Reading & Influences
The papers behind the papers — what I have been reading, and the logic I took from each.
Attention & the Geometry of Focus
Five classic theories each captured one phenomenon and proposed one mechanism. Reading them side by side raised the question my MFA preprint tries to answer: what if they are five views of a single field?
Posner (1980), Orienting of Attention — Covert attention moves while the eyes stay still: valid spatial cues buy speed, invalid ones cost it. Attention behaves like something aimable — the "spotlight" was born here.
Eriksen & St. James (1986), the zoom-lens model — The beam has adjustable width, and resolution trades off against size: widen the field and processing thins out. A fixed resource, variably spread.
Downing (1988), expectancy and visual-spatial attention — Performance declines smoothly with distance from the attended location. Not a hard-edged beam: a gradient.
Lavie (1995), perceptual load theory — Whether distractors get processed depends on how loaded the task is; spare capacity spills outward automatically. Selection is a budget, not a gate.
Amari (1977), dynamics of pattern formation in neural fields — Local excitation plus surround inhibition yields stable bumps of activity on cortex. The mathematics that attention models stand on.
Itti, Koch & Niebur (1998), a saliency-based search model — Bottom-up conspicuity computed as feature contrasts merged into one master map. The ancestor of every entry on today's saliency benchmarks, including mine.
Kümmerer et al., the MIT/Tübingen Saliency Benchmark — How the field scores attention prediction (AUC, NSS, IG) and why the choice of metric reorders the leaderboard. Where my training-free model is evaluated.
Inner Voices & the Dialogical Mind
Vygotsky's social origins of thought, Fernyhough's expanded/condensed continuum, Hermans' society of I-positions, Taylor's imaginary companions, and the voice-hearing continuum — five literatures that converge on the DIVP paper.
The full annotated trail — every paper, and what each one contributes — lives in the interactive framework map.
Mind, Memory & Prediction
How minds model the world — and why remembering, perceiving, and imagining might be the same machinery pointed in different directions.
Clark (2013), Whatever Next? — Brains as hierarchical prediction machines: perception is expectation corrected by error, top-down all the way through.
Friston (2010), the free-energy principle — One quantity, surprisal under a generative model, that perception and action both work to minimize. Ambitious, contested, unforgettable.
Schacter & Addis (2007), constructive episodic simulation — Memory is reconstructive because its job is not archiving the past but simulating futures. Remembering and imagining share parts.
Language & Meaning
Where does meaning come from, if nobody is in charge of it?
Saussure (1916), Course in General Linguistics — Signs are arbitrary; a word's value comes from its differences within the system, not from pointing at the world.
Firth (1957) — "You shall know a word by the company it keeps." Meaning as distribution, stated decades before anyone could compute it.
Lewis (1969), Convention — Language as the stable equilibrium of a coordination game: meanings hold because everyone expects everyone else to keep using them. No legislator required.
Mikolov et al. (2013), word2vec — The distributional hypothesis cashed out as geometry: meaning becomes position in a vector space, and analogy becomes arithmetic.