-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathevolution.html
More file actions
438 lines (386 loc) · 50.9 KB
/
Copy pathevolution.html
File metadata and controls
438 lines (386 loc) · 50.9 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>IV · Evolutionary Cosmology — The Vacuum Game</title>
<meta name="description" content="Once civilizations can choose vacua, universes become organisms with heredity: the five types of cosmological evolution, the master equation M_ab, vacuum ecology, the new selection pressures, and whether the war ever completes.">
<link rel="preconnect" href="https://fonts.googleapis.com">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link href="https://fonts.googleapis.com/css2?family=Spectral:ital,wght@0,500;0,600;0,700;1,500;1,600&family=Source+Serif+4:ital,opsz,wght@0,8..60,400;0,8..60,600;1,8..60,400;1,8..60,600&family=IBM+Plex+Mono:wght@400;500&display=swap" rel="stylesheet">
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/katex@0.16.11/dist/katex.min.css">
<link rel="stylesheet" href="css/main.css">
<style>
/* page-local: the equation-anatomy panel for FIG IV.2 */
.eqa { padding: 1.7rem 1.2rem 0.5rem; text-align: center; }
.eqa .katex { font-size: 1.2em; }
.eqa-kicker {
font-family: var(--font-mono);
font-size: 0.66rem;
letter-spacing: 0.2em;
text-transform: uppercase;
color: var(--muted);
margin: 0 0 0.2rem;
}
.eqa-legend {
display: grid;
grid-template-columns: repeat(auto-fit, minmax(16rem, 1fr));
gap: 0.55rem 1.5rem;
padding: 0.7rem 1.3rem 1.3rem;
border-top: 1px solid var(--hairline);
background: var(--surface-2);
}
.eqa-legend .term {
border-left: 2px solid var(--t-c, var(--muted));
padding-left: 0.7rem;
font-size: 0.9rem;
line-height: 1.5;
}
.eqa-legend .term .sym {
font-family: var(--font-mono);
font-size: 0.72rem;
letter-spacing: 0.14em;
text-transform: uppercase;
color: var(--t-c, var(--muted));
display: block;
}
/* symbols keep their own case — μ must not be uppercased into M */
.eqa-legend .term .sym i, .eqa-kicker i { font-style: normal; text-transform: none; }
.eqa-legend .term p { margin: 0; color: var(--ink-2); }
.eqa-legend .term b { color: var(--ink); font-weight: 600; }
.eqa-legend .k-ascent { --t-c: var(--vac-ascent); }
.eqa-legend .k-light { --t-c: var(--vac-light-bright); }
.eqa-legend .k-false { --t-c: var(--vac-false); }
.eqa-legend .k-terminal { --t-c: var(--vac-terminal); }
.eqa-legend .k-machine { --t-c: var(--vac-machine); }
.eqa-legend .k-harbour { --t-c: var(--vac-harbour); }
</style>
<script defer src="https://cdn.jsdelivr.net/npm/katex@0.16.11/dist/katex.min.js"></script>
<script defer src="https://cdn.jsdelivr.net/npm/katex@0.16.11/dist/contrib/auto-render.min.js"></script>
<script defer src="js/common.js"></script>
<script defer src="js/evolution.js"></script>
</head>
<body>
<header class="site-nav">
<a class="brand" href="index.html">THE VACUUM GAME<span class="tick">_</span></a>
<nav aria-label="Chapters">
<a href="physics.html"><span class="n">I</span>Rulebook</a>
<a href="game.html"><span class="n">II</span>The Game</a>
<a href="survival.html"><span class="n">III</span>Survival</a>
<a href="evolution.html"><span class="n">IV</span>Evolution</a>
<a href="evidence.html"><span class="n">V</span>Evidence</a>
<a href="fiction.html"><span class="n">VI</span>Stories</a>
</nav>
</header>
<div class="hero" data-tint="harbour">
<div class="hero-inner">
<p class="eyebrow">Chapter IV<span class="sep">·</span>The view from outside</p>
<h1>Evolutionary Cosmology</h1>
<p class="thesis">Does vacuum engineering change evolutionary cosmology? <strong>Completely — qualitatively, not merely quantitatively.</strong> Once a civilization can choose which vacuum comes next, universes acquire the one thing Darwin demanded and cosmology has never convincingly supplied: heredity — with a chooser attached. Mutation stops being random, generation time collapses from cosmological to civilizational, and what is under selection is no longer a list of constants but a compound organism: vacuum plus culture plus policy. Descent with modification acquires intent.</p>
</div>
</div>
<div class="page">
<aside class="toc" aria-label="Table of contents"></aside>
<main class="content">
<!-- ================================================================ -->
<section class="sec">
<h2 id="vocab"><span class="secnum">4.1 · Vocabulary</span>What it takes for a universe to have descendants</h2>
<p class="lede">Darwin needs three things: variation, heredity, differential reproduction. Cosmology has spent thirty years arguing whether universes have any of them. Vacuum engineering supplies all three — and adds a fourth biology never had: a chooser.</p>
<div class="box box-vocab">
<p class="box-label">Vocabulary</p>
<dl>
<dt>Heredity</dt>
<dd>Offspring resemble parents <em>because</em> they descend from them. The ingredient the multiverse usually lacks: eternal inflation produces enormous variation, but a pocket universe’s physics is drawn afresh, not inherited from a parent pocket. No heredity, no evolution; only sampling.</dd>
<dt>Unit of selection</dt>
<dd>What actually gets copied, and therefore what natural selection sees. In biology the candidates are gene, organism, group. Here: a set of constants, a vacuum phase, or — this chapter’s answer — a vacuum <em>together with</em> the civilization tending it.</dd>
<dt>Fitness function</dt>
<dd>Whatever quantity gets maximized by the mere passage of generations. Never chosen; it falls out of the reproduction mechanism. Change how universes reproduce and you change what the multiverse optimizes, whether anyone intended it or not.</dd>
<dt>Generation time</dt>
<dd>Interval between a parent’s birth and its offspring’s; the clock speed of all evolution. In <a href="https://en.wikipedia.org/wiki/Cosmological_natural_selection" target="_blank" rel="noopener">Smolin’s scheme</a>, the several billion years a universe needs to make black holes; under engineering, however long a civilization takes to decide.</dd>
<dt>Lamarckian</dt>
<dd>Inheritance of acquired characteristics — the parent passes on what it <em>learned</em>, not just what it was born with. False for genes, true for language, true for vacuum engineering: a civilization that maps the landscape can install its conclusions in its children.</dd>
<dt>Percolation</dt>
<dd>Whether locally growing patches ever join into one connected region spanning everything. Borrowed from materials physics and network theory, and the right question for a universe full of expanding bubbles: does the conversion <em>complete</em>, or does unconverted space survive forever between them? (§4.7.)</dd>
</dl>
</div>
<p>Keep one asymmetry in view throughout. In biology the organism does not choose its offspring’s genome; the copying machinery is blind, the environment edits afterwards. Where <a href="physics.html#rule4">triggering is possible</a> and <a href="physics.html#fineprint">the landscape is mapped</a>, that machinery is a committee with a research programme. Everything strange in this chapter follows from moving the mutation operator inside the mind of the replicator.</p>
</section>
<!-- ================================================================ -->
<section class="sec">
<h2 id="taxonomy"><span class="secnum">4.2 · The taxonomy</span>Five types of cosmological evolution</h2>
<p>The literature is usually presented as a debate — Smolin versus the anthropic principle, Crane versus Smolin — but it reads better as a ladder, each rung adding a cumulative ingredient. Types 0 through II are real published positions. Types III and IV are what those positions become once <a href="game.html#structure">deliberate vacuum engineering</a> enters, and they are where this codex lives.</p>
<h3>Type 0 — the anthropic sieve</h3>
<p>Eternal inflation manufactures an unbounded ensemble of pocket universes with different low-energy physics. Nothing in it reproduces: pockets have no children that resemble them. What looks like selection is entirely <span class="gl" data-g="Observer selection: the observation that conditions are compatible with observers carries no explanatory force beyond the fact that only such conditions can be observed at all.">observer selection</span> — we find ourselves where observers can exist, as you find yourself on a planet rather than the vacuum between planets. A sieve, not a Darwinian process — and a sieve predicts nothing about what lies on the other side of it. (Chapter II’s <a href="game.html#anthropic">anthropic bound</a> is the same reasoning put to different work — bounding how many trigger-capable neighbours we can have.)</p>
<h3>Type I — Smolin’s cosmological natural selection</h3>
<p><a href="https://en.wikipedia.org/wiki/Cosmological_natural_selection" target="_blank" rel="noopener">Lee Smolin (1992)</a> adds the missing ingredient. Suppose every black hole bounces into a new expanding region — a daughter universe — whose constants differ slightly from the parent’s. Now there is heredity (children resemble parents), variation (small random mutation), and differential reproduction (universes whose physics makes many black holes leave more descendants). The prediction is admirably sharp: our constants should sit at a <em>local maximum</em> of black-hole fecundity, so nudging any one reduces the count. The mechanism is purely genealogical — parent and child never share a spacetime, mutation is blind, nobody chooses.</p>
<h3>Type II — the meduso-anthropic principle (Crane, Harrison)</h3>
<p><a href="https://arxiv.org/abs/hep-th/9402104" target="_blank" rel="noopener">Louis Crane (1994)</a> and, independently, <a href="https://ui.adsabs.harvard.edu/abs/1995QJRAS..36..193H" target="_blank" rel="noopener">Edward Harrison (1995)</a> noticed that intelligence is not a bystander in Smolin’s loop. If technological civilizations manufacture black holes — for <a href="physics.html#rule4">starship drives</a>, power, computation, the engineering of it — universes supporting civilizations gain a reproductive subsidy owing nothing to stellar astrophysics. Fitness now runs through biology and technology; the constants drift biophilic as a <em>side effect</em>, not anyone’s purpose. Crane calls the resulting alternation of phases the jellyfish (§4.6). Intelligence has entered the loop but still isn’t steering: it produces offspring without choosing them.</p>
<h3>Type III — engineered</h3>
<p>Now hand that civilization the landscape map from <a href="physics.html#fineprint">Chapter I</a> and the trigger technology of Rule four. Five things break at once, each a change of kind rather than degree:</p>
<ol>
<li><strong>Reproduction decouples from black holes and goes lateral.</strong> Parent-vacuum conversion propagates a chosen phase <em>through the spacetime you already occupy</em>, at lightspeed, no new asymptotic region needed. Descent stops needing a birth canal.</li>
<li><strong>Heredity becomes chosen and high-fidelity.</strong> The blind mutation distribution is replaced by a policy: which vacuum, selected on purpose, targeted as accurately as your cartography allows. This is <span class="gl" data-g="Lamarckian: the parent transmits characteristics it acquired during its own lifetime — here, everything the civilization learned about the landscape.">Lamarckian</span> inheritance at cosmological scale.</li>
<li><strong>Generation time collapses.</strong> From the billions of years a universe needs to make black holes down to the interval between a decision and a launch window. Cosmological evolution acquires a civilizational clock speed — a factor of 10<sup>6</sup> or worse in the exponent of everything.</li>
<li><strong>The unit of selection changes.</strong> What propagates is the vacuum-lineage <em>plus the culture tending it</em> — and neither replicates without the other.</li>
<li><strong>Search replaces drift.</strong> Landscape exploration stops being a random walk and becomes a directed search, with all the acceleration — and all the pathologies of premature convergence — that implies.</li>
</ol>
<p>Fitness in Type III is not black-hole count but <strong>producing civilizations that propagate your vacuum</strong>: a universe is fit if the minds it grows survive long enough, understand enough, and want to reproduce it. A fitness function with opinions in it.</p>
<h3>Type IV — cultivated fixation</h3>
<p>Run Type III forward. Directed search converges: lineages find stable floors — a vacuum nobody can improve on and nothing decays past, the <a href="game.html#cooperation">Schelling floor</a> of Chapter II — and install them. Conversion stops: nothing left to convert, nobody who wants to. Evolution persists in exactly two places: at <a href="physics.html#rule2">frozen fronts</a>, where two fixed vacua meet and structure crosses the border — cosmological horizontal transfer — and in wild volume no civilization reached, where Type 0 physics runs unattended. The multiverse anneals like a polycrystal: large ordered domains, all the interesting chemistry at the grain boundaries.</p>
<figure class="fig">
<div class="fig-panel">
<canvas id="fig-taxonomy"></canvas>
<div class="fig-controls">
<button class="btn" id="tx-0" aria-pressed="false">Type 0</button>
<button class="btn" id="tx-1" aria-pressed="false">Type I</button>
<button class="btn" id="tx-2" aria-pressed="false">Type II</button>
<button class="btn" id="tx-3" aria-pressed="true">Type III</button>
<button class="btn" id="tx-4" aria-pressed="false">Type IV</button>
</div>
</div>
<figcaption class="fig-caption"><span class="fig-no">FIG IV.1 — Five kinds of cosmological descent.</span>
Flip through the rungs and watch two numbers move. <em>Unit of selection</em> starts as nothing (Type 0 has no replicator), becomes a set of constants (I), then constants-plus-biology (II), a vacuum bonded to a culture (III), a frozen domain trading across its borders (IV). <em>Generation time</em> starts undefined, sits at billions of years while black holes do the copying, then collapses to the length of a political argument once reproduction is a decision. Type III is the only rung whose arrows spread sideways through existing space instead of downward into new spacetimes — the change that makes this a war rather than a family tree.</figcaption>
</figure>
<div class="tbl-wrap">
<table class="tbl">
<caption>The ladder, in five rows</caption>
<thead>
<tr><th>Type</th><th>Heredity</th><th>Unit of selection</th><th>Fitness function</th><th>Generation time</th></tr>
</thead>
<tbody>
<tr><td>0 · anthropic sieve</td><td>None — variation drawn afresh</td><td>None (no replicator)</td><td>Undefined; only observability filters</td><td>—</td></tr>
<tr><td>I · Smolin</td><td>Genealogical, blind, low-fidelity</td><td>A set of constants</td><td>Black-hole fecundity</td><td>~10<sup>9</sup>–10<sup>10</sup> yr</td></tr>
<tr><td>II · Crane / Harrison</td><td>Genealogical, blind, biology-amplified</td><td>Constants + the biosphere they permit</td><td>Black-hole fecundity, industrially boosted</td><td>Life of a civilization</td></tr>
<tr><td>III · engineered</td><td>Chosen, high-fidelity, Lamarckian</td><td>Vacuum-lineage + its culture</td><td>Producing civilizations that propagate your vacuum</td><td>A decision cycle</td></tr>
<tr><td>IV · cultivated fixation</td><td>Frozen; horizontal transfer only</td><td>The domain, and its grain boundaries</td><td>Holding your floor; trading at the fronts</td><td>Effectively infinite</td></tr>
</tbody>
</table>
</div>
<p>Underneath the taxonomy sit three distinct physical processes, worth separating because they map onto three different chapters of biology. Only one is a weapon.</p>
<div class="tbl-wrap">
<table class="tbl">
<caption>Three processes, three evolutionary analogues</caption>
<thead>
<tr><th>Process</th><th>What happens physically</th><th>Strategic character</th><th>Evolutionary analogue</th></tr>
</thead>
<tbody>
<tr><td>Parent-vacuum conversion</td><td>A bubble expands through the existing spacetime, replacing one vacuum with another</td><td>Terraforming, conquest, or doomsday weapon</td><td><strong>Ecological competition between phases</strong></td></tr>
<tr><td>Baby-universe production</td><td>An inflating region forms a new asymptotic spacetime and pinches off (<a href="survival.html#baby">Chapter III</a>)</td><td>Reproduction, diversification, possible escape</td><td><strong>Genealogical descent</strong></td></tr>
<tr><td>Artificial black-hole production</td><td>Manufactured holes may, under Smolin-like assumptions, bounce into daughter universes</td><td>Industrial reproduction — perhaps inadvertent</td><td><strong>Crane’s meduso-anthropic selection</strong></td></tr>
</tbody>
</table>
</div>
<div class="box box-caveat">
<p class="box-label">Caveat — how speculative is each rung?</p>
<p>Type 0 needs only eternal inflation and a landscape. Type I needs black-hole interiors to bounce into new expanding regions with slightly altered parameters — no accepted quantum-gravity calculation delivers this, and Smolin’s predictions are contested (notably over whether neutron-star upper mass limits already falsify the local-maximum claim). Type II needs Type I <em>plus</em> industrial black holes counting as reproductive. Types III and IV need the <a href="physics.html#fineprint">full assumption stack</a>: catalysable transitions, targetable destinations, something surviving the wall. None of it is established physics. What is defensible is the conditional: <em>if</em> vacuum engineering is possible at all, the evolutionary picture changes as described — whether or not anyone intends evolution.</p>
</div>
</section>
<!-- ================================================================ -->
<section class="sec">
<h2 id="mab"><span class="secnum">4.3 · The master equation</span>One line that contains all four rungs</h2>
<p>The whole taxonomy collapses into a single reproduction kernel. Write \(M_{ab}\) for the expected number of <em>b</em>-type daughter universes produced by a single <em>a</em>-type universe. Two terms: what happens by itself, and what happens because somebody decided.</p>
<figure class="fig">
<div class="fig-panel">
<div class="eqa">
<p class="eqa-kicker"><i>M<sub>ab</sub></i> — expected number of b-type daughters of an a-type universe</p>
<div class="eq">$$M_{ab} \;=\; \underbrace{\textcolor{#9085e9}{B_a}\,\textcolor{#f2c14e}{\mu_{ab}}}_{\text{Smolin: blind genealogy}} \;+\; \underbrace{\textcolor{#3987e5}{I_a}\,\textcolor{#e66767}{S_a}\,\textcolor{#d55181}{E_a}\,\textcolor{#199e70}{P_{ab}}}_{\text{engineered: chosen genealogy}}$$</div>
</div>
<div class="eqa-legend">
<div class="term k-ascent"><span class="sym"><i>B<sub>a</sub></i> — fecundity</span><p><b>Naturally generated daughters.</b> How many new universes vacuum <em>a</em> produces on its own, e.g. through black holes. Crane’s move is to raise this term with industry.</p></div>
<div class="term k-light"><span class="sym"><i>μ<sub>ab</sub></i> — mutation</span><p><b>Natural mutation distribution.</b> Given a parent of type <em>a</em>, the blind spread of daughter types <em>b</em>. Narrow and unsteerable; nobody’s policy.</p></div>
<div class="term k-false"><span class="sym"><i>I<sub>a</sub></i> — intelligence</span><p><b>Probability intelligence arises.</b> Chemistry, stars, time, a habitable window. The term all conventional fine-tuning arguments are about.</p></div>
<div class="term k-terminal"><span class="sym"><i>S<sub>a</sub></i> — survival</span><p><b>Probability it survives to universe-engineering capability.</b> The Great Filter as a coefficient: every self-destruction, every <a href="game.html#casec">race to the bottom</a>, sits in here.</p></div>
<div class="term k-machine"><span class="sym"><i>E<sub>a</sub></i> — effort</span><p><b>Number of engineered offspring.</b> Industrial capacity times the will to use it: foundries, membranes, launch windows, centuries of nerve.</p></div>
<div class="term k-harbour"><span class="sym"><i>P<sub>ab</sub></i> — policy</span><p><b>The policy matrix.</b> The distribution of vacua its civilizations <em>deliberately choose</em>. The term with no analogue anywhere in biology.</p></div>
</div>
</div>
<figcaption class="fig-caption"><span class="fig-no">FIG IV.2 — Anatomy of the reproduction kernel.</span>
Read the two terms as two theories. Smolin is the first alone; Crane raises \(B_a\) by routing it through intelligence; engineering adds the whole second term — and with it \(P_{ab}\), a mutation operator that has read the literature. Note what multiplies what: the engineered term is a <em>product</em> of four probabilities and counts, so it dies if any factor dies. A universe of brilliant civilizations that all destroy themselves has \(S_a \approx 0\) and leaves no descendants, however large \(I_a\) and \(E_a\) look.</figcaption>
</figure>
<p>Now take the limit that matters. If engineered offspring dominate — if the second term outweighs the first — the long-run distribution over vacua is governed not by which universes manufacture the most stellar black holes but by which produce civilizations that <strong>survive</strong>, <strong>understand the landscape</strong>, <strong>wish to reproduce</strong>, <strong>choose productive daughter vacua</strong>, and <strong>successfully seed their descendants</strong>. Five conditions, each cultural as much as physical. The effective replicator is therefore not a set of constants but a compound entity:</p>
<div class="eq">$$\boxed{\;\text{vacuum} \;+\; \text{civilizational phenotype} \;+\; \text{reproductive policy}\;}$$</div>
<p>Much closer to gene–culture coevolution than to blind selection. Take the comparison literally: in human evolution, language and the brain that speaks it selected on each other for a million years, neither intelligible as the unit. A vacuum and its civilization stand in the same relation: the vacuum sets what minds are possible, the minds decide what vacuum comes next.</p>
<div class="box box-physics">
<p class="box-label">Physics — when is it genuinely Lamarckian?</p>
<p>The Lamarckian upgrade is conditional on information transfer. If civilizations can seed memories, values or machinery into their babies — <a href="survival.html#baby">pre-attack seeding</a> of an inflating child universe, initial conditions written on purpose — acquired knowledge is inherited and the process is Lamarckian in the strict sense. If nothing crosses the throat, selection acts only <em>indirectly</em>, on physical laws that happen to generate universe-makers. Same equation, radically different tempo: directed inheritance improves within a generation, indirect selection waits for chemistry and civilization to be re-derived from scratch each time.</p>
</div>
<div class="box box-story">
<p class="box-label">Story hook — manufactured kinship</p>
<p>If \(P_{ab}\) is a policy, someone writes it down, teaches it, enforces it across deep time. A gardener lineage seeding successors with its own values and decision procedures is not being sentimental — it is <em>manufacturing relatedness</em>, converting strangers into kin so <a href="https://doi.org/10.1016/0022-5193(64)90038-4" target="_blank" rel="noopener">Hamilton’s rule</a> applies at cosmological scale (see <a href="game.html#cooperation">Chapter II § acausal cooperation</a>). So the real vacuum engineering is value engineering, and the most consequential document in the universe is a curriculum. Fiction that takes this seriously gets an institution: drafters of the policy matrix arguing over a clause that will decide the physics of 10<sup>40</sup> descendants — and a heretic who edits one line.</p>
</div>
</section>
<!-- ================================================================ -->
<section class="sec">
<h2 id="ecology"><span class="secnum">4.4 · Vacuum ecology</span>Competition inside one spacetime</h2>
<p>Genealogy is only half of it. Because parent-vacuum conversion happens <em>in place</em>, several vacua can compete for the same spacetime at once — an ecology, not a family tree. Track the volume \(V_a\) occupied by vacuum <em>a</em>: each phase inflates at its own rate, gains volume from phases converting into it, loses volume to phases it converts into.</p>
<div class="eq">$$\dot V_a \;=\; 3H_a V_a \;+\; \sum_b \kappa_{b\to a} V_b \;-\; \sum_b \kappa_{a\to b} V_a, \qquad \kappa_{a\to b} = \kappa^{\text{natural}}_{a\to b} + \kappa^{\text{engineered}}_{a\to b}$$</div>
<p>The first term is expansion: a high-Λ phase manufactures space exponentially fast, \(H_a\) its Hubble rate. The middle and last terms are conversion — bubbles of <em>a</em> nucleating inside <em>b</em> and vice versa. And the last symbol is the premise of this codex: the transition rate has a second contribution that is <strong>somebody’s decision</strong>. Civilizations enter the equations not as observers but as coefficients.</p>
<p>What makes this hard to reason about: several plausible definitions of “winning” are simultaneously true, and they point in different directions:</p>
<ul>
<li><strong>High-Λ vacua</strong> manufacture volume fastest — they win by sheer inflation, and they are the worst places to live.</li>
<li><strong>Lower vacua</strong> win domain-wall collisions: <a href="physics.html#rule2">deeper eats shallower</a>, so a small deep domain can chew through a large shallow one.</li>
<li><strong>Near-zero-Λ vacua</strong> support the longest-lived civilizations, and therefore the most <em>engineering</em> — a fitness route running through minds rather than volume.</li>
<li><strong>Negative-Λ vacua</strong> are terminal sinks: whatever falls in recollapses to a crunch and contributes nothing further. Volume that has no future.</li>
<li><strong>Intelligent vacua</strong> steer \(\kappa\) itself, raising transition rates toward the phases they prefer and suppressing the rest. Policy as a rate constant.</li>
</ul>
<figure class="fig">
<div class="fig-panel">
<canvas id="fig-ecology"></canvas>
<div class="fig-controls">
<span class="ctl"><label for="ec-eng">κ<sub>eng</sub> ours→harbour</label><input type="range" id="ec-eng" min="0" max="0.3" step="0.005" value="0.06" data-out="ec-eng-out"><output id="ec-eng-out"></output></span>
<span class="ctl"><label for="ec-reck">Recklessness ours→terminal</label><input type="range" id="ec-reck" min="0" max="0.25" step="0.005" value="0.02" data-out="ec-reck-out"><output id="ec-reck-out"></output></span>
<span class="chip ascent"><i></i>nursery (high Λ)</span>
<span class="chip false"><i></i>ours</span>
<span class="chip harbour"><i></i>harbour (Λ≈0)</span>
<span class="chip terminal"><i></i>terminal (Λ<0)</span>
</div>
</div>
<figcaption class="fig-caption"><span class="fig-no">FIG IV.3 — The policy matrix, made visible.</span>
Four compartments integrated forward from the volume equation, plotted as <em>share</em> of total volume: the raw volumes explode exponentially, so a stacked chart would show nothing but the nursery. Every non-terminal phase still grows without bound in absolute terms; the sliders change only the <em>mix</em>. Push κ<sub>eng</sub> up and the teal harbour band swells at the expense of ours: a civilization converting its own spacetime on purpose. Push recklessness up and the red band grows instead — the same capability, aimed badly, producing volume with no future. Note the violet floor that never disappears: in eternal inflation the high-Λ background keeps expanding whatever the bubbles inside it do, so there need be <strong>no final global victor</strong>.</figcaption>
</figure>
<p>That last point defuses a natural misreading: vacuum engineering does not select <em>the</em> vacuum of the multiverse, only shares inside a background running away from everyone. The war is not for the universe; it is for the fraction.</p>
</section>
<!-- ================================================================ -->
<section class="sec">
<h2 id="pressures"><span class="secnum">4.5 · New pressures</span>Five traits that become cosmologically selectable</h2>
<p>Here is where the framework starts paying rent. Once \(M_{ab}\) has an engineered term, properties of a universe previously irrelevant to reproduction become <em>directly</em> selectable — and several are ones we can measure. <a href="evidence.html#frame">Chapter V</a> is the audit; this is the list of what to audit.</p>
<div class="card-grid">
<div class="card c-false">
<span class="card-no">Pressure one</span><span class="card-title">Metastability with a maturity switch</span>
<p>A fit vacuum must be stable long enough for galaxies, chemistry and intelligence, yet triggerable enough that mature civilizations can reproduce it. Too unstable and it dies before growing gardeners; too stable and it is <em>reproductively sterile</em> — a perfect prison. Selection favours a narrow metastable sliver, and our measured Higgs and top masses put us in one, lifetime near 10<sup>161</sup> years. That is <a href="evidence.html#higgs">Exhibit A</a>.</p>
</div>
<div class="card c-light">
<span class="card-no">Pressure two</span><span class="card-title">Legibility</span>
<p>You cannot target a transition you cannot compute. Vacua whose inhabitants can infer the landscape — whose physics is guessable from inside, whose effective theories are separable, whose constants are measurable to enough digits — leave more engineered descendants. <strong>The comprehensibility of physics becomes a fitness trait.</strong> The unreasonable effectiveness of mathematics stops being a philosophical puzzle and becomes an inherited character.</p>
</div>
<div class="card c-harbour">
<span class="card-no">Pressure three</span><span class="card-title">Governability</span>
<p>Reproduction requires \(S_a\): surviving to capability. Universes producing civilizations that can coordinate vacuum technology outperform those whose civilizations reliably self-destruct first — and since <a href="game.html#governance">the binding constraint is internal</a>, this selects on politics. Whatever makes coordination easy at planetary scale is, here, a cosmologically heritable trait.</p>
</div>
<div class="card c-ascent">
<span class="card-no">Pressure four</span><span class="card-title">Causal compartmentalization</span>
<p>A small positive Λ separates descendants into horizon patches before they can overwrite one another — <a href="game.html#casee">causal secession</a>, enforced by relativity rather than treaty. Lineage survival goes up even though per-patch resources go down. Λ = 10<sup>−122</sup> stops looking like an embarrassment and starts looking like <a href="evidence.html#lambda">a setting</a>.</p>
</div>
<div class="card c-machine">
<span class="card-no">Pressure five</span><span class="card-title">Strategic depth</span>
<p>Since deeper vacua win collisions, civilizations install or reproduce the deepest <em>habitable</em> vacuum rather than the most fertile. Selection for defensibility: the equilibrium is not the vacuum with maximal complexity or computation but the deepest survivable non-negative floor (<a href="game.html#casec">the depth arms race</a>). Fitness rewards being unattackable, not lush.</p>
</div>
</div>
<div class="box box-physics">
<p class="box-label">Physics — heat death as an immune system</p>
<p>Pressures three and four combine into the most counterintuitive claim in this chapter. Zero Λ gives a civilization unbounded computation — and gives any one deranged descendant an unbounded domain to rewrite. Small positive Λ trades some of that future for <em>eventually partitioning its descendants beyond mutual reach</em>. If lineage-level selection operates, it favours the second: universes that make competent universe-makers <em>and then keep too many of them from sharing one future light cone</em>. Heat death is then not a tragedy in the design but a quarantine — an evolved firewall, the cosmic equivalent of an air gap. Note the shape of the argument: it predicts what ordinary anthropics predicts (small positive Λ) for a different reason, and the two can in principle be told apart by <em>how tuned</em> the value is relative to what galaxy formation alone requires.</p>
</div>
</section>
<!-- ================================================================ -->
<section class="sec">
<h2 id="lifecycle"><span class="secnum">4.6 · The life cycle</span>Crane’s jellyfish</h2>
<p>Put the pieces in a loop and something biological falls out. <a href="https://arxiv.org/abs/hep-th/9402104" target="_blank" rel="noopener">Crane’s metaphor</a> is a jellyfish, whose life alternates between a sessile polyp that buds and a free-swimming medusa that reproduces sexually — two body plans, one organism, alternating generations. The cosmological version has four stages:</p>
<ol>
<li>A high-Λ <strong>nursery</strong> vacuum rapidly generates space and daughter bubbles. Uninhabitable, enormously fertile.</li>
<li>Some of those bubbles land on a low-Λ <strong>adult</strong> vacuum, supporting long-lived intelligence. Barely fertile, enormously habitable.</li>
<li>That intelligence manufactures high-Λ <strong>baby universes</strong> behind black holes — deliberately or as a by-product of its power and propulsion industry.</li>
<li>Those babies decay into new low-Λ <strong>habitats</strong>, and the cycle closes.</li>
</ol>
<p>Reproductive and mature phases alternate, and intelligence’s role is unambiguous once you see the diagram: <strong>intelligence is the universe’s reproductive organ.</strong> Not its purpose, not its crown — its gonad. Everything a civilization thinks it is doing (exploring, computing, escaping heat death) is, at the level of the cycle, gamete production.</p>
<figure class="fig">
<div class="fig-panel">
<canvas id="fig-lifecycle"></canvas>
<div class="fig-controls">
<span class="chip ascent"><i></i>nursery · high Λ</span>
<span class="chip harbour"><i></i>adult · low Λ</span>
<span class="chip light"><i></i>manufactured black holes</span>
<span class="chip false"><i></i>new habitats</span>
</div>
</div>
<figcaption class="fig-caption"><span class="fig-no">FIG IV.4 — Alternation of generations, cosmological edition.</span>
The cycle running, particles carrying the lineage clockwise between stages. Two stages are fertile and uninhabitable, two habitable and nearly sterile; the loop closes only because intelligence bridges them. Notice that stage 3 is <em>the same technology</em> as the weapon in <a href="physics.html#rule4">Rule four</a> and the drive in <a href="survival.html#dispersal">Chapter III</a> — manufactured black holes below ~10<sup>9</sup> kg. Reproduction, propulsion and omnicide are one industry, distinguished only by where you point it.</figcaption>
</figure>
<div class="box box-story">
<p class="box-label">Story hook — the fertility cult that is also a weapons programme</p>
<p>A faction that believes stage 3 is a sacred duty builds exactly the hardware a faction that fears stage 3 must destroy — and both are right about the physics. The theological argument (are we obliged to reproduce the universe?) and the arms-control argument (can we allow anyone to hold a black-hole foundry?) are the same argument, conducted by people who don’t know they are having it. Give the Gardeners a liturgy that is also a manufacturing spec.</p>
</div>
</section>
<!-- ================================================================ -->
<section class="sec">
<h2 id="percolation"><span class="secnum">4.7 · Percolation</span>Does the war ever complete?</h2>
<p class="lede">The statistics of a spacetime filling up with bubbles are not new physics. They are metallurgy.</p>
<p>When a metal crystallizes, nuclei appear at random in the melt and grow until the grains meet. <a href="https://en.wikipedia.org/wiki/Avrami_equation" target="_blank" rel="noopener">Kolmogorov, Johnson, Mehl and Avrami</a> worked out the statistics in the 1930s, and the result transfers to cosmology essentially unchanged: the probability a point is <em>still</em> in the old phase is the exponential of minus the rate times the four-volume of its past light cone. Every bubble that could have swallowed you had to nucleate somewhere in that cone; the bigger the cone, the worse your odds.</p>
<div class="box box-toolkit">
<p class="box-label">Toolkit — KJMA survival and the Guth–Weinberg threshold</p>
<p>Survival probability of a point in the false vacuum, with nucleation rate \(\Gamma\) per unit four-volume:</p>
<div class="eq">$$P_{\text{false}}(x) \;=\; \exp\!\left[-\,\Gamma \times V_4\big(\text{past light cone of } x\big)\right]$$</div>
<p>In flat space the cone grows without bound and the exponent runs away: the transition always completes. In de Sitter space it does not. Expansion caps how much comoving volume any bubble can ever claim (the same \(c/H\) cap as <a href="physics.html#rule3">Rule three</a>, ≈ 17 Gly for our Λ), so later bubbles are comovingly <em>smaller</em>. <a href="https://inspirehep.net/literature?q=guth%20weinberg%20recovered%20slow%20first-order" target="_blank" rel="noopener">Guth and Weinberg (1983)</a> turned this into a threshold on the dimensionless rate:</p>
<div class="eq">$$\frac{\Gamma}{H^4} \;\gtrsim\; 0.1 \quad \Longrightarrow \quad \text{bubbles percolate; the transition completes}$$</div>
<p>Below it, false vacuum persists eternally: bubbles keep forming forever, and the unconverted volume between them keeps growing too. This is the technical statement behind “old inflation doesn’t end properly” — and, here, behind something stranger.</p>
</div>
<p>Because in an engineered multiverse \(\Gamma\) is not a property of the vacuum but of the <em>civilization density</em>. Nucleation happens where somebody builds a foundry, so \(\Gamma/H^4\) is essentially the number of trigger-capable civilizations per Hubble four-volume. And the statistics governing civilization spacing in <a href="https://arxiv.org/abs/2102.01522" target="_blank" rel="noopener">Hanson’s grabby-aliens model</a> transfer wholesale: hard-step emergence, expansion at a fixed fraction of lightspeed, a characteristic separation between origins.</p>
<p>Put those together and you get the sentence this chapter exists to deliver: <strong>wild vacuum survives forever because assassins are rare.</strong> Eternal inflation is not sustained by any deep principle, but by the sparseness of the civilizations capable of ending it.</p>
<figure class="fig">
<div class="fig-panel">
<canvas id="fig-percolation"></canvas>
<div class="fig-controls">
<span class="ctl"><label for="pc-lam">Γ/H⁴ (log)</label><input type="range" id="pc-lam" min="0" max="1" step="0.005" value="0.667" data-out="pc-lam-out"><output id="pc-lam-out"></output></span>
<button class="btn primary" id="pc-reset">Reset field</button>
<span class="chip false"><i></i>wild false vacuum</span>
<span class="chip harbour"><i></i>converted domains</span>
<span class="chip light"><i></i>frozen fronts</span>
</div>
</div>
<figcaption class="fig-caption"><span class="fig-no">FIG IV.5 — The assassin-rarity threshold.</span>
A comoving slice: galaxies hold still, bubbles do the moving, each bubble’s comoving radius freezing at \(c/H\) as in <a href="physics.html#rule3">FIG I.4</a>. Later bubbles start from a smaller cap and end comovingly tinier — which is why coverage can saturate below 100%. Sweep Γ/H⁴ across the Guth–Weinberg value of ≈ 0.1 and watch the verdict flip: below it, blue wild vacuum survives with room to spare and keeps inflating between the domains; above it, the domains merge and the transition completes. The surviving fraction follows KJMA, \(P = \exp[-\Gamma V_4]\), sampled from the field. The gold seams where domains touch are the <a href="physics.html#rule2">frozen fronts</a> — permanent borders in a war with no other stable boundary.</figcaption>
</figure>
<div class="box box-story">
<p class="box-label">Story hook — the polycrystal borderlands</p>
<p>Below threshold, the endgame map is not a victor’s empire but a polycrystal: ordered domains of settled physics, wild inflating gaps nobody reached, grain boundaries where two incompatible constitutions touch and neither advances. All the traffic is at the seams — <a href="fiction.html#engines">frozen-front borderlands</a> where wall-bound structures carry information between phases, the only remaining evolution is horizontal, and the descendants of enemies trade. A war that fossilized into a trading map, and a border you can stand on because both sides are equally strong. See also <a href="game.html#race">the war map</a>: the same tessellation, read forward instead of back.</p>
</div>
</section>
<!-- ================================================================ -->
<section class="sec">
<h2 id="measure"><span class="secnum">4.8 · The measure problem</span>Counting is not neutral</h2>
<p>Every claim in this chapter of the form “vacuum X dominates the long run” hides a question cosmology has not answered: <em>dominates in what count?</em> With infinite volumes, infinitely many pocket universes and infinitely many observers, ratios are undefined until you specify how to take the limit — and different, equally reasonable rules give incompatible answers. Quantum mechanics has the Born rule for weighting branches; cosmology has no agreed analogue. This is <strong>the measure problem</strong>, and <a href="https://www.dwarkesh.com/p/adam-brown" target="_blank" rel="noopener">Adam Brown</a> flags it precisely: we lack a weighting rule.</p>
<div class="box box-caveat">
<p class="box-label">Caveat — how much of this survives the measure problem?</p>
<p>The <em>local</em> physics here is measure-independent: bubbles nucleate, walls collide, deeper eats shallower, and the KJMA survival formula holds inside any causal patch. The measure problem attacks every <em>global</em> superlative — “most universes are like ours,” “engineered vacua dominate,” “selection has been running long enough to matter” — exactly the sentences a Type III framework wants to say. Read them as conditional on a measure, not as facts. The safe form of the thesis is comparative: engineering changes the reproduction kernel in specified ways; what that implies about the census is a separate, unsolved problem.</p>
</div>
<p>Now notice what the measure problem does to <em>decisions</em>, where it stops being an academic nuisance. <a href="physics.html#rule5">Rule five</a> already put infinite payoffs on the table, and with infinite payoffs how you count determines optimal play: a measure weighting physical volume rewards inflating fast; one weighting observer-moments rewards making minds; one weighting lineages rewards spreading thin. Same physics, same landscape, three incompatible strategies — all rational, none refutable from inside.</p>
<div class="box box-story">
<p class="box-label">Story hook — ideology equals measure</p>
<p>So the factions of <a href="fiction.html#factions">Chapter VI</a> are not arguing about values in any ordinary sense but about a normalization constant: their ethics are downstream of their arithmetic. A measure war is a war over how to count infinity, fought with vacuum weapons, by people who each believe the others are committing a category error — the rare fictional conflict where the theological dispute is also a mathematics seminar, and where converting an opponent to your measure genuinely changes what they will do with a foundry. Give one faction a proof, and give the proof a flaw only the reader sees.</p>
</div>
</section>
<!-- ================================================================ -->
<section class="sec">
<h2 id="ethics"><span class="secnum">4.9 · The payoff</span>The multiverse selects on decision theories</h2>
<p>Assemble the chapter. Reproduction requires \(S_a\): survival to capability. Survival requires not losing the <a href="game.html#casec">race to the bottom</a>. And the race is lost by a specific <em>kind of reasoning</em> — the kind that treats other civilizations’ choices as fixed facts rather than as correlated with its own. Causal-decision-theory racers preempt, dig deeper, mistarget, and end in crunch tilings: negative-Λ domains that recollapse, leaving no descendants and no policy matrix. Their \(E_a\) may be enormous; their \(M_{ab}\) is zero.</p>
<p>Cooperators who reason acausally — treating their own decision procedure as evidence about the procedures of civilizations they will never meet, in the sense <a href="https://longtermrisk.org/multiverse-wide-cooperation-via-correlated-decision-making/" target="_blank" rel="noopener">Oesterheld formalized</a> — converge on survivable floors, keep their industry, and reproduce. So the trait under selection, at the top of the hierarchy, is not a constant, not a chemistry, not even a culture. It is a <strong>decision theory</strong>.</p>
<div class="box box-story">
<p class="box-label">Story hook — evolution derives cosmic ethics</p>
<p>This is the moral spine, and it is unusually strong because it asserts nothing about what is good — only about what is left. The multiverse is a filter with a peculiar output: after enough generations the survivors are disproportionately lineages that cooperate without communication, keep promises no one can verify, and refuse advantages that would foreclose others’ futures — not because those things are virtuous but because the alternative leaves no descendants. Ethics as a fixed point of cosmological selection. A character can derive the golden rule from a percolation calculation and be shaken by it, a scene almost nothing else in science fiction can offer. See <a href="game.html#cooperation">Chapter II § what is left for cooperation</a>, and <a href="evidence.html#frame">Chapter V</a> for whether our universe looks like a survivor of exactly this filter.</p>
</div>
</section>
<!-- ================================================================ -->
<section class="sec sources">
<h4>Sources & further anchors</h4>
<ul>
<li><a href="https://en.wikipedia.org/wiki/Cosmological_natural_selection" target="_blank" rel="noopener">Smolin (1992, 1997)</a> — cosmological natural selection: black-hole bounces, mutated constants, fecundity as fitness; and the contested falsification tests.</li>
<li><a href="https://arxiv.org/abs/hep-th/9402104" target="_blank" rel="noopener">Crane (1994)</a> — the meduso-anthropic principle and the jellyfish life cycle; <a href="https://ui.adsabs.harvard.edu/abs/1995QJRAS..36..193H" target="_blank" rel="noopener">Harrison (1995)</a> — independent formulation with intelligence as the reproductive agent; <a href="https://arxiv.org/abs/0908.1803" target="_blank" rel="noopener">Crane & Westmoreland (2009)</a> — black-hole starships, the industrial motive for stage 3.</li>
<li><a href="https://inspirehep.net/literature?q=guth%20weinberg%20recovered%20slow%20first-order" target="_blank" rel="noopener">Guth & Weinberg (1983)</a> — percolation of true-vacuum bubbles and the Γ/H⁴ threshold for completing a transition.</li>
<li><a href="https://en.wikipedia.org/wiki/Avrami_equation" target="_blank" rel="noopener">Kolmogorov (1937), Johnson & Mehl (1939), Avrami (1939–41)</a> — KJMA crystallization statistics; survival probability as exp(−Γ × swept four-volume).</li>
<li><a href="https://arxiv.org/abs/2102.01522" target="_blank" rel="noopener">Hanson, Martin, McCarter & Paulson (2021)</a> — grabby aliens: hard-step emergence rates and civilization spacing, transferable to nucleation-site density.</li>
<li><a href="https://longtermrisk.org/multiverse-wide-cooperation-via-correlated-decision-making/" target="_blank" rel="noopener">Oesterheld (2017)</a> — evidential cooperation in large worlds: acausal cooperation between civilizations that never meet; <a href="https://en.wikipedia.org/wiki/Superrationality" target="_blank" rel="noopener">Hofstadter’s superrationality</a> is the ancestor of the idea.</li>
<li><a href="https://www.dwarkesh.com/p/adam-brown" target="_blank" rel="noopener">Adam Brown on the Dwarkesh Podcast (2024)</a> — engineered vacuum transitions; the missing multiverse measure and the absence of a Born-rule analogue.</li>
<li><a href="https://arxiv.org/abs/1205.6497" target="_blank" rel="noopener">Degrassi <em>et al.</em> (2012)</a>, <a href="https://arxiv.org/abs/1307.3536" target="_blank" rel="noopener">Buttazzo <em>et al.</em> (2013)</a> — near-critical Standard Model metastability, lifetime ~10<sup>161</sup> yr: the measured input to Pressure one.</li>
<li><a href="https://inspirehep.net/literature?q=farhi%20guth%20guven%20obstacle" target="_blank" rel="noopener">Farhi, Guth & Guven (1990)</a>; <a href="https://arxiv.org/abs/1512.01819" target="_blank" rel="noopener">Garriga, Vilenkin & Zhang (2016)</a> — baby-universe production behind black holes, the mechanism stages 3–4 of the life cycle assume.</li>
</ul>
</section>
<nav class="chapter-nav">
<a class="card c-light" href="survival.html"><span class="dir">← Prev · Chapter III</span><span class="card-title">Surviving the Wall</span></a>
<a class="card c-ascent next" href="evidence.html"><span class="dir">Next · Chapter V →</span><span class="card-title">The Suspicious Universe</span></a>
</nav>
</main>
</div>
<footer class="site-footer">
<p>THE VACUUM GAME · CHAPTER IV — EVOLUTIONARY COSMOLOGY</p>
</footer>
</body>
</html>