If listed as part of a class’s description, it is your class level in that class. Rage powers would therefore mean Barbarian class level when they refer to level. Exceptions will specify another level explicitly. “Character level” is used to indicate your total, overall level.
If the ability in question is not a class feature, but rather a feat, item, or something else, then that is probably an oversight; such things should always specify. Check any errata files, and if there’s nothing, ask your DM (or if you are the DM, make a judgment call).
With thanks to doppelgreener for finding the link, here is a statement by Jason Buhlmann, lead designer, confirming this.
A 3.5 perspective - the background section of Druid in Player's Handbook:
Though their organization is invisible to most outsiders, who consider druids to be loners, druids are actually part of a society that spans the land, ignoring political borders. A prospective druid is inducted into this society through secret rituals, including tests that not all survive. Only after achieving some level of competence is the druid allowed to strike out on her own.
All druids are nominally members of this druidic society, though
some individuals are so isolated that they have never seen any highranking
members of the society or participated in druidic
gatherings.
Player's Handbook 2 gives another bit of detail:
Your secret instruction occurred in caves and forests sacred
to life. You learned clandestine verses, the names of stars
and constellations, the cycle of the seasons, sacred songs,
formulas for prayers and incantations, rules of divination
and magic, and the language of animals.
One more clue, and possibly the biggest one, is the druidic language (which remains in Pathfinder with similar restrictions). No druid can teach it to non-druids, yet all druids know it. While I cannot comment on whether this druidic society exists in Pathfinder, at least the language part must be taught by another druid. Though, generally, I'd guess a druid is similarly trained by other druids. Keep in mind this is all subject to change according to the DM's views of the setting.
As to becoming one in the game, mechanics-wise, there shouldn't be any restrictions, so long as you meet the alignment requirements. One would assume that you found a druid and got inducted into the society "off the screen". It should not be hard to pass any tests when you have character levels, and the society would be happy to have a competent and resourceful member. It is probably similar for quite a few of the classes. After all, if you multiclass as a wizard, you probably did not just buy a blank spellbook and reinvent magic?
The main consideration here is probably the convenience and enjoyment of other players - this is most likely a solo adventure, and everyone else would be just twiddling their thumbs bored.
The distinction between drawing power and being granted power is very blurry, opinion based, and subject to DM's and everyone else's interpretation of deities and the divine. Additionally, 3.5 similarly allows clerics to be dedicated to a cause, alignment or idea (and glosses over on how the non-deity clerics come to be), just as druids can be dedicated to nature, which may imply the existence of some differences.
With that out of the way, I would doubt clerics are "granted" power, except in the interpretation where gods appear to you personally or operate call centers. It's probably more a matter of establishing a divine connection, and then drawing/calling on that connection to grant you power/miracles. Which would be rather similar to druids and their connection to nature. I'd guess the "rewarded for their devotion" quote should be interpreted less directly - nobody specifically rewarded you with the power, the consequence of the process of devoting yourself to nature is access to the power of nature. The druid class then provides the combination of a connection to nature and knowledge on how to channel the power of that connection.
Best Answer
You misunderstand what the class tiers are, and how they are supposed to be used. This is unfortunately very common.
The tiers are intended to be both level and optimization agnostic. That means, in general, that given the same amount of effort put into optimizing, and at the same level (whatever level that is) the higher-tier class will be more capable. That isn’t just power; a lot of it is versatility, flexibility, and so on. Because Pathfinder characters have to handle a lot of very different challenges and problems, the tiers penalize one-trick ponies even if they are very good at that thing.
To do this, the tier system basically treats all levels equally. A class that sucks until 20th level is still rated poorly, no matter how good that capstone feature is (see the 3.5 truenamer, which was nearly unplayable for most of the game but at 20th got one of the most broken class features in the game). Likewise, a class that is really good at 1st level but gets nothing much else from then on does not do very well (especially in 3.5, a number of classes only got good features in their first level or two, and even though those features are good, the tier accounts for them over the course of all 20 levels—being good only for 5% or 10% of the game doesn’t rate very highly).
And this system works in part because of how Pathfinder (and D&D 3.5 before it) progresses. People talk a lot about linear warriors, quadratic mages—which is a common theme in a lot of games, not just these—but that term does not apply all that well in 3.5 or Pathfinder. Usually, linear warriors start out stronger than the quadratic mages, and then grow more slowly, so that later in the game the quadratic mages end up ahead. D&D 3.5 and Pathfinder do not actually do this.
Instead, the quadratic mages are already ahead at the beginning of the game, and then also grow faster, so the modest gap early on in the game becomes massive by the highest levels. In a sense, the idea of linear warriors and quadratic mages kind of does apply—warriors’ progression is almost-linear, while mages’ growth is actually more accurately called exponential, which is even greater than quadratic. But since the mages are ahead even at early levels, the trope of linear warriors, quadratic mages really doesn’t apply.
So it is important to keep in mind that tier does not change with level. But we can still talk about power and versatility gaps, and how they change as level does. Specifically, the advantages of magic and spellcasting grow rapidly, so while classes are nearly peers of one another at low levels, by high levels it’s as if mundane and magic classes are literally playing different games entirely.
A 1st-level wizard has some advantages over a 1st-level fighter, but the difference is not too huge to be overcome. In a duel, a 1st-level wizard could give himself a statistical advantage over a 1st-level fighter (i.e. more likely to win than not), but could not guarantee victory. As levels go up, it gets easier and easier to do so, however. In fact, a particularly notorious series of tests demonstrated that a 13th-level wizard could generally beat a 20th-level fighter (barring stuff like Leadership for a 17th-level wizard cohort).
Likewise with clerics and monks. Clerics definitely get more and more powerful toys later on, and the first level or two of monk are actually fairly good, so at low levels they are much, much closer in power and versatility than they otherwise would be—but I would still rate the cleric as being better, because of its flexibility.
But I do want to focus on the claim of useless—that isn’t really a claim that the tier system makes. We can only really discuss power level relative to other things. In a party of experts and warriors (the NPC classes), a monk looks really, really good. And in a campaign designed for experts and warriors and challenges appropriate to those classes, a monk could easily be considered overpowered. Instead, when we say the monk is useless, we are talking about its ability relative to other classes. It is deep into the weaker side of things, which means it struggles at challenges reasonable for more average classes, and something that challenges the most powerful classes is probably impossible for a monk. In that sense, the monk is “useless”—unable to contribute anything that another class couldn’t have done better, or done in addition to other things. As an option, it fails because other options exist that do things better. (The monk also has other design problems are more inherent to the class itself, but that’s not really what I’m talking about here.)
So is a low-level monk “useless”? Maybe; there are other options that can do the things the monk does, but do them better or also do other things, so if that makes it a “useless” option to you, then the answer is yes. On the other hand, the gaps between various tiers are much smaller at low levels, and monk is a pretty front-loaded class. So if your idea of “useless” is “literally unable to contribute, would be better if the monk wasn’t there at all,” then no, at low levels the monk is not that. As levels go up, the risk of the monk fitting into that definition does go up quite dramatically, but again, that does depend some on the standards of the rest of the party.
The traditional cut-off between “spellcasters have some advantages but it’s not insurmountable” and “okay, now spellcasters and warriors are in totally different leagues” is 7th level. At 6th level, warriors have their first iterative attack, and all saving throws just got a bump, so they’re relatively high. At 7th level, spellcasters can get 4th-level spells—and some of those are just ridiculous. For example, drop a solid fog on a warrior, and he’s basically incapable of doing anything about it. And that’s not even getting into the spells that start warping plot and narrative in complicated ways.
This is why the E6 variant is popular. 6th-level characters have plenty of toys to enjoy, without getting into higher-level magical effects that just completely change the game (and leave mundane characters behind). And the E6 system gives them bonus feats as they continue to gain XP, so your game can continue on for a long time after you hit 6th level, and you still get that sense of progression, and new toys to play with, but again, you leave out broken high-level stuff.