docs(mental-models): add ch-07 Symbols & Connectionism

[PanQiWei]

Summary

• Add ch-07 (符号与联结 / Symbols & Connectionism) to the mental-models series — 6 articles + index, bilingual (ZH + EN independently written)
• Core thesis: symbolic and connectionist representations are irreducible; the harness is the translation layer between them; translation has structural costs distributed along a tension axis (constraint strictness vs. generative freedom)
• Ingest ~20 new wiki pages (Kautz, Fodor & Pylyshyn, Newell & Simon, Park et al., Gurnee & Tegmark, Othello-GPT cluster, RANLP trajectory bias, OpenAI/Anthropic structured outputs)
• Generate 50 new/refreshed wiki infograph cards (28 missing concepts + 12 missing entities + 10 stale)

Articles

#	ZH	EN	Core
01	两条路	Two Roads	Irreducible capabilities of symbols vs neural nets
02	奇怪的中间物	The Strange Hybrid	LLM's representational duality
03	翻译层	The Translation Layer	Naming the harness as neuro-symbolic bridge
04	翻译的代价	The Cost of Translation	Five structural impedance mismatches
05	张力轴	The Tension Axis	Constraint strictness vs generative freedom
06	符号与联结之道	The Dao	Resonance with cybernetics, entropy, fractals, causality

Review & verification

• Opus editorial review (writing rules, claim verification, cross-document consistency)
• 6 independent fact-check agents (one per article, web-search verified)
• All critical/important issues fixed (Deep Blue, CFG vs FSM, author attributions, source separation, OWASP reframing)
• ZH + EN strict build: 0 warnings

Test plan

• uv run mkdocs build --strict -f mkdocs.zh.yml — passes
• uv run mkdocs build --strict -f mkdocs.en.yml — passes
• Visual review of article cards in browser
• Cross-chapter link verification in rendered site

🤖 Generated with Claude Code