1) Fact-check: what the texts actually say
✅ The “Goguryeo 690,000 households / Baekje 760,000 households” claim is real
In Xin Tang Shu (新唐書), the entry states that Tang “took five divisions, 176 fortified places/cities, and 690,000 households” for Goguryeo, and for Baekje “five divisions, 37 prefectures, 200 fortified places/cities, and 760,000 households.” (zh.wikisource.org)
So the numbers exist in the source—your post isn’t inventing them.
✅ The “176 (Goguryeo) vs 200 (Baekje) fortresses/cities” figure is also in the same passage
Same Xin Tang Shu section gives 176 and 200 in the same breath as the household totals. (zh.wikisource.org)
✅ Samguk Sagi does preserve lists of forts that surrendered and forts that did not
In the Samguk Sagi geography/miscellany (잡지), there is an entry explicitly titled “Fortresses north of the Amnok that surrendered” and another titled “Fortresses north of the Amnok that did not surrender.” (한국사데이터베이스)
What I can verify from the official DB pages is that there are categorized lists (surrendered vs not surrendered). (한국사데이터베이스)
But the exact “11 resisted / 7 fled” summary (as phrased in the gallery post) looks like it may be a misread or a compression of those lists and later events.
People love clean numbers because they feel like truth.
But in wartime, numbers often measure something else: reach.
One line in the New Book of Tang (Xin Tang Shu) has become a quiet trap for modern readers. It records that after Goguryeo’s fall, Tang “took” 176 fortified places and counted 690,000 households, while Baekje—often portrayed as smaller in territory and military weight—appears with 760,000 households and 200 fortified places. (zh.wikisource.org)
At first glance, it reads like a scandal: Was Baekje secretly larger? Was Goguryeo exaggerated? Did someone lie?
A better question is simpler—and sharper:
What exactly is being counted?
Households don’t equal “the whole population”
The word “household” in historical sources is rarely a neutral demographic unit. It’s usually a state-facing unit: a taxable, registrable, governable cell. In peaceful times, that can approximate population. In conquest conditions, it often becomes something narrower:
people who didn’t flee
people living in places the new authority can reach
people who accept registration (voluntarily or under pressure)
people in administrative zones the conqueror can actually administer
That means the “household” figure may be less a mirror of society and more a map of control.
Fortress wars create blind spots by design
Goguryeo wasn’t a soft-bodied kingdom. It was a fortress ecology—terrain, walls, garrisons, supply lines, and local power nodes.
And fortress wars don’t work like modern “paint the map” conquest. They work like cutting arteries.
You don’t need to occupy every valley to break a kingdom. You need to seize the route, the gateway, the hinge fortresses that make movement possible. Once those fall, a capital can be isolated and struck—even if large areas remain messy, resisting, or simply unregistered.
That’s why it matters that Samguk Sagi preserves categorized lists of fortresses that surrendered and those that did not in the north of the Amnok region. (한국사데이터베이스)
Even without dramatizing the exact headcount, the structure of the record points to a reality modern readers intuitively recognize: collapse is not the same as total capture.
A practical hypothesis: the Tang number is a “registration footprint”
Here is the hypothesis worth publishing:
The “690,000 households” recorded for Goguryeo may represent the households Tang could realistically register immediately after victory—centered on the places it had taken, secured, and reorganized—rather than the total number of households that had ever existed in Goguryeo’s full prewar footprint. (zh.wikisource.org)
This does not require conspiracy. It only requires friction: refugees, relocation, holdouts, border peoples, and bureaucratic limits.
Then why does Baekje look larger?
This is the part that makes the puzzle emotionally satisfying.
Baekje’s core territory, especially late Baekje, is often imagined as more tightly centered on the peninsula’s southwest and its agricultural base—areas that can be administratively “counted” in a different way after conquest. Meanwhile Goguryeo’s footprint includes borderlands and fortress zones where registration is hardest precisely when war has just ended.
So the comparison may be unfair, not because one kingdom was “smaller,” but because the conqueror’s clipboard sees different things in different landscapes.
If you want this post to feel like a “mini research report,” give readers tests, not just vibes:
Compare multiple Tang-era records (where possible) for consistency of the figures and terminology.
Track fortress counts and administrative reorganizations: when a place becomes a prefecture/commandery, it becomes countable.
Treat household figures as “state capture,” not “life on the ground.” Then the paradox becomes a feature, not a bug.
The real mystery isn’t population. It’s state vision.
In the end, the most interesting story here isn’t “who had more people.”
It’s this:
When an empire says it counted you, it may only mean it counted the part of you it could touch.
And sometimes, that difference—between a living society and a registered society—is where history hides its biggest lies.
2) Media & games that fit this theme
Watch/read vibes (easy entry)
Korean historical dramas that touch the post-Goguryeo → Balhae memory line: Dae Jo Young (대조영) is the obvious gateway (migration, remnants, frontier state-building).
Goguryeo war focus: dramas/films centered on fortress defense and Sui/Tang pressure (even if stylized) help readers “feel” why corridor conquest matters.
Games (theme-aligned)
Civilization: the “fortress corridor vs total territory” idea is perfect for Civ-style mechanics.
Paradox-style grand strategy: this topic screams state capacity (administration), control vs autonomy, fort networks, refugee movement, census mismatch—all core Paradox DNA.
3) Civilization mod idea: Wonder design (plug-and-play)
If you want one Wonder that matches this article’s thesis, make it about fortress administration + census reach (not just “more soldiers”).
Wonder: Liaodong Fortress Belt (요동 방어선)
Era: Classical (or Medieval if your mod’s pacing is slower)
Cost: 250 (Classical) / 400 (Medieval) production
Requires: Walls in the city (and Construction tech equivalent)
Effects (Civ V-ish, readable and strong):
City gains +25% Ranged Combat Strength and +1 Range for city attacks (fortress city fantasy)
Free Great General
+1 Gold and +1 Culture for each Fort you control (makes “fort network” matter)
When completed, instantly reveals a “Registration Footprint” UI concept (optional): +10% production when building Courthouse / administrative buildings (state vision theme)
Why it fits: This wonder turns the essay’s point into gameplay: control is nodes + administration, not just land paint.
Alternate (more thematic, less combat):
Wonder: “Household Register Office (호적청)”
Era: Medieval
Cost: 350
Effect: +15% Gold, +15% Production in the city; +1 Happiness per 2 population empire-wide for 20 turns (a “state consolidation surge”)
4) Paradox mod pitch: what to build from this
If Civ feels too “boardgame clean,” Paradox is where this idea becomes art.
Core mechanic: Control vs Count
Provinces have:
Control (military presence)
Registration (taxable/administrable population)
Conquest gives you control first; registration lags unless you invest in:
forts, roads, bureaucrats, hostages/elite co-optation, resettlement policy
Result: you can “win” a war and still have low usable manpower and tax—which mirrors the entire household-number paradox.
Event chain ideas
“Fortress Corridor Secured” → capital vulnerable
“Unregistered Remnants” → rebels, migration, frontier state seed (hello Balhae-style emergence)
“Census Shock” → official numbers jump later, revealing that early counts were partial
No comments:
Post a Comment