The book predominantly rejects 2 with lesser claims about one. She didn't dispute the weakness of central authority in France (and slightly less in Germany), but she pointed out that England had very strong central authority. Italy, of course, had strong municipal authority but the area was not unified. Ironically, the area that's not thought of as being the core of Feudalism (Germany), which used few of the commonly understood Feudal terms, would probably be the best candidate for the traditional definition of Feudalism.
Essentially, schoolbook feudalism is the feudal pyramid and the concept of the fief. The idea is that the King gives land to his vassals (Lords) to hold in fief. That means the King is the owner of the property and holds it in fee simple (meaning he holds it absolutely). The fiefholder (the Lord) "pays" the King for the land by offering military service. In turn, he subdivides his land into lesser fiefs to be given to lesser vassals who owe him military service in rent (which he can then use for his own benefit or to fulfill his obligation to his King to "pay" his rent). Reynolds's book essentially destroys this entire concept.
It's just very interesting to me that something which was supposedly non-homogenous somehow turned out more or less homogenous (really, does anyone claim that it was exactly and precisely the same everywhere in the Middle Ages? I feel like she's arguing with a strawman) by the time it was abolished in the 19th Century.
I think you lack understanding of what the evidence is for Feudalism. For the evidence to work, it assumes words were used consistently over a wide geographic range and a wide time period. Then, to fill it gaps, it'll use similarities of ceremonies (or assumed ceremonies) to assume other words also mean that same thing (for example, Benefice is used as a Carolingian version of Fief). In one source, a person would be described as placing himself in his Lord's hands (or paying him homage). Therefore, it is assumed that "placing himself in his Lord's hands" means the same thing as homage and any time one does either, one is doing it for the purposes of fulfilling their feudal obligations. One fiefholder is said to do that. Therefore, all fiefholders do that. Furthermore, if someone who is a holder of a Benefice does it, a Benefice suddenly becomes a Fief. They do this even when the context makes it clear it isn't a ceremony. For example, if a brother rebels against his King and loses a battle and is trapped and about to be killed, when he throws himself in the hands of his King, that isn't a Feudal ceremony, that's a plea for mercy. But Feudal theory doesn't treat it that way. That isn't a strawman, that's the fundamental underpinnings of the entire concept besides one book, the Libri Feodorum, which most people who talked about Fiefs never read.
Do I think Reynolds was arguing against a strawman? I think she acknowledged that she had to choose a generic definition of feudalism because there are so many shifting definitions that it's virtually impossible to respond to all of them (although, as Elizabeth A. R. Brown pointed out in
Tyranny of a Construct, the fact that historians use feudalism to mean whatever they want it to mean is another problem with using the term at all). But her main point was that the evidence didn't exist when the sources are properly read.
The Fief was essentially invented as a legal concept in Italy as a method for cities there to increase their power over the countryside. As their power grew, they convinced landholders to give up absolute control over the land in exchange for holding it for the city (since land is never held in the absolute when it's subject to government regulation and control, these landholders lost little at first). Lawyers, on the other hand, began to classify and quantify and come up with neat definitions for what Fiefs meant (and then found lots of Fiefs had nothing to do with that, so they had to distinguish between "true Fiefs" and other Fiefs). In France, on the other hand, Fief was just a status symbol. If you were rich and important, your land was called a Fief (except, of course, when lower status people happened to have land called Fiefs). So as time went on, people had their land converted to Fiefs, which they held absolutely just as before, except where Royal authority limited their control (just as royal authority did over non-Fiefs). In England, the word Fief is where we get the modern "Fee," which, in legal tradition, means land held absolutely by the owner subject to no other restrictions.
The only area where the more traditional Fief system (where land is held conditionally in exchange for military service or other obligations to the true owner of the land) is true is with the Church (something we generally don't think of because the traditional feudal pyramid goes from King to Lord, not church to lord). Canon law prohibited the alienation of church land. Therefore, the church would grant land in a Life Estate for three generations (meaning three generations get to hold the land, then it reverts automatically back to the church) in exchange for the landholder to provide service to the church. Of course, once the great-grandchild finds out he has to give up the land that has been in his family forever, he's none too pleased and there's a big dispute. A lot of royal proclamations concern these lands (iirc, the proclamations of a Holy Roman Emperor, who I am drawing a blank on his name, that are the basis of the Libri Feodorum come from a dispute between a land holder and a Bishop in Milan).