[hyperthesis]

The two instances of "instead of" have opposing senses. The grammar can't be interpreted without the semantics. Anyway, it seems other people find this OK, which is interesting to me.

Below is the context in TFA, and the quote's form with show/tell placeholders:

  she informed one writer that their “story is certainly worth telling,” but they
  "[tell] instead of [show], [show] instead of [tell]."

I can read the second part as a directive (I'd use a semi-colon instead of a comma to show this myself). I bet the original letter's connection (instead of "but they" in TFA) supports this.

It feeds into my nascent theory that grammar is not necessary for everyday human communication, and its function is ornamental, like patterns on clothing, signifying social position. (Of course, it is useful in precise technical exposition and argument; though there's still greater precision in numerical quantification and first order logic.)

Grammar is the greatest joy in life, don't you find?

[soneca]

I don’t know English grammar, but I don’t think inverting the position of the show/tell would be against Brazilian Portuguese grammar.

I think a grammar is certainly necessary for everyday communication though. Not the academically decided normative grammar for a whole language. But every verbal communication follows grammatical rules, most of the time different from the written grammar and normative grammar, but rather a grammar agreed among the people that form a group that speaks that local grammar.

[hyperthesis]

My grammatical issue is not the swapping itself, but the introduction of "but they" (indicating a problem with what the author has done) being applied to the swapping. That makes sense for the first part, tell not show, but not for the second part, show not tell (which is what the author should do).

Is grammar necessary? The original "lingua franca", an English pidgin, was a trade language around the Mediterranean, therefore economically very advantageous, yet has minimal grammar.

People can communicate with gestures.

One does need a way to qualify some specific part, e.g. if we negate something, what exactly are we negating? Maybe I have too restricted a definition of "grammar", but I think of the Chomsky Hierarchy, which are all sequential, i.e., with order dependence. I think the target of negation can be indicated without position, e.g. so "not tree" and "tree not" are not a sequential grammar (a commutative grammar?) Negation of a part could also be denoted with inflection, or gestures.

One could argue all this is a kind of grammar; but it's so far from what we normally call grammar that I think it's overstretching the concept.