A lap wouldn't be a bad idea, I don't know how to predict how much 3V will shrink in heat treat as I've never worked with it. If it were A2, I would say it likely wouldn't shrink at all, 440V, I'd say it'd shrink a small bit, but my experience in precision holemaking is related to tool and die, not blades, so the blade steels I just don't have a lot of experience with. Your anecdote about adding more as a chef is true. You can always make a hole bigger. Sometimes you can make it smaller, by heating it, but not much smaller, and again, depending on the material and whether heating it again ruins your final product. So the .0011 under reamer with a lap is a good plan.
As for the question about drilled hole size, the answer is simple. Physics apply in machining like they do everywhere else in the world. For every action, there is a reaction. One of the principle goals of machine tool makers is to reduce and contain that reaction as much as possible, but they cannot eliminate it. They might reduce it to a point you cannot practically measure, but it's still there. It's the underlying reason behind tolerances, and why they must exist. So when you squeeze that reamer into an under size hole, the reamer wants to stay it's nominal diameter, and the work piece wants to resist being cut and stay it's current size. This creates flex in the tool that's commonly called deflection. Reamers, of all the various cutting tools, are quite good at resisting it. They're supported by the work on both sides, they have shallow, straight flutes, they are fed axially while cutting radially, and they're generally quite sharp and free cutting. Still, they deflect. So generally, increasing the pre-hole size will increase the reamed hole size, decreasing it will decrease the reamed hole size. It's not always measurable, it's not the ideal way of tweaking a final reamed hole size, but it is a trick that can work. For a practical example, my shop had a job some time back which required to precision 1.181" holes, the diameter tolerance was +/- .0002". We bored these holes for a long time with single point boring bars, but the variability in that process led to a lot of rework parts over time. I put in a solid carbide Komet reamer, to ream that hole, a production tool that could make that hole to size and round 5 to 6 thousand times before needing replacement (at least that's what the Komet salesman told me). We used their recommended drill size, which was -.007 the diameter of the reamer. And not two weeks later, our holes are coming out small. Only .001" small, but that is 5x the tolerance allowed. We did not ream 6,000 holes in two weeks, so I called our salesman and asked what the deal was, this $3,000 reamer is already making small holes!. He asked me to check the drilled hole size. I did. It was .015" under, rather than the .007" specified. There you go he said, change those inserts and you'll be back to nominal. So I did, and we were. That's where I learned that the pre-hole size drilled can have a significant impact on reamed hole size.
Lastly, using lockers or adhesive is always a good practice in my opinion if the application can tolerate it. Some can't, but that's generally not a concern with knifemaking I imagine. I think loctite type products are preferable to adhesives in this application, and if you do choose to use an adhesive I would recommend Rocksett, because it's thin like CA glue, but far far more durable. Loctite 609 is a retaining compound for making up loose fitting slip fits up to .005" oversize. It works very, very well at that. For press fits, I generally use red 262 threadlocker, as that is meant to work in tighter spaces. Some of the carbide punches I used to build were cylindrical sleeves with a punch body cutout EDM cut into them at a .0005" slip fit, we would assemble them with 262 loctite and they would run for 100s of thousands of hits without moving. When they were sharpened to short, we could heat them up, push them apart, and insert the punch body half way and get another few runs out of them. So it works well in tight fits.