I believe Keith Duckworth's feet were set on the path to the modern racing engine more by chance than by a blinding flash of insight.
The earliest Cosworth engines were based on the old English Ford 105E, which had bore and stroke dimensions of (if memory serves) 81 x 48 millimeters, in round numbers. This Ford engine was fitted with a single-cam cylinderhead for what was, for a time, Formula III, and my impression is that Duckworth did not bother with a twin-cam layout with inclined valves because there was room -- in the 105E's huge bore -- for straight-up valves of adequate size.
Subsequent Cosworth experience with small engines built on the Ford big-bore block provided more evidence that it is possible to use large bore/stroke ratios in combination with narrow valve angles and get superior results.
It must be noted here that Duckworth was building on the work of other engineers at Coventry, and at Weslake. The development of the old Coventry Climax racing engines showed that there exist optimum valve and port sizes, and that both absolute output and (especially) power range can suffer when those sizes are exceeded. As it happened, the optimum sizes could be fitted over the Ford's large bores with the valves set almost flat over the piston crown.
So, without intending to minimize Keith Duckworth's considerable contribution to racing engine technology, I believe the essence of what he did was to see, before others did, beyond the old pattern of bore/stroke ratios close to unity and included valve angles of 60- to 80-degrees. He saw, I believe, balanced and rational integration of breathing, combustion and target-horsepower factors, and his was a vision of breadth and penetrating insight.