It was and it wasn't a change.
It was fairly different from 3e where mechanics were fairly consistent for monsters and PCs.
For OD&D, Basic, and AD&D 1e and 2e monsters were different from PCs who also had some differences from NPCs.
The thing to remember here is that, in many of these cases, the differences were reflective of in-character aspects of what the system was telling us (i.e. what it said about the "implied setting"). In AD&D 1E, for instance, Gary Gygax talks about the humanocentricism that the game assumes on page 21 of the DMG. But while he doesn't expressly delineate how this manifests in decreasing potential for improvement the further away you get from humanity, we can see it in action:
- Humans are the
saiyans wunderkind of the game world, able to advance without limit in all classes, save for those which only have a limited number of levels to advance through.
- Demihumans have some degree of selection, but there are many classes for which they're simply not human enough to qualify, and others for which their lack of humanity (even for half-human races) limits how high their potential can reach. Thieves are an exception, but I always saw that as having to do with how most thief abilities are static percentages rather than opposed checks (i.e. they're about personal skill rather than mastering external forces such as magic or overcoming direct opposition by other beings in combat, etc.; half-orcs have unlimited advancement as assassins, but assassins top out at 15th level, so that's moot).
- Humanoids such as orcs, hobgoblins, gnolls, bugbears, etc. have very little potential. A few reference having a chieftain or local leader with slightly better attacks and Hit Dice, and some have very limited magical potential (i.e. shamans and witch doctors), but other than noting how there'll often be children with weaker statistics, humanoids can't really become more than they are.
- Other monsters don't even have that. They might be impressively powerful (e.g. a beholder), but other than growing from childhood to maturity, their abilities and powers are fixed. Dragons aren't so much an exception to this as they are a twist on the "childhood to maturity" part, continually growing into more power over time. But that's really it.
Of course, the standard started breaking down almost immediately. The reasons for the class and level restrictions for non-humans weren't well understood (which was partially because those reasons weren't explicitly stated, and Gary's essay on page 21 of the DMG only addressed them obliquely, treating the game's humanocentricism as
fait accompli), and even if they had been explained (i.e. by saying that it wasn't that every other race was incompetent, but that the demihuman gods were more skilled at creating mortal races, and that humans – who had no creator god – were some sort of anomaly that deities were interested in recruiting to their cause, hence why there were so many human pantheons), it doesn't seem likely that many people would have accepted that, since it brought forth a lot of presumptions that I'm guessing a great deal of players weren't necessarily on board with.
The result was that UA pushed level limits for AD&D 1E, and 2E began offering supplements that rolled back a lot of demihuman restrictions. Basic introduced new "attack ranks" for demihumans (in the
Master Set, as I recall) by calculating accumulated XP beyond level limits, along with introducing new races-as-classes in supplements. And of course
Dragon was replete with alternative rules (or so it seemed, at least).
A PC halfling has a Charisma score stat. A halfling in the monster sections does not. The 1e PH says that a PC dwarf cannot be a cleric, but an NPC one can.
With regard to monsters (and NPCs presented in the MM), it's clear that they did have the same ability scores as PCs; it's just that they weren't actually presented in their MM entries. For instance, the entry for giants notes that they have Strength scores of 19 to 25 ("as compared with humans"), and when standard gnolls are said to deal 2-8 points of damage, the fact that gnoll guards deal 3-9 and chieftans deal 4-10 suggests that (in addition to their other noted stat improvements) their Strength score is higher.
With regard to certain demihuman race/class combinations being permissible but restricted to NPCs only, that looks like a holdover from the notations in
Supplement I: Greyhawk:
Among the dwarves themselves, but never as a player, there are clerical types. Dwarf clerics are found as high as 7th level (Lama), and they can cure and resurrect their own. These clerics are also fighters.
Among the elves there are clerical types as high as 6th level (Bishop) who interact only with their own kind. These clerics (fighter/magic-user/cleric types) have magical ability limited to the 6th level (Magician).
Which is to say, it looks like this is hard-coding in a cultural practice (explicitly so among the elves, but only implied for dwarves, i.e. that odd note about how they can cure and resurrect their own; the implication being that no one else can?) into the rules, something else I suspect that players chafed at.
With regard to the notation in the AD&D 1E DMG (page 100) that listed alternative ability score modifiers for non-human races, that's interesting because, taken in context with how there are
also ability score modifiers presented for (some) classes that NPCs have, it seems to suggest a shift in presentation. Whereas before the Core Rulebooks presented the rules for PC races and classes as telling us how things worked from an in-character standpoint (i.e. simulationism), here the DMG looks to be giving GMs a shortcut for developing NPCs (still using randomly-generated ability scores, notice) that can serve as foils to the PCs (i.e. gamism); there's a reason why it even offers NPC-only "occupations" that further alter ability scores.
Of course, this served to widen the cognitive gap even more, which is why this also seemed to be something (along with things like the one-minute round) that was either overlooked or ignored by a lot of people (though the fact that most GMs didn't want to determine their NPCs' ability scores randomly certainly helped). It's no surprise that these alternative ability score modifiers vanished in 2E.