We can barely think about abstractions (time, argument, ideas) directly; we can only borrow one concrete domain to understand an abstract one. "ARGUMENT IS WAR" — so we "attack" positions, "defend" claims, "demolish" the other side; arguments have winners and losers. "TIME IS MONEY" — so time can be "spent," "saved," "wasted." This mapping from a concrete source domain to an abstract target domain is systematic yet invisible: say casually "our relationship has reached a crossroads," and the whole "LOVE IS A JOURNEY" metaphor is quietly running underneath.
Non-trivial: (1) The metaphor isn't in the words, it's in the thought. "ARGUMENT IS WAR" isn't a few figurative terms but an entire treat-the-other-as-enemy cognitive stance — it pre-decides what you can see and do. (2) The most powerful consequence: change the metaphor and you change the whole space of available actions. Imagine "ARGUMENT IS DANCE" — the goal becomes jointly completing a graceful dance; no winners, only how well you coordinate. Same situation, every possibility transformed. (3) This is core evidence for embodied cognition: abstract thought is not free-floating logic but rooted in sensorimotor experience — "important is big," "intimate is close," "moral is clean," all grown from infant bodily experience. The mind is not a disembodied symbol processor but a body shaped over and over by metaphor.
Practical: when something stalls you, first ask "which metaphor am I thinking about it with?" — surfacing the metaphor is what earns you the right to swap it.
"ARGUMENT IS WAR" vs "ARGUMENT IS DANCE." The first makes you want to win and overpower; the second makes you want to dance together and coordinate. Language is only the surface — what's really swapped is your perception of the whole thing and the actions available to you: one conversation, two utterly different sets of possibilities.
(1) AI / cognitive science: "the mind is a computer" is a metaphor that ruled for half a century — it makes "memory," "storage," "bandwidth," "running a program" feel self-evident, while quietly erasing emotion, body, and embodiment. Switch to "the brain is a prediction machine" or "the mind is an ecosystem" and the research directions shift instantly. (2) Every technical abstraction you make is a metaphor choice — "neural network," "attention," "training" — each illuminating and concealing. Noticing which metaphor you're thinking with is a meta-skill of the AI super-individual.
A frame is the default structure for understanding something. The real battlefield is usually not the facts but which frame becomes default. "Tax relief" buries an entire frame in two words — tax is an affliction, relief is rescue, the taxer a villain. Before you've made any argument, your stance is pre-decided by the term. A framing war is the fight over which frame gets to name the battlefield.
Non-trivial: (1) The most counter-intuitive law: negating a frame activates it. Tell someone "don't think of an elephant" and an elephant appears in their head. So when you rebut with the opponent's term ("I'm against so-called tax relief"), every repetition exposes and reinforces their frame. (2) The real victory isn't winning the debate inside their frame but making your own frame the unthinking default — the same money called a "membership fee" or a "fine," the same tax called "estate tax" or "death tax." The war ends at the naming stage. (3) Same root as anchoring, but deeper: anchoring shifts a number, framing shifts the whole problem space. (4) Be honest about the boundary: a frame is a neutral tool, usable to reveal or to manipulate. Spotting framing wars is first of all a defense — seeing which frame is doing your thinking for you.
Practical: when a word makes you instantly take a side, don't rush to agree or rebut — ask "what does this word treat as a given?" Unpack the naming first, then decide whether to accept the battlefield.
The same tax is "estate tax" to one side and "death tax" to the other; on the same issue, the two sides call themselves "pro-life" and "pro-choice." Neither will fight under the other's name — because accepting the other's word is accepting their entire worldview, and the stance is half-lost before a single argument.
(1) Technical narratives: the same AI technology framed as "AI replaces humans" or "AI augments humans into super-individuals" steers regulation, investment, and adoption in opposite directions — the narrative war is a framing war. (2) Engineering review: framing "eventual consistency" as "weak consistency" (sounds like a defect) or "highly available and scalable" (sounds like a strength) often produces opposite verdicts on identical underlying mechanics. See through this and you can drag the debate back from "whose word sounds nicer" to "what's the actual underlying tradeoff."
Reframing is the constructive use of framing: facing a stuck situation, don't try harder inside the old frame — change the frame organizing the problem so that new options appear. The classic "slow elevator problem": tenants complain the elevator is too slow, engineers price every speed-up and all are too expensive; then someone reframes "the elevator is too slow" into "the wait is too boring" and installs mirrors by the doors — the complaints vanish. The elevator never got faster; the problem changed.
Non-trivial: (1) The frame determines the entire solution space. Under "speed it up," mirrors aren't even a candidate; reframe and they become optimal. That's why reframing is the highest-leverage move — it doesn't search within the solutions, it shifts the space the solutions live in (cf. leverage points, Day 3: the highest leverage is changing the paradigm itself). (2) It's not spin or self-deception: reframing requires the same facts to still hold under the new frame, only reorganized. In negotiation, reframing "a clash of positions" into "an alignment of interests," or "split the pie" into "grow the pie," releases real options, not rhetoric. (3) Name the shadow side: reframing that slides into denying the real problem degenerates into toxic positivity — framing "the system actually crashed" as "a growth opportunity." The discipline: reframe to open options, not to evade reality.
Mirrors for the slow elevator. Reframe "an engineering speed problem" into "a psychological waiting problem" and a near-zero-cost solution appears instantly — one that simply didn't exist in the original frame. Change the problem, and the answer goes from nothing to obvious.
(1) Engineering: instead of asking "why is the system so slow" (endless performance tuning), reframe as "which SLO is unmet, and for whom" — you often find you needn't speed up the whole thing at all. Reframing "the model isn't accurate enough" as "the task boundary is drawn wrong" often beats labeling another hundred thousand examples. (2) Buddhism: the "turning of thought" and prajñā wisdom you're drawn to are reframing at its core — the same circumstance, viewed through a different frame, and the entire psychological solution space changes. Reframe impermanence from "threat" to "the inherent flow of things," and the structure of suffering loosens.
Metaphors have a life cycle: vivid at first, everyone feels them as figures of speech; used long enough they "die" into literal meaning — "the leg of a table," "the foot of a mountain," "grasp an idea," "the heart of the matter." No one feels these as figurative anymore. A dead metaphor still carries its original mapping; the mapping has merely gone invisible and is taken as reality itself.
Non-trivial: (1) Dead metaphors are far more dangerous than live ones: because they no longer look like metaphors, they escape scrutiny. "Time is money" is so thoroughly dead that we no longer notice it shaping our whole relationship to time — idleness as "waste," life as a "ledger." (2) The cure is to revive the dead metaphor: treat it as a figure of speech again and its smuggled assumptions resurface, open to challenge (same root as "the map is not the territory," Day 43 — a dead metaphor is a map mistaken for the territory itself). (3) Nominalization is the worst zone: freezing a process into a thing — "self," "mind," "memory" are dynamic processes that language solidifies into something that seems to sit there, prompting grammar-manufactured pseudo-questions like "where is my true self?"
Practical: periodically audit the words you use most and most unthinkingly — the more a word feels like "that's just what it means," the more likely it's a dead metaphor presetting the world for you.
"The leg of a table," "a mountain ridge," "time is flowing" — completely dead, no one reads them as figures. "The invisible hand of the market" is half-dead: it still quietly personifies the market into an intentional coordinator; revive it and ask "whose hand, and by what right does it coordinate?" and a lot of intuitions about markets start to loosen.
(1) AI: "training" a model is a dead metaphor that makes RLHF sound like animal taming or teaching, hiding that underneath it's just gradient descent fitting a statistical surface in high dimensions. Revive the metaphor and your intuitions about "alignment" and "will the model deceive" shift wholesale — you're not disciplining a student, you're sculpting a surface. Tech jargon is almost all dead metaphors — "cloud," "memory," "virus," "firewall," "neural network" — each quietly presetting your assumptions. (2) Consciousness / Buddhism: the "no-self" (anatta) you're drawn to is precisely the reviving of the biggest dead metaphor — "the self." Restore the noun "I" to its underlying verb-like process (a ceaselessly arising-and-passing stream of awareness) and the question "who am I really?" reopens.