require stratified dating and interface sampling.
Scale inflation: normalize reported dimensions by functional comparanda—privilege pattern coherence over literal numbers.
Confounds: elite hoards or specialty depots misread as staple stores.
Falsifiers: persistent absence of cereal/storage signatures in primary contexts; residue profiles incompatible with staple provisioning; canal chronologies inconsistent with 118e’s wet/dry framework.

2.18 Controlled Terms (Quick Reference)

Staple (provisioning): bulk, dry-storable cereal base (σῖτος domain).
Ospria: pulses/legumes, the named companion domain.
Provisioning Complex: the coupled reading of 115a–b (consumption grammar) and 118e (seasonality) carried from Order-1 onward.
DB/SU: Dual-Basket vs Single-Umbrella parses.
All-B policy: tag all of 115a–b as Timeline-B.

3. Orders 1–3 Workflow

3.1 Overview

This section operationalizes the study from text to reconstruction, enforcing the all-B policy for Critias 115a–b and carrying 118e as a textual seasonality constraint from Order-1 onward. Together, 115a–b + 118e constitute the Provisioning Complex: a property-defined staple base with a named companion domain operating on a wet/dry, twice-yearly schedule. All steps below carry Dual-Basket (DB) and Single-Umbrella (SU) in parallel and evaluate candidates by consilience (IC/EC) rather than assertion.

3.2 Inputs & Outputs

Inputs. §1 Problem Definition (context-clue reading; all-B tagging; Provisioning Complex; DB/SU); §2 Methods (guardrails; structural tests; evidence classes; Puzzle Model; 17-piece catalogue; IC/EC rubric).

Outputs. (i) Order-1 reading notes and Provisioning Complex statement; (ii) Order-2 audience/pragmatic memo; (iii) Order-3 sampling plan; (iv) a populated Evidence Log; (v) DB/SU IC/EC tables; (vi) a parse decision or explicit parallel carriage; (vii) handoff to §4 Integrated Results.

3.3 Gate 1 — Pre-registration & Normalization

Before any fieldwork or labwork:

register hypotheses (DB/SU), acceptance rule, and scoring thresholds (no negative IC; mean IC ≥ +1 across the hydraulic–storage–cereal core; PP3 seasonality must be satisfied);
lock controlled terms (§2.18);
state the Provisioning Complex as the textual premise;
confirm all-B tagging for §1.5 references;
publish the Evidence Log template.

3.4 Order-1 Workflow — Text-Only (No Geography, No Species)

Step O1-1. Literal notes. Record the Greek and targeted glosses for 115a–b and 118e; highlight the property label for the base, the Greek class-name for the companion, and the present-tense framing.
Step O1-2. Provisioning Complex statement. In one paragraph, state that 115a–b (base–companion grammar) and 118e (wet/dry schedule) are carried together from this point forward as the textual premise.
Step O1-3. Parse statements. Write the DB and SU statements you will carry; include the cross-clause coherence requirement (a valid parse honors both the meal grammar and 118e’s schedule).

Deliverable O1: a one-page Order-1 sheet (placed at the head of the Evidence Log and cited in §4.2).

3.5 Order-2 Workflow — Audience/Pragmatics (Still Text-Side)

Step O2-1. Context-clue memo. Explain why property + Greek class preserves intelligibility when specific names were unfamiliar or lexically unavailable; reiterate all-B tagging for 115a–b.
Step O2-2. Structural tests. Apply syntagmatic balance, paradigmatic non-redundancy, commutation, and pragmatics. Document how προσχρᾶσθαι natively supports DB while SU remains legitimate by enumerative economy.

Deliverable O2: a two-paragraph memo to be cited in §4.3 and appended to the Evidence Log.

3.6 Order-3 Workflow — Reconstruction (Geography & Material Culture Enter)

At Order-3 we test the Provisioning Complex materially. Run the pipelines below in parallel; each datum maps to PP#, IC (DB/SU), and EC in the Evidence Log.

Hydraulics & Landscape Pipeline. Map perimeter/inland/transverse canals; mountain inflows; sea outlets; bunds/levees; drying floors; sluices/gates and level controls. Cross-check spacing/geometry against functional comparanda and tidal/monsoon regimes. (Anchors: PP5, PP7–PP9.)
Storage & Logistics Pipeline. Locate granaries (ventilation, raised floors), drying platforms, standardized containers (metrology), canal-node wharfage, tallies/seals; estimate provisioning capacity and dominance. (Anchors: PP10–PP11, PP16.)
Plant Remains & Residues Pipeline. Recover micro-botanical (phytoliths—including rice husk/bulliform—starch, weed spectra) and macro-botanical (charred grains/chaff; chaff temper) signals; sample interiors of cooking vessels/tools for starch/protein/lipid films and co-films. (Anchors: PP2–PP3, PP12, PP17.)
Plant Remains & Residues Pipeline. Apply AMS to plant remains; use OSL (or equivalent) on canal fills, bunds, drying surfaces, and gate foundations. Time-stamp hydraulic operation relative to 118e’s wet/dry cycle. (Anchors: PP3, PP8–PP9.)

3.7 Evidence Log — Template & Tagging

Use a standardized table so each datum carries its interpretive status. Timeline tag = B for any assumption derived from 115a–b. Chronology determines how (or whether) categories can be projected toward the remote epoch. Suggested fields: (i) ID; (ii) Provenience/ Context; (iii) Clause Anchor (115a–b/118e); (iv) Puzzle Piece (PP#); (v) Indicator(s); (vi) Parse Target (DB/SU); (vii) IC (DB); (viii) IC (SU); (ix) EC (A/B/C); (x) Dating Method/ID; (xi) Analyst Notes; (xii) Ref/Figure.

3.8 Parse Gate & Decision Rules (DB vs SU)

Rule 1 — Evidence-led. Prefer DB when legume + cereal indicators co-occur in provisioning contexts and improve interlock (raises PP17); prefer SU when cereal signals dominate and legume traces are sparse or secondary.
Rule 2 — No contradiction. Reject parses that introduce negative IC against the hydraulic–storage–cereal core or violate PP3 seasonality.
Rule 3 — Transparency. If undecided, carry both parses with parallel justifications and report Σ(IC) and mean IC for each.

3.9 Scoring & Thresholds (IC/EC)

Score each piece for IC (−2…+2) and EC (A/B/C). Adopt into synthesis only when there is no negative IC and the mean IC ≥ +1 across the hydraulic–storage–cereal core (PP5, PP8–PP11, PP16) with PP3 satisfied. Report totals for DB and SU separately. Flag PP3 as a text-driven hard constraint: contradiction implies IC = −2 for both parses.

4. Integrated Analyses & Results (Orders 1–3)

4.1 Overview & Conventions

This section consolidates the outputs of Order-1 (text), Order-2 (audience/pragmatics), and Order-3 (reconstruction). We proceed from the Provisioning Complex—Critias 115a–b (base–companion consumption grammar) coupled with 118e (wet/dry, twice-yearly operating calendar)—adopted from Order-1 onward. The all-B policy holds for 115a–b (present-tense audience gloss). Both Dual-Basket (DB) and Single-Umbrella (SU) parses are carried in parallel. Scoring follows §2.16: Internal Coherence (IC) on −2…+2; External Consistency (EC) at A/B/C; adoption requires no negative IC and mean IC ≥ +1 across the hydraulic–storage–cereal core, with PP3 (seasonality) treated as a text-driven hard constraint.

4.2 Order-1 Outputs (Carried Forward)

4.2.1 Greek & Literal (Targeted Clauses)

115a–b: “… τὸν ἥμερον καρπόν, τὸν τε ξηρόν, ὃς ἡμῖν τῆς τροφῆς ἕνεκα ἐστιν, καὶ ὅσοις χάριν τοῦ σίτου προσχρώμεθα — καλοῦμεν δὲ αὐτοῦ τὰ μέρη σύμπαντα ὄσπρια.” Literal: “… the cultivated produce, and the dry [kind] which is for us for the sake of nourishment, and the things that we use in addition for the sake of the staple — we call all its kinds ‘ospria’ (pulses/legumes).”
118e: δύο γὰρ τοῦ ἐνιαυτοῦ θερισμοί — τὸν μὲν ἐν χειμῶνι ὑπὸ τοῦ οὐρανοῦ ὑετοῦ, τὸν δ᾽ ἐν θέρει ὑπὸ τῶν ἐκ τῶν ὀχετῶν ὑδάτων. Literal: “twice in the year they gathered the fruits—in winter by the rains of heaven and in summer by water from canals.”

4.2.2 Final Order-1 Reading

Staple base (property label): “dry … sustenance” = dry, storable provisioning base (σῖτος domain).
Companion (named class): ὄσπρια (‘pulses/legumes’) introduced by προσχρᾶσθαι (“use in addition”).
Seasonality: a wet/dry, twice-yearly operating schedule (118e) belongs to the same textual case.
Tagging: all of 115a–b is B (audience gloss). Species/geography remain outside Order-1.

4.3 Order-2 Outputs (Audience/Pragmatics)

4.3.1 Context-Clue Outcome

The text functions as a context clue for a Classical audience: a property term for the staple plus a Greek class-name for the companion conveys the provisioning logic without presuming that specific crop names from the source region were present in the Greek lexicon.

4.3.2 Structural Tests — Verdict

Test	Question	Verdict	Implication
Syntagmatic	Balanced coordination vs head + gloss?	Balanced triad; not mere gloss	Supports base + companion schema
Paradigmatic	Does the companion duplicate the staple umbrella?	No; pulses are distinct	DB favored by non-redundancy
Commutation	Do alternate companions preserve provisioning sense?	Dry, stored ‘with-foods’ fit	Legumes are natural fit
Pragmatics	Does wording serve audience clarity?	Property + Greek class	Context-clue confirmed
Cross-Clause	Does the parse honor 118e seasonality?	Required from Order-1	Hard constraint on candidates

4.3.3 Parse Statements (Carried Forward)

DB: staple cereal umbrella (property-defined) + distinct legume complement (ospria) used “in addition.”
SU: one dry-seed staple umbrella; ospria is a Greek-familiar example inside the umbrella.

4.4 Bridge — Hypotheses & Sampling Plan (Order-3)

H-DB1 (Co-residues): robust cereal + legume co-films in provisioning contexts (granaries, storage vessels) raise PP17 and favor DB.
H-SU1 (Cereal dominance): cereal-dominant residues with sparse legume traces favor SU.
H-HYD: canal geometries and gate features operate within the wet/dry framework (PP3, PP8–PP9).
H-STOR: storage/drying infrastructure and logistics standardization show staple provisioning dominance (PP10–PP11/PP16).

Sampling priorities: stratified granary floors and drying yards; canal margins/bund crowns; interior surfaces of cooking vessels/tools; foundation trenches of sluices/gates. Each sample is logged with PP#, IC (DB/SU), EC, and dating.

4.5 Clause-to-Feature Mapping (southern Kalimantan)

Plato clause	Key feature	Southern Kalimantan analogue	Relevance to staple reconstruction
115a: τὸν ἥμερον καρπόν	Cultivated produce	Long-standing managed lowlands; intensive agriculture	Supports intensive cultivation context
115a: τὸν τε ξηρόν … τῆς τροφῆς ἕνεκα ἐστιν	Staple (dry, storable base)	Granaries; drying platforms; rice export tradition	Matches provisioning & storability requirements
115a–b: ὅσοις … προσχρώμεθα — … ὄσπρια	Companion domain (pulses)	SEA legume complement, lauk (unspecified; e.g., Vigna spp.; soybean later historically)	Operationalizes base–complement pairing (PP17); follows base placement in the Puzzle Model
118e: δύο θερισμοί	Two harvests per year	Wet/dry calendar; tidal irrigation (anjir–handil-saka)	Consistent with double-cropping and managed water
118c–e: canals	Waterways & irrigation	Barito–Kapuas–Kahayan; transverse–inland–irrigation (anjir–handil–saka)	Hydraulic capacity for wet cultivation & transport (PP8–PP9)
118e: winter rains/ summer canals	Seasonal water sources	Monsoon rains + regulated canal flows	Implements 118e schedule (PP3)
—	Staple provisioning dominance	Granary/export role; standardized sacks/containers	Strengthens PP16 (logistics/capacity)

4.6 Puzzle Piece Scores (Catalogue with PP3 dual role; PP17 pre-scored)

Clarification: The IC/EC scores in this table evaluate the Order-3 base staple reconstruction (rice-inclusive umbrella). The companion identity is intentionally left unspecified at Order-3 and handled generically as legumes (ὄσπρια) in line with Order-1. Its interlock with the base is captured by PP17 (base–complement pairing); no separate companion species scoring is required.

PP#	Name	IC (DB)	IC (SU)	EC (A/B/C)	Notes / Anchor
PP1	Within tropical belt (≤ ~23.5°)	+2	+2	A
PP2	Wet cultivation capacity (paddy suitability)	+2	+2	B
PP3	Twice-a-year harvest (Critias 118e)	+2	+2	B	Seasonality (118e); text-driven hard constraint—if contradicted, set IC = −2 for both parses.
PP4	Co-occurring palms incl. coconut	+2	+2	A
PP5	Advanced waterways & irrigation on the plain	+2	+2	B
PP6	Alluvial/deltaic lowlands	+2	+2	B
PP7	Cereal storage infrastructure	+1	+1	C
PP8	Pulse complementarity (ospria)	+1	0	B	Textual complement (ospria) present at Order-1; not species-specific.
PP9	Elephant habitat in fauna	+1	+1	B
PP10	Aromatics/fragrant products	+2	+2	A
PP11	Navigation & hydraulic engineering competence	+2	+2	B
PP12	Monsoon rainfall reliability	+2	+2	A
PP13	Tidal modulation potential	+2	+2	B
PP14	Rice origins proximity (E/SE/S Asia)	+2	+2	A
PP15	Indian Ocean trade connectivity	+2	+2	A
PP16	Staple provisioning dominance (bulk-calorie economy)	+2	+2	C
PP17	Base–complement pairing (σῖτος + ὄσπρια)	+2	+1	A	Textual complementarity at 115a–b (προσχρᾶσθαι → ὄσπρια); companion follows base placement.

With the 17-piece catalogue and the scored table, the integrated totals are: Σ(IC) DB = 31 (mean ≈ 1.82), Σ(IC) SU = 29 (mean ≈ 1.71).

4.7 Narrative Assessment

The combined Order-1/-2 outputs articulate a base–supplement economy voiced for a Classical audience: a dry, storable staple base paired with a named legume domain, operating under a wet/dry, twice-yearly schedule. In Order-3, the southern Kalimantan application shows pattern-level concordance across hydraulics (transverse/inland/ irrigation canals with plausible tidal modulation), seasonality (118e), provisioning infrastructure (granaries, drying yards, standardized containers), and regional plausibility for a rice-inclusive staple umbrella. These strands improve puzzle interlock; decisive adjudication still rests on dated micro-/macro-botanical signatures and residue profiles in primary provisioning contexts.

4.8 Parse Decision & Sensitivity

Stance: carry DB and SU in parallel.
Preference rules: favor DB where cereal + legume co-residues appear in provisioning contexts and logistical organization indicates deliberate pairing (raises PP17); favor SU where cereal signals dominate and legume traces are sparse or secondary.
Flip conditions: robust, dated co-residues in storage/transport contexts strengthen DB; sustained cereal-only provisioning signatures strengthen SU.

4.9 Risks & Falsifiers (Results-Side)

Equifinality (canal look-alikes), temporal palimpsests (over-cut channels), scale inflation in reportage, category drift between parses. Falsifiers: absence of cereal/storage signatures in primary contexts; residue profiles incompatible with staple provisioning; canal chronologies inconsistent with 118e.

5. Discussion

5.1 Purpose & Scope

This section interprets the integrated results (§4) against the methodological railings (§2) and the problem definition (§1). We keep the Provisioning Complex (Critias 115a–b + 118e) in view: a property-defined staple base paired with a named legume domain and operating on a wet/dry, twice-yearly schedule carried from Order-1 onward. The goals are to explain why the Greek wording presents a legume class rather than a named cereal, to formalize the dependency between base and companion in provisioning terms, to weigh the DB/SU parses in light of the evidence, and to set out what would strengthen or falsify the reconstruction.

5.2 The Textual Premise Restated (Order-1 carried forward)

The staple is expressed as a property label—“the dry sort that is our sustenance”—while the companion is expressed as a Greek class-name, ὄσπρια (pulses), introduced with προσχρᾶσθαι (“use in addition”). This asymmetry is a deliberate context clue to keep the provisioning logic intelligible for a Classical audience when specific pre-Solon crop names are unavailable or unsafe to assert. Clause 118e supplies the operating calendar—two harvests per year, in winter by rains and in summer by canal water—which functions as a text-driven hard constraint on reconstruction (§4.6, PP3).

5.3 Why “Pulses” Appears (and Not “Rice”) in the Greek Wording

Under the context-clue hypothesis (§1.4; §2.8), the priest chooses audience-familiar categories rather than species names from the remote region. A property term for the staple (“dry … sustenance”) communicates function without committing to a specific cereal; a Greek class-name for the companion (“we call them ospria”) communicates the meal grammar without naming a foreign legume species. The combination is precise enough to carry provisioning sense and robust enough to survive transmission (§5.8), yet neutral as to species.

5.4 Staple in the Provisioning Sense (Quantity, Storability, Logistics)

Throughout this study, staple is used in the provisioning sense (§1.2; §2.7): the bulk, dry-storable base that governs quantity, storage, and logistics. This definition harmonizes the Order-1 property label with Order-3 infrastructure: drying floors, granaries, standardized containers, and canal nodes (§4.5; §4.6 PP10–PP11, PP16). Rice qualifies on processing logic—cultivation → harvest → drying → storage—but the wording at Order-1 remains non-diagnostic among cereals.

5.5 Follow-On Dependency: Why the Companion “Tracks” the Base

In provisioning systems, the companion is functionally dependent on where and how the base is produced and stored. The text marks this with προσχρᾶσθαι (“use in addition”), and the Puzzle Model captures it as PP17 (base–complement pairing). Practically, once the base is placed (Order-3), the companion follows: its procurement, processing, and storage nodes are co-located with, or logistically tethered to, the base’s provisioning network. This is why we do not need to fix a species-level identity for the companion in Order-3 scoring; the interlock is measured at the relationship level (DB) rather than at the species level.

5.6 Species-Level Agnosticism for the Companion (and What Would Decide It)

Order-1 gives a class (ὄσπρια), not a species. Order-3 therefore keeps the companion deliberately agnostic at the species level while retaining a strong interlock signal via PP17. The decision is evidentiary: a species claim requires co-occurring indicators in primary provisioning contexts—e.g., cereal phytoliths/chaff together with legume residues (starch morphometry; paleoproteomic markers) in granaries or storage vessels, or macro-remains (cotyledon/seed coat) with direct AMS dates. Absent such signals, the class-level reading remains both faithful to the text and sufficient for provisioning analysis.

5.7 DB vs SU: Consequences of the Two Parses

DB (Dual-Basket). Interprets the triad as a base + complement pairing. It leverages the native force of προσχρᾶσθαι and aligns with Classical meal grammar. In the results (§4.6), PP17 carries IC(DB)=+2, reflecting strong textual support; the consilience totals favor DB when provisioning contexts show co-residues or paired logistics (§4.9).

SU (Single-Umbrella). Interprets “dry … sustenance” as a single dry-seed staple umbrella and treats ospria as an illustrative example rather than a distinct basket. SU maximizes enumerative economy and transmission robustness but flattens the cereal–legume contrast. It is preferred only where cereal signals dominate and legume traces are persistently sparse (§4.9).

5.8 Timeline Discipline & Legendization in Transmission

We maintain the all-B tagging for 115a–b and carry 118e as a textual seasonality constraint from Order-1. The long transmission chain—Sais temple registers → Sonchis’ narration → Solon → oral circulation in Athens → Plato—invites localization, metaphorization, and personalization (§1.4 context-clue; the legendization account). The context-clue strategy (property + Greek class) is precisely what we would expect to survive such a chain: intelligible to the audience, resilient under retelling, and non-committal at species resolution.

5.9 Sensitivities, Limits, and Falsifiers

Sensitivities. Wetland geomorphology can mimic canals; later re-cuts overprint earlier networks; storage features can be misread as elite hoards.

Limits. Text cannot identify species; only material signatures can.

Falsifiers. (i) Absence of cereal/storage signals in primary contexts; (ii) residue profiles incompatible with staple provisioning; (iii) canal chronologies inconsistent with the 118e schedule; (iv) robust cereal-only provisioning with no legume co-signals in relevant contexts (would weaken PP17 and favor SU).

5.10 Implications for the Sundaland Application (Southern Kalimantan)

The Sundaland case (§4.5) coheres at the pattern level: canals (transverse/inland/ irrigation) with tidal modulation, wet/dry seasonality, drying floors and granaries, and a logistics system suitable for a dry-storable cereal base with legume companions. In this framework, the companion is expected to track the base’s nodes and flows (PP17). Species-level claims for the companion remain open until residues or macro-remains dictate otherwise; the reconstruction does not require that specificity to evaluate provisioning fitness.

Anjir System: a unique ancestral heritage tidal irrigation network in southern Kalimantan. Primary canals called “anjir” or “antasan” were constructed traversing two tidal rivers, also used as navigation purposes. Inland canals were built to irrigate and drain the fields from and to the anjir: secondary canals called “handil” or “tatah” and tertiary canals called “saka”. During low tides, the canals drain toxic water while during high tides fresh water enters the canals and conveyed to the fields. The system yields two rice crops in a year.

5.11 Forward Tests & Predictions

Predictions. (1) Double-harvest signatures in stratified contexts (wet/dry) aligned to gate/sluice phases; (2) co-films of cereal + legume residues on storage or transport containers; (3) granary microstratigraphy showing drying → storage sequences; (4) container metrology clustering around standardized provisioning units; (5) dateable hydraulic episodes that synchronize with the agricultural calendar implied by 118e.

Deciders. Positive detection of legume co-residues in provisioning contexts will raise PP17, strengthen DB, and narrow the companion’s species set; stable cereal-only provisioning would instead favor SU.

5.12 Closing Synthesis

The Greek wording gives us a functional grammar—a dry, storable staple base with a named legume domain—operating on a wet/dry year. The Puzzle Model turns that grammar into a testable reconstruction: once the base is placed, the companion follows operationally, whether or not we can yet name its species. In this light, rice remains a strong, testable base candidate; the companion stays intentionally species-agnostic until residues decide. This division of labor—text for grammar, material for species—keeps the analysis disciplined and cumulative.

6. Conclusion

6.1 What the Text Can Bear

Plato’s wording at Critias 115a–b gives a functional grammar—a dry, storable staple base (“the dry sort that is our sustenance”) paired with a named companion class (ὄσπρια, pulses), introduced with προσχρᾶσθαι (“use in addition”). Coupled with 118e (“twice yearly”: winter by rains, summer by canal water), we treat these clauses as a single Provisioning Complex carried from Order-1 onward. We adopt the all-B policy for 115a–b (audience-directed gloss). The text is non-diagnostic among cereals and species-neutral for the companion—by design, as a context clue for a Classical Greek audience when specific crop names from the source region were unavailable or unsafe to assert.

6.2 Methodological Outcome

A disciplined semiotic–philological reading (Orders-1/-2) sets the gates; reconstruction (Order-3) proceeds by a Puzzle Model that scores independent, interlocking properties (IC −2…+2, EC A/B/C). We carry Dual-Basket (DB) and Single-Umbrella (SU) parses in parallel; PP3 (seasonality) is a text-driven hard constraint, and PP17 (base–complement pairing) measures the relationship-level interlock without forcing species identity for the companion.

6.3 Reconstruction Verdict (with Scores)

Applying the model to the southern Kalimantan case yields strong pattern-level concordance across hydraulics, seasonality, storage/logistics, and regional cereal plausibility (rice-inclusive umbrella). With the 17-piece catalogue and the scored table, the integrated totals are: Σ(IC) DB = 31 (mean ≈ 1.82), Σ(IC) SU = 29 (mean ≈ 1.71). These meet the adoption threshold (no negative IC and mean IC ≥ +1 across the hydraulic–storage–cereal core) and favor DB when provisioning contexts produce cereal + legume co-signals; SU remains viable where cereal dominance persists.

6.4 Companion Policy (Follow-On Dependency, Species-Agnostic)

Operationally, the companion tracks the base: procurement, processing, and storage are co-located or tethered to base provisioning nodes. This “follow-on” dependency is encoded textually (προσχρᾶσθαι) and structurally (PP17). Accordingly, Order-3 keeps the companion species-agnostic (legume-class) until residues/macros decide. The reconstruction does not require a species name to evaluate provisioning fitness; it requires co-residue and logistics signatures that demonstrate pairing.

6.5 Implications for the Sundaland Application

The Sundaland (southern Kalimantan) application exhibits the right shape: transverse/ inland/irrigation canals with plausible tidal modulation, a wet/dry agricultural calendar consistent with 118e, evidence of drying/storage infrastructure and standardized logistics, and regional plausibility for a rice-inclusive staple umbrella. Within this frame, a legume companion is expected to follow the base through the provisioning network. Species-level identification for the companion is deliberately left open pending dated residues or diagnostic macro-remains.

6.6 Limits, Risks, and Decisive Tests

Limits. The text cannot assign species; transmission invites localization.

Risks. Equifinality in canal landscapes, temporal palimpsests, and scale inflation in reportage.

Decisive tests. (i) Double-harvest stratigraphy aligned with hydraulic phases; (ii) co-films of cereal + legume on provisioning containers (starch morphometry; paleoproteomics; targeted biomarkers); (iii) granary microstratigraphy that records drying → storage sequences; (iv) container metrology clustering; (v) well-dated hydraulic episodes consistent with 118e. Positive legume co-signals in provisioning contexts will raise PP17 and favor DB; sustained cereal-only provisioning would strengthen SU.

6.7 Final Statement

Within the textual limits of Critias 115a–b and 118e, the most economical, testable reading is that “the dry sort … for our sustenance” names a dry, storable staple umbrella into which rice plausibly fits on processing and provisioning logic, while ὄσπρια marks the companion class. Order-3 consilience in the Sundaland case supports this reconstruction at the pattern level and prefers DB under co-residue evidence. Species-level identity for the companion is deliberately left open pending residues; when the analytics decide it, PP17 will register the interlock, and the reconstruction will tighten without rewriting the textual premise.

References

Primary sources

1929. Timaeus. Critias. Cleitophon. Menexenus. Epistles. Translated by R. G. Bury. Loeb Classical Library 234. Cambridge, MA: Harvard University Press.
1916. Enquiry into Plants, Vols. I–II. Translated by Arthur F. Hort. Loeb Classical Library 70 & 79. Cambridge, MA: Harvard University Press.

Lexica & digital tools

Liddell, H. G., R. Scott, and H. S. Jones. 1940. A Greek–English Lexicon, 9th ed., with revised supplement. Oxford: Clarendon Press. (LSJ)
LSJ Online. Thesaurus Linguae Graecae (TLG)/Scaife resources for quick lexical checks and Greek text access.

Greek foodways, agriculture, and categories (background)

Dalby, Andrew. 2003. Food in the Ancient World from A to Z. London & New York: Routledge.
Garnsey, Peter. 1999. Food and Society in Classical Antiquity. Cambridge: Cambridge University Press.
Isager, Signe, and Jens Erik Skydsgaard. 1992. Ancient Greek Agriculture: An Introduction. London & New York: Routledge.
Wilkins, John, and Robin Nadeau, eds. 2015. A Companion to Food in the Ancient World. Chichester: Wiley-Blackwell.

Rice origins, domestication, and spread (for Order-2/3 context)

Choi, J. Y., et al. 2017. “Multiple Origins but Single Domestication in Asian Rice.” Molecular Biology and Evolution 34(4): 969–979.
Fuller, Dorian Q. 2011. “Pathways to Asian Civilizations: Tracing the Origins and Spread of Rice and Rice Cultures.” Rice 4: 78–92.
Fuller, Dorian Q., et al. 2009. “The Domestication Process and Domestication Rate in Rice: Spikelet Bases from the Lower Yangtze.” Science 323(5919): 1607–1610.
Fuller, Dorian Q., Allaby, R., and Willcox, G. 2011. “Consilience of Genetics and Archaeobotany in the Entangled History of Rice.” Archaeological and Anthropological Sciences 2: 115–131.
Higham, Charles. 2014. Early Mainland Southeast Asia: From First Humans to Angkor. Bangkok: River Books.

Southeast Asian tidal irrigation & Kalimantan canal traditions (Order-3 instantiation)

Hatta, M., et al. 2023. “Managing the Newly-Opened Tidal Paddy Fields for Rice Production in Indonesia.” Agronomy 13(2): 327. (Open access.)
Niatika, U., et al. 2025. “Towards Sustainable Water Management in the Terusan Tengah Swamp Irrigation System, Central Kalimantan.” In IOP Conf. Ser.: Earth & Environmental Science.
Revina, S., et al. 2022. “The Effect of Tides on Water Salinity and Acidity in the Main Channel of Anjir Serapat Lowland Irrigation Area, Central Kalimantan.” Atlantis Press/Proc. ICOSEAT 2022.
Sulaiman, A. A., Sulaeman, Y., and Minasny, B. 2019. “A Framework for the Development of Wetland for Agricultural Use in Indonesia.” Resources 8(1): 34. (Includes sketches of anjir–handil–saka )
Sriyono, E. 2021. “Toward a Social Construction of Water Resources Management: The Case of Kalimantan.” Cogent Social Sciences 7(1). (Notes anjir, handil, saka canal typology.)
Workshop on Research Priorities in Tidal Swamp Rice. 1993. Manila: IRRI. (Background on Indonesian tidal-swamp rice; South Kalimantan emphasis.)

Project articles cited (for readers’ orientation)

Atlantis Java Sea Project. 2021. “Anjir System: An Atlantis Legacy Canal System for Flood Regulation in Flat Areas.” (Blog article; field context for tidal canals.)