Comment by 4cao | Hacker News Reader

Comment by 4cao | Hacker News Reader

> Jack is looking at Anne, and Anne is looking at George. Jack is married, and George is unmarried. Is a married person looking at an unmarried person? (Yes/No/Indeterminate)

The answer to this is not so obvious to me. What if Anne is divorced or or widowed? Unmarried can mean "not married," but it also could mean "never married," as in the following example:

> Crude incidence rates of stroke (per 1000 per year) among the four marital status categories were as follows: 1.96 (married), 1.52 (unmarried), 2.36 (divorced), and 5.43 (widowed). Compared to married persons, adjusted risk of stroke was significantly increased for divorced (RR 1.23; CI 1.19-1.27) and unmarried men (RR 1.07; CI 1.03-1.11) but not for widowed men (RR 1.02; CI 0.98-1.06); risk was slightly increased for divorced women (RR 1.10; CI 1.06-1.15) while not for widowed (RR 1.0; CI 0.97-1.03) and unmarried women (RR 0.97; CI 0.97-1.03).

https://pubmed.ncbi.nlm.nih.gov/29492956/

Ok? I feel like it's relatively clear from the question that in this hypothetical world there are only 2 states, married or unmarried.

Logically the question can be cut down to:

Given an alphabet {A, B} is there an symbol x such that for the string "AxB" an A is not followed by a B.

It's also just "not a question", but really a puzzle. And I think that's what's happening in "discourse-decoupling" also. I do think they might well just be the same.

Ie., in ordinary conversation when one says, roughly, "heres some information, what do you think?" one doesnt expect the question itself to be a puzzle -- ie., that the answer is in some formal sense about the information as information.

I think likewise, in ordinary conversation, "shall we just kill all the ants?" means kill all the ants. It is a very unusual (academic) conversation in which the question is read as a sort of "ethical puzzle".

If you just told people in both cases, "this question is a puzzle" -- then I imagine it would trigger the decoupling mechanism more reliably.

I admit I failed that question, and I understand the impulse to then challenge the way the question was constructed. In school, it's something I did regularly, partly because some of the questions were indeed weakly constructed, but also partly because I was reluctant to admit I actually misunderstood.

In this case, it's not a trick question per se, but it certainly does something to trick the mind into not enumerating the possibilities. It would be interesting to see how the majority of programmers respond to it, as compared to say the majority of mathematicians. I suspect the question tricks people into abstracting too early: 'one variable in this system is undefined, hence I can't say!'

Reply on news.ycombinator.com