Parsimony as a side dish—a game to play on meetups?
Tl;dr: we seem to (naively?) compound parsimony with other heuristics.
There was a quote from R. Burton, provided by JQuinton, offering an amusing exercise. The reader had to guess if an excerpt of text was nonsense or merely a rambling presentation of something, and what that could be. Upon learning a single word of explanation, it became difficult to read it as anything than the object’s description, even if the reader was told there were other possible answers (beyond ‘ridiculous.’)
Here’s an example:
> Reconstruction probably won’t even account for all that escaped the initial search. Of course it doesn’t have as much lead as the stained variety. Please don’t follow ‘old traditions’ again, it’s a cultural thing anyway. The little ones are easily lost in water. The dog won’t know how to find them, but the problem is rather that it won’t know how to avoid them. The high singing note I will miss. Be careful not to step on it. It was to harmony what it is to ruin. A different brand is used in electron microscopy, to make ’em smooth and even. There’s plenty under the bench in the park. I had cherished it since my wedding.
(I’m not a writer, so that was rather clumsy. Please post better examples in comments.)
But what if we are offered to brainstorm before answering, and to try viewing the excerpt as a collection of true facts, just not necessarily a coherent story? Here are some of our possible approaches (heuristics):
- it’s a picture;
- it’s an instruction;
-it’s paraphyletic (e.g., it is about a single thing which has more than one cause);
- it’s alive!
- it’s dangerous!
- it’s an advertisement;
- it’s a single thing better described by more than one word, though still recognizable from the best match;
- it’s a compilation of distant phenomena related to whatsitname, the everyday thing;
- it’s all true, BUT there are qualifiers (which could make it more plausible-sounding);
- it’s just not a physical body;
- it’s an extreme case etc.
...and then we try to guess again.
The number of hypotheses now should be more than one, but why? What changed?
I think we start out with expectations nearest to ‘picture’ and most removed from ‘qualifiers’, and maybe it was useful in ancestral environment, but I would not expect them to be most fruitful. Maybe if enough people played such scenarios out, we would experimentally obtain a set of more useful ones?
Is there a way to prove that one and only one interpretation is true (allows for all statements to be true)? What would you expect such excerpts (sets?) have in common?
And more importantly: is there any way to weaken priming’s hold on us?
(I am not a native speaker, so if there are any mistakes, please point them out to me. Thank you.)
Since you’re talking about a game most people haven’t played before, presenting a concrete example is important to do early. I recommend the mystery passage should (a) be blockquoted so it stands out more and (b) moved to before you start talking about heuristics.
Done, thank you.
Reminds me of a game called Ein solches Ding… i.e. “such a thing...” which consisted of lots of cards specifying some rather vague property like “weights more than a barrel”, “has feather, fur or skin”, “is used by musicians”, “has handle, hole or key” and so on. This accompanied by illustrations ranging from creative to bizarre always involving a question-mark. The goal if to think of a thing matching all the requirements so far and to add one further criteria—until someone calls it a bluff upon which a matching real thing must be named.
Like Ghost, but for things instead of words!
It’s a bit more complicated, but yes.
The paragraph seems to be mostly about this, although there were a few lines that don’t quite seem to make sense in context...
Well, you see, there’s been an accident:) I would happily rephrase it to make it more relevant. It’s rather like constructing a Bongard problem, when you might want to hint at’other explanations’, only in a BP you know there must not be any, and here they are actually advisable.