Questions for September 2023

Inspired by similar pages from Alex Guzey, Gwern, Patrick Collison, & Suspended Reason, here’s a list of some random open questions (as of September 2023).

These are questions/puzzles for which I haven’t yet found (sufficiently clear) explanations or formulations.

They’ve been bothering me lately.

1. Narrative Voice

When considering narrative voice & point of view, how have tastes changed in the consumption of popular prose fiction over time?

  • Some extremely vague impressions that I’d like to test:
    • The most read novels of the 2020s are much more likely to be narrated from a first person POV than those of the 1920s.
    • Second-person narration has always been mostly artsy & weird.
    • 3rd person omniscience is in decline, but is mostly being replaced by more limited modes (not by first person narration).
    • Free indirect speech is in decline relative to a mid-1900s peak, and was never as important for mass-market readers as it was/is for the highbrow lit crit set.
      • (Same for stream of conciousness? I’m less sure.)
  • A basic start here would be:
    1. consider some full-text corpus of novels published in the anglophone world over the last ~200 years;
    2. construct a reception dataset (eg borrowing records from libraries);
    3. measure narrative perspective in the novels;
    4. weight by borrowing by year.
  • Recent work on measuring the passage of time in fiction seems encouraging.
  • I’m only interested in reading communities, not texts by year of publication.

2. Hedgerow Biodiversity

Hedgerow biodiversity in the United Kingdom is a big deal. Is the apparent density of species actually unusually high, or is this a selection effect?

  • Or, something like: “How much of this biodiversity is ‘real’; how much is a product of the cultural importance of hedgerows — which is huge! — leading to greater attention paid by scientists, and hence a mistaken impression that hedgerows are ecologically vital?”

3. The Moche

What was the deal with the Moche civilisation?

  • ~100-700 AD, northern Peru.
  • As far as I can tell, the material evidence points pretty definitively towards religious practices that were extremely human-sacrifice-centric, and also towards clear social stratification.
  • It seems like there’s also some disagreement about the details of the human sacrifice.
  • Big impressive technologies include:
    • Mass-produced mold-made ceramics;
    • Surface enrichment depletion gilding;
    • Electroplating. (!!)
  • Specific questions:
    • How much of an outlier is this mode of religious and social organisation?
    • What led to the prevalence of human sacrifice among the Moche (and some other empires, cultures, & civilisations in South America)?
    • Should I think of the combination of ‘extreme ritual violence’ (probably religious) and ‘extreme technological/craft progress’ (in at least some domains) as making the Moche case extra unusual?
  • The “Decapitator” God seems pretty dope, ngl.

4. Moral ‘Realist’ In What Sense?

When philosophers claim to be moral realists, in what sense are they using the word ‘realist’?

  • The 2020 PhilPapers Survey shows ~62.07% of respondents “Accept or lean towards” moral realism.
  • I’ve had various conversations with people — ranging from antirealist neopragmatists with a sociological bent to self-proclaimed ‘hardcore’ moral realists who endorse hedonic utilitarianism — and I’m still unable to fully pin down the view.
  • Specifics:
    • Why don’t deflationary approaches to ontology also generate a pressure towards moral antirealism?
    • If one accepts basic naturalist premises — which most philosophers seem to — why would we expect our ‘morality’ concepts to track a subject-independent ‘real’ thing (‘The Good’)?

5. Bushtucker Commercialisation

If the Macadamia Nut, why not the Kakadu Plum or the Warrigal Green?

  • Native to the East Coast of Australia, commercial cultivation of the macadamia nut by white invaders began in the 1880s, and the species was introduced to Hawaii for this purpose.
  • It’s the most expensive nut in the world — it’s slow to grow & hard to harvest — but it sells well and is a commercial success.
  • Compare to some other Australian edible plants which are only now even in Australia getting a limited market, despite being easier to cultivate:
    • Warrigal greens (Tetragonia tetragonioides) are a hardy halophyte.
    • Tasmanian mountain pepperberries (Tasmannia lanceolata) make for a fucking great spice.
    • Kakadu plums (Terminalia ferdinandiana) seem like an obvious candidate for ‘cool superfood’.
    • Wild oranges (Capparis mitchellii) are, despite the name, related to capers, and are extremely tasty.
    • Pencil yams (Vigna lanceolata) are fantastic, year-round desert yams.
  • …and so many other edible & medicinal Australian plants! Why such a lacklustre commercialisation & cultivation story?

6. FP + REPLs + ADHD

How much is programmer cognition impacted by a REPL and tightness of feedback loop? How does this vary with ‘neurodiversity’?

  • Anecdotally, when I talk to colleagues, some report that the combination of a functional programming paradigm and sensible tooling makes them feel substantially ‘less anxious’ or ‘more able to manage [their] attention’.
    • This is particularly noticeable with everyone’s favourite modern lisp dialect (Clojure).
    • The colleagues who tell me this are almost always ‘neurodiverse’.
  • Personally, it feels like there’s a class of thinking that I can ‘only’ have when tightly entangled with a REPL interface.
    • This does feel intuitively related to ADHD & Autism qualities in my thought.
  • (Standard EMT & Embodied Cognition citations apply.)
  • (Also: why are all human-computer interaction studies so artificial and crap?)

7. Do Yetis Have Society?

Are elves the only example from folklore of a supernatural being/race ‘with their own society’?

  • Is my factoring of ‘supernatural beings in folklore’ (as a category) bad?
  • Is ‘supernatural creatures with their own civilisations’ a distinctively Western European tradition?

8. Optical Illusions As Adversarials

At what level of granularity does this analogy between optical illusions (in humans) and adversarial examples (in ML) break down?

9. No Lichen

Imagine all lichen magically disappeared from the planet instantly. What would happen, in what order, and how long could humans survive?

  • Lichens are composite organisms, formed by a mutualistic relationship between algae or cyanobacteria and filaments of fungi species.
  • The role of lichen is pretty wild.
  • 6-8% of the land surface of the Earth is covered by lichens, and the nutrient cycling function is pretty indispensable.
  • What’s the state of the art for modelling trophic cascade, and how far can I push into the tails?

10. Saladin in Jerusalem

As stylised metaphor, how true is it to claim: “when Richard the Lionheart took Jerusalem, he washed the streets with blood; when Saladin took Jerusalem, he washed the streets with rose water”?

11. Curing Aphantasia

Can we cure aphantasia? If so, what would that tell us about hyperfantasia?

12. Modelling ‘ex-situ in-toto’

When attempting to conserve endangered species for which there is only a single remaining population, a ‘last resort’ is always to take that entire population into captivity (‘ex-situ, in-toto’). Sometimes this works, sometimes it’s a catastrophe. When is it a good idea?

  • Browne et al model this ‘trigger-point’ problem as an optimal stopping problem, which differs “substantially”, as they say, from the “adaptive management methods” more common to conservation biology.
    • This method could be modified to include a species-specific population viability model.

13. Lateralization Updates

How deep does the lateralization of brain function rabbit hole go?

  • Does my strong lefthandedness predict outcomes for some mental health treatments?
  • If I ever need trauma treatment, should I avoid EMDR?
  • Some people I know vaguely think this shit is important & non-trivial & surprising.
  • See, eg

14. Quaker Decline

What happened to the Quakers?

  • I know they are still around.
    • (I have met and am related to many.)
  • I mean: why are they less effective than they used to be?

15. Does Tetris Treat Trauma?

Some work indicates that playing Tetris soon after experiencing a traumatic event functions as a kind of cognitive ‘therapeutic vaccine’, reducing the incidence of flashbacks etc. See:

If so, this ought to generalise, as it indicates useful things about memory formation (and trauma).

  • Can we engineer some kind of optimised Therapeutic Super Tetris?
  • Someone let me A/B test interfaces (at an extremely high tempo) on trauma patients plz.

16. TurnItIn, AI … Perplexity?

If they don’t scale back, will TurnItIn’s grandiose claims about their tool’s ability to detect AI generated assignment submissions cause students to submit more interesting prose over the next 1-3 years? Would that be good for all involved?

  • Despite my (and others’) attempts to explain otherwise, many colleagues outside of CS/AI continue to believe the “TurnItIn can kinda-sorta detect generated content” line.

17. BJJ’s Evolving Meta

Among the serious BJJ practitioners I know, many describe an evolving meta that vacillates between ‘invention of a genuinely new technique’ & ‘high-precision execution of fundamental techniques’. Is this real?

  • If so, my intuition is that this indicates BJJ has not yet fully developed a canon of techniques at the frontier (after which I would expect the meta to evolve principally w/r/t precision in execution and sequencing of those techniques).

18. Optimal Octopus Breeding

If I wanted to begin an octopus intelligence breeding program in order to gain information about alignment of & communication with ‘alien’ and largely asocial agents, what would be the best species?

  • Related: I have some experience with dingoes & ‘camp dogs’ (with mixed-dingo genetics), and much more limited experience with alpine dingoes. In every case, the apparent intelligence of these dogs was remarkable compared to every other non-human animal I’ve interacted with. Would extended interaction with an octopus change my mind about this?

19. Steelmanning Representationalism

Does ‘representationalism’ have a coherent defence?

  • I take representationalism to be the position that ‘models’ in science are meaningfully (if imperfectly) analogous to scientific instruments such as microscopes, telescopes, etc.
  • That is, models
    1. enable abstraction and generalisation over
    2. neutral observations of
    3. an already-given external world by
    4. a separable (human) observer.
  • Many low-resolution/naive accounts of scientific modelling seem to assume some version of representationalism to be true.
  • Basically no serious contemporary work in philosophy of science, computer science, applied mathematics, or science & technology studies endorses this view.
    • That is: anglophone ‘analytic’ philosophers don’t believe it; serious abstraction-manipulators don’t believe it; Continental-inflected sociologists and soft humanities folks don’t believe it.
      • (Remarkable levels of agreement, frankly.)
  • Where is the Steel Man and can I meet him?

(If you think you can answer a question or dissolve my confusion, please contact me.)