Am I wrong that a non-bolt on guitar would be lighter?
Steve

I don't think it matters as long as you get good fit and contact between the neck and the joint.

I could be wrong, but I think if you're gluing the neck with hot hide glue, a dovetail joint won't hold unless you've got really good fit and contact, so it's a guarantee of a good neck joint in a way. You can always make an ill-fitting neck hold with a bolt.

On higher quality acoustics where the neck joints are equally good in both cases, I think the differences are much less important and not worth worrying about. After all, you can change bridge pins and what not to tweak your favourite guitar . . . but it's not like you can change your neck joint . . . or can you?

I like to pay my attention to soundboard design, guitar shape, bridge placement, woods, etc

Possibly. A bit of glue probably weighs less than a couple of steel bolts.
Won't be more than a few grams though.

If you use a bolt on butt joint you can make the neck block smaller, perhaps that's enough to make the bolt on lighter than a typical dovetail neck and block assembly, I don't really know. The weight difference is pretty insignificant either way, I've got a dread with a bolt on neck that is lighter than a Larrivee OM (dovetail).

You are describing a sound I associate with Graph Tech saddles. Maybe you could try one on the Alvarez and report back.
I never have liked the butt joint, because that leaves very little wood in the heel to attach a bolt. Those threaded inserts that Taylor used can pull out, because they are so shallow, and threaded into end grain. If you insist on a butt joint, insert a hardwood dowel vertically in the heel and thread lag screws into it, ala Collings.
In addition, I tended to make stronger neck blocks when I used bolt-ons, because the stress is concentrated at the bolts. It does not necessarily mean a larger block, however. I got the strength by laminating a 1/4" thick cross-ply of maple on the face of a thinner, wider neck block. The total weight was no different than a dovetail block, but the weight of the bolts did make a difference.
As I indicated, I prefer smaller lag screws (like Collings), versus the 1/4" hex head machine screws that Martin and Taylor use. IMHO, the machine screws are probably 10 times stronger than they need to be, and much heavier than the lag screws. For that matter, you could use #10 or #8 machine screws and save considerable weight.

Hide glue gives you those toasted wheat underpinnings; must be true, I read it on UMGF:D
Bolt-on necks are the work of Satan-apparently...

In all seriousness though I have owned great-sounding instruments with various neck joints. All were chosen because of their tones and not the building method. Also, how can a couple of microns thickness of glue-assuming a very accurately cut dovetail-have an effect on tone? There aren't great gobs of the stuff in there-especially given the amount that gets squeezed out when the neck is fitted.

If one had dozens of guitars made of the same wood species, of the same wood thickness and with the same bracing but half of them with bolt-on necks and half with glued necks; you might be able to draw some conclusions about how the different attachment methods affect sound. Otherwise, there are so many variables that any conclusions or assumptions are meaningless.

Most of the sound that the guitar amplifies comes from the vibration transmitted by the bridge. The extent to which the neck transmits audible sound to the soundboard would depend on the phase relationship. If the vibration from the neck is out of phase with the sound at the bridge, then improved neck joint transmission would have a damping effect. If the sound is in phase, then improved transmissio might have a positive effect. The point is that assumptions are being made in the absence of enough knowledge of how the whole guitar works to amplify string vibration.

.