The basic idea of "embedded commands", for those who haven't heard of them, is that you embed commands in ordinary language so that people will subconsciously pick up on the command and do what you tell them to without consciously being aware of it (and therefore "unable to resist"). For example, you might say "that's cool. I wouldn't expect you to make an exception for me" while subtly emphasizing the command "make an exception for me". And if it "works", they'll make an exception for you even though they wouldn't have if you simply told them "make an exception for me".
No, it's not nearly as powerful or mind-controlley as people
looking into covert hypnosis hope, but yes, there's
something to it. In fact, like always, it's something that fits
naturally into how we use language.
If you're sitting at the table you might say "could you pass the milk?",
for example. Super sneaky mind control, eh? You asked a question, and
instead of answering it, they succumbed to your embedded command!
Well, except that it's not sneaky. Yes, there's a command there,
and yes they followed it. And yes, they might not have if you just said
"pass the milk". But not because they have anything against passing the
milk - just against the presupposition that you get to boss them around.
Niceties like "please" and "could you?" still communicate the same
command, but they also change the frame to acknowledge the favor making
them a lot more willing to do it. If they catch you talking about "hah, I
so tricked my brother into passing the milk! He had no idea I
mind-controlled him!", then that frame is gone and guess what? It won't
work next time.
The "wrapping" around the command provides context in which the
command may make sense. Wrapping it with "you might want to ____" gives
them information that you think they'd get something out of the command
as opposed to doing it because you said so and you're the boss. "you
don't have to ____" lets them know that you won't pressure them into
doing it - but still implies that there's some reason for them to do it
and brings the idea to mind.
And that's all it is, really - bringing the idea to mind without
being forceful and turn-offish about it by supplying appropriate
context. It doesn't provide any significant impetus to actually do what
you asked for. They'll do it if it feels appealing in context, and they
won't if it don't. No amount of stealthiness can make them want to any
more. All you can hope for with embedded commands is to not shoot
yourself in the foot.
To prove this, let's look at the results of a trick I came up
with. It's a trick to give suggestions and cover it with amnesia without
getting any real hypnotic responsiveness to mess with things. It makes
the perfect test ground for things like this. Basically, I'd suggest
that when I scratch my nose, they'd laugh, but that they wouldn't know
why because they'd forget that I ever told them any of this as soon as I
snap my fingers. I'd make sure they were actually listening and not
pre-rejecting the idea, but neither building up any other motives or
detailed frames to encourage them to actually do it. Then I'd snap my
fingers and distract them for long enough for my suggestion for them to
not remember take effect. And guess what happened? They'd laugh when I
scratched my nose. They wouldn't know why.
But... they wouldn't laugh stupidly hard, even if that's
what I suggested. They'd only laugh within the boundaries that that
context allows. They're kinda curious about the mind trick, so I have
some compliance. They can explain some laughter as something goofy or
another, so they'll allow themselves to laugh a bit, but they can't
explain and have no reason to laugh stupidly hard, so it won't happen.
Doesn't happen. If you're sufficiently sneaky seamless/context-fitting with your embedded suggestions, you still hit this limit.
Neat to know about, I guess, but scary mind control it is not.