In other words, the marginal emissions caused by the extra load is the actual consequence, not the emissions proportional to that load.

But PHEVs in my opinion have an overlooked utility for demand dispatch and battery to grid transfers, even when compared to BEVs.

The entirety of a PHEVs battery can be used for either task.

A long range BEV that you use to commute may be put in a mode where the power company is free to decide when to charge the battery once it's above a user set minimum. This allows off demand to be shifted to follow renewable production.

But for a PHEV, that minimum is zero. While the battery isn't as big as a BEV, the portion set aside for this purpose may be as large. During a heat wave, they don't have to pull anything from the grid, and can in fact contribute. Dirty power, but much less impactful than building power stations for a few days a year.

The net result is a vehicle that acts as a BEV for commuting, is able to travel long distances without a supercharger, and allows the grid to tolerate higher renewable mixes.

I think it makes a ton of sense today. But the market is shifting so quickly, it may only for a short window.

I think that this is unnecessarily cautious, to the point of being misleading.

The right answer is "Hybrids ARE lower than ICEs when it comes to GHG emissions" - this is true in general and of all hybrids collectively. It may not be true in some rare individual cases, where the hybrid is not particularly efficient and the electricity is generated from coal (see figure ES1 in the Argonne paper you linked above).