It used to be thought that scoring or 'toothing' the surface to be glued made the joint stronger. Partly it was done to 'give the glue someplace to go' when the joint was clamped, and partly there was a belief that most of the strength of the glue line is mechanical, so having more glue in the joint made it stronger. It turns out that a major part of the strength of a glue line (and just how much is debated...) is chemical. In that model what you want is the thinnest possible glue line, with as much wood as close to touching across the joint as you can get. Toothing actually makes the joint weaker, since it reduces the close fitting area of of the glue line, and the thick glue in the groves is brittle. Once the glue at the back of the bridge, where the stress is highest, lets go, the joint becomes even weaker, and catastrophic failure is not far behind.
In addition (as if that's not enough) knife scoring the wood of the top introduces 'stress risers'; places where the stress in higher than it 'could' be. Once the joint starts to let go, the wood can simply peel up to the next scribe line, even if the actual glue line doesn't fail. This is not uncommon on cedar tops if the maker gets a little too zealous scoring around the back edge of the bridge to clean up the finish. Glue can't get into that narrow cut, so the wood is already 'broken' there.
As I say, there is not total agreement on this in the woodworking community, and I would expect some push-back. It's mostly a matter of balance; which mechanism you think is more important. There is no doubt that most of the volume of wood, and particularly softwood, is air. Filling in that air space with almost anything will make it stronger, so there is a role for mechanical strength. Most of the research I've seen says that tightly fitting smooth glue joints that have been freshly planed or scraped are significantly stronger than ones that have been sanded smooth, or left to oxidize for more than about 15 minutes. That research goes back to WW II, and was done in the US Dept. of Agriculture Forest Products lab, so it's public domain.