A subst-antial question

I've used [subst] before, and thought I understood it, but today it
surprised me. I guess subconsciously I was expecting it to work like
the parser, preserving word boundaries, but it doesn't. Here's a toy
example:

% package require comm
4.3
% set pairs {a 1 b 2 c 3}
a 1 b 2 c 3

Now I can create a local array:

% array set foo $pairs
% parray foo
foo(a) = 1
foo(b) = 2
foo(c) = 3

Here's a naive attempt to create a remote array using the same code:

% ::comm::comm send 51395 {array set foo $pairs}
can't read "pairs": no such variable

That's expected, since "pairs" isn't defined in the remote interpreter.
We have to substitute it here, and evaluate the results there. Here's
another naive attempt:

% ::comm::comm send 51395 array set foo $pairs
wrong # args: should be "array set arrayName list"

That's also expected, because of the implicit [concat] behavior of
[comm send]. How 'bout keeping the code in braces (to concat-proof it)
and substituting it?

% ::comm::comm send 51395 [subst {array set foo $pairs}]
wrong # args: should be "array set arrayName list"

Hmm... what's going on?

% subst {array set foo $pairs}
array set foo a 1 b 2 c 3

I was expecting a list of four words, which is what the parser
generates from the same expression. I can think of ways around this,
but am I missing something simple?

.