The Two Waters and the Sanctuary — The Time of the Return

I · Two Conflations to Undo SPINE

Before the location — two taught errors to clear.

The mystery is old: Ezekiel measures a temple with a river that heals the Dead Sea (Ezekiel 40–47); Zechariah sees living waters going out from Jerusalem, half to one sea and half to another (Zechariah 14:8). It is commonly taught that these describe one and the same water, on the Temple Mount, foursquare like the New Jerusalem. That teaching rests on two conflations — and once they are cleared, the two prophecies do not compete. They harmonize.

The first conflation reads Ezekiel's temple as a scaled version of the Revelation-21 cube. But Revelation 21 is a city, foursquare and cubic — "the length and the breadth and the height of it are equal" — and it contains no temple at all: "I saw no temple therein: for the Lord God Almighty and the Lamb are the temple of it."^[9] Ezekiel's structure has a sanctuary, an altar, priestly service, and a water-source. To read one as a scaled version of the other is a category error — and it imports a foursquare symmetry that Ezekiel's own text never asks for.

The second conflation fuses Zechariah's source with Ezekiel's. But Zechariah names his source plainly: the living waters go out "from Jerusalem"(miyrushalaim)^[7]. Ezekiel's water issues from the sanctuary threshold (47:1), wherever that sanctuary stands. Two prophets, two sources. The teaching that makes them one is the thing to question — not the texts.

II · What Ezekiel Actually Measures SPINE

Read on its own terms — a directional sanctuary.

Strip the cube and read the reed. Ezekiel is set on a very high mountain as a vantage — "set me upon a very high mountain, by which was the frame of a city on the south" — the place the seer stands to look out, not where the temple sits and not where the water rises.^[1] From there he measures a specific, directional building:

The one place the text says "foursquare" (ravua) is the inner court — a hundred cubits by a hundred, ~52 m square^†^[2]. Not the whole complex. The outer boundary (42:20) is a wall "to make a separation between the sanctuary and the profane place" — a boundary of unstated shape, which may follow the ground^†. And the west end is closed: the separate place (gizrah) sits there, and the gates are on the east, north, and south — there is no west gate^[3]. The glory comes "from the way of the east," the east gate is the principal gate and is shut after the glory enters^[4], and the water issues from under the threshold, bound to the east^[5].

Ezekiel's own measurements, drawn to scale. Read without the Revelation-21 cube, the text gives a directional sanctuary: gated on east, north, and south; the west end closed by the separate place (no west gate); foursquare only at the inner court (40:47); the house on an east–west spine; — a compound of distinct cells, not a foursquare cube — with water issuing from under the threshold on the east side.^[2]^[3]^[5]

This is not a symmetrical cube. It is an east-facing, west-backed sanctuary — the exact shape a cliff-terrace at a spring-head produces. The topography a real site would need is not a vast level square; it is a height closed on one side, open toward the water on the other.

III · The Water of the Eastern Quarter SCAFFOLD

The living water of the east.

One word carries weight, and it must be read exactly. In Ezekiel 47:1 the Septuagint has both the water and the house's front as κατ' ἀνατολάς (kat' anatolas)^†. Anatolas is a noun — "the east" — and kata with the accusative is orientational: "toward, along, in the eastern quarter." It is the same phrase that describes the house facing east. So the text does not say the water "flows eastward" as a vector — that is the translators' inference — and it does not say the water comes "from the east," which would take a different preposition. What the word carries is orientation to the eastern quarter: the water is of the east^[6].

And qadim is a place-noun elsewhere in Ezekiel's own hand — "the east side," "the east border" (47:18; 48:1). Where the vision does speak of motion from the east, it is the glory: "the glory of the God of Israel came from the way of the east"^[4]. So the vision holds two easts at once — the glory from the east, and the water of the eastern quarter — and both point the search the same way: toward the eastern source, the great mountain.

The sea it heals is not in doubt. The waters "go down into the desert (the Arabah)… and the waters shall be healed"; the fishermen stand "from Engedi even unto Eneglaim" — both Dead Sea shore sites^[10]. The healed sea is the Dead Sea, the landmark every writer knew.

IV · The Reconciliation SPINE

Two sources — one healed sea.

Here is how both prophecies stand true at the same time. They describe two distinct living-water features, not one river forced through one channel:

Zechariah's waters go out from Jerusalem, dividing to two seas, summer and winter — Jerusalem's own drainage, reaching the Dead Sea from the west^[7]. Ezekiel's waters issue from the sanctuary threshold at the foot of the great mountain, running down the Jordan from the north to the same sea^[5]. They do not share a watershed. They converge at the Dead Sea — the one basin both are sent to heal.

The "former sea" and "hinder sea" of the divided-waters verse read cleanly on the front/back, leading/following sense of the words: the leading sea (Galilee, alive, the Jordan's upper reach) and the following sea (the Dead Sea, the terminal one)^[8]. The great flow runs the north–south spine — Hermon to Galilee to Jordan to Dead Sea — while Jerusalem's living water comes in from the side. Both arrive; the sea rises and freshens.

Two waters, one healed sea. The great flow originates in the east — Hermon's snowpack, the water coming from the eastern mountain (qadim) — and runs the north–south spine: Hermon → Galilee → Jordan → Dead Sea, the leading sea and the following sea of the divided-waters verse. From the west, Jerusalem's own living water enters as a side-feed. They do not share a channel; they converge at the sea they both heal, and the Jericho ford — Joshua's crossing, John's baptism — sits on the Jordan at the meeting-reach.^[14]^[7]^[10]

And the meeting-place has a name you already know. On the Jordan axis, just above the Dead Sea, sits the Jericho ford — Joshua's crossing into the land, and John's baptism, where the companion exhibit Rivers of Living Water maps the two wadis converging. It is the reach where the northern living water meets its terminus, above the sea both waters heal. The ~550 CE high-stand and the freshening to Ein Gedi are the physical trace of a basin receiving more than it loses — the healing made visible in the record.

V · The Candidate at the Foot of Hermon SCAFFOLD

Where the sanctuary-source could stand — Caesarea Philippi.

If Ezekiel's water is "the living water of the east," from the great mountain, one site answers the description more nearly than any other: Banias / Caesarea Philippi, at the foot of Hermon. A powerful spring issues from a cavern at the base of a sheer cliff — one of the Jordan's sources, fed by Hermon's snowpack from the east^[11]. The sacred precinct sits on a terrace with the cliff as its back wall. The water starts shallow at the ledge, deepens, and runs down to the Jordan and the Dead Sea.

The type fits with unusual precision: a cliff-backed spring-sanctuary, closed on the backed side, opening to the plain, living water from the east, at the foot of the very-high-mountain of the vantage verse. The civic ground is large enough to hold the measured precinct; a monumental north gateway opened toward the grotto — the "gate of Hades." And it is here that Christ stood and said, "upon this rock I will build my church; and the gates of hell shall not prevail against it" — and the archaeology shows exactly that: a church built over the Pan temple, the idol's niche become the apse, the pagan altar broken into the wall, in the very era the corpus reads as the serpent's binding^[13].

One geometric note the survey makes possible. If a sanctuary stood on the plateau above the grotto — the high ground from whose northeastern edge the spring emerges^[16] — then water issuing from beneath it would emerge on the structure's eastern side, at the plateau's edge over the grotto. That matches Ezekiel's kat' anatolas exactly: not a claim about which way the water travels, but where it issues — from under the house, on the east. The plateau geometry and the Greek orientation agree.

And there is a further mark on this water, dated and dramatic. In 1033 the great Jordan Valley earthquake collapsed much of the Banias cave and blocked the spring that once gushed from it — the water that had issued from under the rock was reduced to a seep at the terrace foot, where it remains^[15]. A near-contemporary chronicler recorded a "wonder" at Banias in that era, when "two mountains met together." If the living water of the sanctuary-source ran during the age the corpus measures, here is a documented moment when it was stopped: the threshold ceased to issue, in the eleventh century, when the ground itself moved.

— The standing seam, held in the open —

One thing does not fit, and it is not hidden. Ezekiel's sanctuary faces east — the glory from the east, the east gate, the principal axis. Banias is backed by cliff on the north and opens south; its building-axis is rotated ~90°^†^[12]. The lexical reading — living water of the east, from Hermon — resolves the water's source-direction in the site's favor. But the gate-axis orientation remains a genuine seam. And the documented remains are pagan-then-church, not a recovered temple of the LORD. So "Ezekiel's temple stood here" is held as conviction, fenced — not shown. What Christ promised at this rock was a church, and a church is what the ground delivers; whether it is also Ezekiel's sanctuary is the mystery this exhibit does not close.

— The ground, and the test that would settle it —

Honesty first about the ledge itself: the sacred terrace is the rubble-floor of a collapsed karstic cave (~100 × 40 m), and the cliff is that cave's original back wall — a natural feature, not an engineered platform^[16]. The documented structures on it are the pagan-then-church sequence. But the survey also records what has not been dug: the spring emerges at the edge of a broad plateau above the grotto, and the eastern suburb and the cliff-top corridor remain unexcavated. So the honest claim is narrow and testable: the upland and eastern ground here has room for a larger structure and has not been systematically examined. If it were surveyed — spade or LiDAR — and found to hold a west-closed, east-oriented sanctuary on Ezekiel's ~52 m module with a threshold water-channel, the identification would move from conviction to demonstrated. If it holds only more of the documented Roman-Byzantine city, or nothing, the identification falls — and the reconciliation still stands on the texts alone.

VI · The Attested Class, and the Dated Vision SPINE

Sanctuaries for a season — a documented category, and a timestamped blueprint.

The claim-shape "a sanctuary of the LORD existed for a season, for a purpose, away from the expected mountain" is not a novelty this exhibit invents. It is a class with three documented members. At Elephantine, a temple of YHW with altar and offerings stood on an island in the Nile through the fifth century BCE — older than Cambyses' arrival by the community's own petition — until it was destroyed in 410 BCE; we know it from the congregation's own papyri, including their correspondence with Jerusalem about rebuilding^[17]. At Leontopolis, a Zadokite priest built a temple in Egypt around 160 BCE and ran sacrifices for some two hundred thirty years until Vespasian closed it — and he built it on a proof-text: "In that day shall there be an altar to the LORD in the midst of the land of Egypt" (Isa 19:19). Scripture itself chartered a for-a-season altar in the least obvious location on the map, and history records a priest acting on the charter^[18]. And on Gerizim the Samaritan sanctuary stood from the Persian period to 110 BCE — real enough that the location question was still live at a well in Sychar, where the Lord's answer relativized both expected mountains: neither in this mountain, nor yet at Jerusalem… in spirit and in truth^[19]. Three sanctuaries, three seasons, three endings. The category is attested. What this does for the Banias case is precise and limited: it converts "Ezekiel's temple existed for a season, elsewhere" from a claim without a genus into a proposed fourth member of a documented class — which raises the prior and proves nothing. The eastern-terrace test remains the test; the class carries the plausibility, the spade carries the proof.

And the blueprint itself bears a date. Ezekiel 40:1: "In the beginning of the year, in the tenth day of the month… in the selfsame day the hand of the LORD was upon me." The readings of rosh hashanah divide — the seventh-month head makes the date Yom Kippur (the one day a year's "head" can fall on a tenth is the Jubilee, proclaimed by trumpet on the Day of Atonement, Lev 25:9); the first-month reading makes it Nisan 10, lamb-selection day, the Jordan-crossing day. The division is real and printed. The remarkable fact survives it: both candidates are the calendar's two tenths — the vision of the temple whose location was never the obvious one arrived timestamped, on either reading, to a Tenth^[20]. On the atonement-day reading, the builder's further step — that the tenth of the seventh month is the logical day for the heavens to rend and the heavenly Zion to come into view, the congregation's day of watching the door — is his held reading, CONVICTION by this exhibit's own tiers, resting on the emergence-grammar of Hebrews 9:28 and carried with the fences of the companion Day of Atonement exhibit. The date in the verse is Spine. The year it points to is not in the verse.

VII · Where the Weight Sits SPINE

The joints, and what each costs if pulled.

the Rev-21 conflation, undone

Ezekiel's temple is not a scaled Revelation-21 cube.

SPINE: Rev 21 is a CITY, foursquare and cubic (21:16), containing NO temple (21:22). Ezekiel 40–47 is a TEMPLE with sanctuary, altar, priestly service, and a water-source. The one 'foursquare' in Ezekiel is the inner court (40:47), not the whole complex; the outer boundary (42:20) is a separation-wall of unstated shape. Reading Ezekiel through Rev 21 imports a symmetry the text does not require.

If pulled: If the conflation stands, every 'temple location' argument inherits a foursquare cube that Ezekiel never describes — and rules out real, directional, cliff-backed sites on a shape borrowed from a different vision. Undoing it costs nothing and frees the reading to follow Ezekiel's own measurements.

the two sources, distinct

Zechariah's Jerusalem water and Ezekiel's sanctuary water are not the same feature.

SPINE: Zech 14:8 says the living waters go out 'from Jerusalem' — named plainly. Ezekiel's water issues from the sanctuary threshold (47:1), wherever that sanctuary sits. The common teaching that fuses them is itself a conflation. Separated, they harmonize: two sources, one healed sea.

If pulled: If they must be one source, the two texts collide (one river vs. a divided one; Jerusalem vs. the threshold) and force a single impossible channel. Separated, both are literally true at once — the reconciliation is Spine, resting on the plain texts, not on any site identification.

the convergence at the ford

The two waters meet at the Dead Sea and along the Jordan — at the Joshua crossing / John's baptism.

SCAFFOLD: Ezekiel's water reaches the Dead Sea down the Jordan from the north; Jerusalem's reaches it from the west by its own drainage. They converge in the sea they both heal; the Jericho ford — Joshua's crossing, John's baptism — sits on the Jordan axis at the meeting-reach. The ~550 CE high-stand and the freshening to Ein Gedi are the physical trace.

If pulled: If the reader demands both waters flow through the ford from ONE spring, that re-conflates the sources and the topography breaks (Jerusalem sits on the divide; its water cannot join the northern Jordan). The honest claim is convergence AT THE SEA and along the axis — which holds, and which the ford marks.

Banias as the sanctuary site

Banias is the strongest candidate — but the identification is fenced.

SCAFFOLD/CONVICTION: the type fits (cliff-backed spring-sanctuary at Hermon's foot; living water from the east; scale that holds the ~52 m core and ~104 m spine; a north gateway toward the gate of Hades; the Matthew-16 church on the rock in the binding era). The GATE-axis orientation is the standing seam. The documented remains are pagan-then-church, not a recovered hêkâl.

If pulled: If Banias is not Ezekiel's temple, the RECONCILIATION is untouched — two distinct waters healing one sea stands on the texts alone. Only the site-identification falls. The eastern-terrace excavation is the named test; until it is run, 'Ezekiel's temple was here' is held, not shown.

the attested class

Sanctuaries for a season, away from the expected mountain, are a documented category — three members and a charter verse.

SPINE: Elephantine (YHW temple with offerings, 5th c. BCE, destroyed 410, known from its own papyri); Leontopolis (temple of Onias, ~160 BCE–73 CE, built on Isa 19:19's altar-in-Egypt and closed by Vespasian, per Josephus); Gerizim (Persian period–110 BCE, the 'this mountain' of John 4). The claim 'it existed for a season, for a purpose, not where expected' has a genus with three attested members and a prophetic charter.

If pulled: The class raises the PRIOR of the Banias identification; it cannot substitute for the dig. 'Existed for a season' and 'was here' share the same named test — the eastern terrace. If the reader inflates precedent into proof, the exhibit's own standard (consistent-with is never proven-by) is the first casualty; the honest gain is plausibility of kind, not demonstration of instance.

the vision's date

Ezekiel's temple vision is timestamped, on either reading of 40:1, to one of the calendar's two Tenths.

SPINE for the date-arithmetic; the readings divide honestly. Rosh-hashanah-as-seventh-month yields Yom Kippur of a Jubilee-marked year (Lev 25:9 — the only 'head of the year' that functions on a tenth); rosh-hashanah-as-first-month yields Nisan 10, the lamb-selection and Jordan-crossing day. Both candidate dates are the two tenths this corpus tracks. The builder's atonement-day preference, and his reading of that day as the day the heavens rend and the heavenly Zion comes into view, are CONVICTION, carried under Hebrews 9:28's emergence-grammar and the companion exhibit's fences.

If pulled: If the Nisan reading is right, the Yom Kippur resonance moves from the verse to the type — nothing else in this exhibit shifts, since the reconciliation and the site case never rested on the date. If the reader converts the timestamp into a year-prediction, Mark 13:32 is pulled out from under the whole corpus; the date in the verse names a day of the calendar, never a turn of it.

The Search forEzekiel's Templeand the Living Waters of Zechariah