DM Howard
Explorer
Gygax-style AD&D (as presented in the rulebooks, and as discussed in the RPG magazines of the period) is based around dungeon exploration.
I certainly can't contend otherwise on this point!

So, let me concede this debate from a historical standpoint, where games were focused mostly on dungeoneering. I would like to widen the discussion, a tiny bit, to encompass the slightly more modern approach (say, 2E) of both world and dungeon encounters, so take the following with that in mind.
The GM builds the dungeon, but the players are the ones who (via their PCs) scout it out and choose which bits to try and assault, which bits to avoid, etc. If you look at Gygax's discussion of "Successful Adventures" on pp 107-9 of his PHB, you will see this very much in evidence. For instance,
Notice the emphasis on the players being the ones to set an objective, to choose which encounters to engage and which to avoid, etc. This is very different from (eg) the DL modules, or a more contemporary module like (eg) Speaker in Dreams, which assume that the players will proceed through a series of encounters chosen by the GM.
Right, the PCs choose what to take on. AD&D modules are full of encounters, monster or otherwise, that the party should really bypass or nullify in some way, but at the end of the day the DM is still the creator of the milleu (unless you are running a published module, and even then) and has ultimate authority over what encounters the party comes upon. I definitely understand your point, and concede that point, for the most part, if we are talking about running published modules.
This is why AD&D doesn't need encounter-building guidelines, other than the very basic idea that the 1st level of a dungeon should have mostly 1st level monsters on it.
I agree, but only because it is up to the party to decided what is too dangerous. Yes, of course, the DM shouldn't slap a dragon right in front of a 1st level party, but if they decide to wander into a dragon's hunting grounds then that is unfortunate for the characters.
AD&D does not assume that there will be encounters that the PCs can neither defeat nor evade. Ie that there will be hopeless situations in which the PCs find themselves. (At least in general - especially for low-level PCs this may sometimes happen through sheer misfortune, but it's not meant to be a ubiquitous feature of the game.)
I wouldn't argue that it is ubiquitous, but I think AD&D very much assumes the idea that if the party acts stupid, then that's to the detriment of their characters.
I don't think this is right. There are also enforced "social" encounters - eg the GM has a red dragon turn up to talk to a group of 1st level PCs. And there are enforced "escape" encounters - eg this is part of the first DL module, I think.
True, but I find those tend to be in service to the plot of the story or module than anything else.
The difference between these and the Gygaxian style is that the GM has deliberately placed the encounter to generate a certain sort of response from the players. (And you can see plenty of posters still advocating this approach in 5e threads.) Whereas in the Gygaxian style, if the encounter is a placed one - typically, a dungeon room occupant - then the players are expected to control their interaction with it (via scouting, planning etc); and if it is a random one then there are general reaction rules, evasion rules etc to make the outcome a combination of player choice and random chance rather than GM pre-determination.
I agree with the first sentence, I think those type of encounters have their place, even in AD&D, if it serves the story the DM is trying to tell, but there is certainly a fine line between that and railroading. I tend to think that the expectation, in the Gygaxian style, is the same be it a dungeon encounter or a wilderness encounter: solve the encounter. There can be many types of solutions.