The issue with building a slimmed down camera, without the video features, GPS, WiFi and NFC connectivity is that the video section is pretty much just software, leveraging the existing hardware, unless you want to lose Liveview too? Ok there is some added hardware for the audio, and the wireless connectivity and GPS receiver. The problem is that all of it together only adds about £20 to the total parts bill for the whole camera. I have recently been working on some computerised control systems, and I can get single chip WiFi, NFC, and GPS units for around £10 each for the WiFi and NFC, and about £35 for the GPS, and that is buying single items. Given the quantities that Canon build I'm pretty sure that it is not costing them more than I suggested for the parts.
Once you put that extra bit of hardware in the camera, everything else is programming, programming that has already been done. You don't think they write completely new specific software every time the develop a new camera do you? Mostly it is simply a matter of integrating existing modules. Even when a new generation processor is introduced, it seems to use mostly the same instruction set, and all it requires is that the software source code is recompiled for the new camera. So adding new software based features is generally a cost effective measure. What really makes 35mm format cameras so much more expensive than APS-C bodies is that the 35mm sensor is 2.56× larger in area. From what I recall from doing a basic IC design course at university, as part of my electronics degree, is that yield rates work at roughly the square of the chip size. So double the area of a chip, based on the same production techniques, and they are four times more likely to fail, not to mention that doubling the area doubles the cost, whether or not the chip actually works. Additionally most of Canon's R&D costs are going on things like sensor development, along with technologies like on sensor DPAF, and then being able to use this data to do light field calculations to correct for lens deficiencies. Fortunately for Canon R&D costs are not normally directly related to the sale price of a new model featuring the new technology. R&D budgets are usually set up to be some relatively constant percentage of profits, and are simply seen as a cost of doing that sort of business. Building better products usually means more profits some of which can be plowed back into more R&D. This is good, because sometimes you spend ages developing a new technology, only for someone else to beat you to it, and patent it out from underneath you, and sometimes the new stuff simply doesn't work, or costs too much to implement.