My lunchtime reading is currently "Civilizations" by Filipe Fernández-Armesto.
He makes an interesting point, which I'd seen before, which is that hunter-gatherers appear to have better lives than arable farmers. What is significant historically, he argues, is not at what point farming is "discovered", but at what point, and for what reasons, a society chooses to give up its easy gathering lifestyle in exchange for hard work in the fields.
My assumption was always that farming is inevitable because a farming society, with its high population density, will defeat any low-density gatherer society it comes into conflict with. But this can be seen as analogous to the erroneous group-selection arguments in evolutionary biology; a particular behaviour may be beneficial to a group, but if is detrimental to the individual, individuals without it will take over groups faster than groups with it will out-compete groups without. It is not enough for a farming society to defeat a non-farming one, if farmers can have a better lifestyle by abandoning their fields. My explanation needs a few gaps filled in.
What Fernández-Armesto doesn't quite say, but suggests strongly, is that the change from gathering to farming particularly benefits leaders. By making underlings absolutely dependent on central infrastructure (cleared land, irrigation), the leader increases his control over them. Public choice theory, neolithic edition. (That sounds so good I'm putting it in the title).
Fernández-Armesto makes an analogy with 19th-century industrialisation. Landowners benefited, but workers didn't. (Of course, in the long run we all benefited from both farming and industrialisation, but as Fernández-Armesto correctly points out, that could hardly justify them at the time if they made most peoples' lives worse).
The point here is that central control is the mechanism for preventing "defection" back to more pleasant lifestyles.
(Someone may point out if Marx said something similar - I don't know whether he did, but his lot were generally quite good on this kind of historical speculation, for what it's worth. Quite a lot of the history of hitherto existing societies is, to a significant degree, the history of class struggle).
There is a second explanation, which is randomness plus a ratchet - even if going from gathering to farming is unpleasant, going back, once population has increased, is likely to be much worse. So for whatever freakish reason agriculture starts, it's likely to stay, and spread.
There's no immediate practical point to all this, but if we're looking at changing the structure of society, as I have been
of late, anything to do with the mechanisms and reasons for major change is a good thing to have rattling around our toolbox.