I don't see how you can fulfill that responsibility if you don't understand its manufacture.
In this context,
understand is a parameter - there are infinite degrees possible between full understanding (expertise) and no understanding at all. IMO one can be perfectly responsible for their product with a good understanding at some medium detail-level; it's not necessary to master all the nitty gritty.
V&V done right, confirms that the design is a good idea.
This is a great example why the term "V&V" annoys me so much. The second V typically stands for
Validation, and here it apparently refers to
Design Validation. While the latter is a very common term, fairly-well understood in the industry, it makes little sense if you look at what it actually means. One can't really validate a pure design (which is an abstract object; drawings are not the design, they're just a medium to convey it). To "validate a design" (in the sense regulators intend, and industry typically applies) you need a physical product, one that users can use. So a much more accurate term would be
Product Validation (in contrast with Process Validation, which is a different beast).
In that sense Design Validation, or Product Validation, can't be completed without some fairly-good understanding of the manufacturing process, and especially it's parameters, noises and limitations, but the step where a very intimate knowledge of the manufacturing processes is required is actually Process Validation. Luckily, Process Validation can usually be scoped as a turn-key and outsourced to a Contract Manufacturer. Again, it doesn't mean one should shut their eyes and minds to it - it's always better to have some understanding, involvement and critical thinking even if one doesn't have the expertise; but it's usually unnecessary (and maybe even damaging) to control every little detail in this activity (where outsourced).
When products fail at scale up, it often turns out that the development team didn't include manufacturing expertise.
I tend to agree, and we have a name for it - DFM (Design For Manufacturability). But that should come quite before one even starts to develop the manufacturing process in any detail. It's an essential part of
product design, not something external to it. When design engineers fail at that (and they do, quite a lot, sadly), it is a design failure first and foremost. The level of expertise that a CM typically has, the one that IMO the Specification Developer doesn't need to have, is at much finer levels of detail. It can, and usually should, be incorporated into the product realization process, but it can happen in later stages.
I think contract manufacturing makes good business sense from the perspective of not having to make a huge investment in capital equipment and labor. It is essentially a means of achieving economies of scale. I think outsourcing expertise is quite a different matter.
I think you might have written this wearing your Class III devices hat. When considering lower-risk device types, outsourcing manufacturing (Private Labeling) is, not in a few cases, a way to gain access to technical expertise while focusing on core strengths such as marketing. I'm nothing near a business expert, but it seems to me that it works all right, at least in some of the cases.