The humanist axis of stress in Latin typefaces dictates that the diagonals leaning to the right (/) are thinner than the ones leaning to the left (\). This stress axis also applies to circular shapes; the thinnest parts in the letter 'o' are in the upper-left and lower-right quadrants, and the thickest parts in the other two quadrants.
This is illustrated in typical humanist sans serif designs such as Lucida Sans.
However, there are geometric or constructed sans serif typefaces that, on first glance at least, do not follow this stress angle. Such designs could be considered to have vertical stress (few designs are perfectly monolinear). Unlike in the Lucida Sans example above, the 'w' will have apparent left-right mirror symmetry.
So in designing a highly constructed neo-grotesque, which I recently put up for critique (see thread here), I designed the 'v', 'w', and 'x' to have perfect mirror symmetry.
But then Satya showed that even in some designs with apparent vertical stress, the weights of diagonals varied subtly following the humanist axis. Here is his picture of the 'w' from Univers LT Std 55 Roman (I had suggested Univers as an example of a design employing vertical stress):
I examined the letter 'w' from different fonts in my system, and found examples of both perfect symmetry and subtle humanist asymmetry. The 'w's in Corbel and Microsoft Sans Serif are symmetric, while those in Arial and IPA Proportional Gothic (a Japanese font with a grotesque Latin) are asymmetric. In the Thai font Cordia new, the right-leaning diagonals are just one unit (1/4096 of an em) thinner than the left-leaning ones.
What I would like to know is the reason for the slight asymmetry.
Is it a subtle nod to the humanist axis borne out of a desire not to make the design 100% mechanical?
Or is it for optical correction? Are we so accustomed to the humanist axis that given a symmetric 'w', the right-leaning diagonals will look thicker even though there is no difference in the actual thickness?
If the latter is true, then it must surely be a culture-specific case of optical illusion, since many other writing systems have reversed angles of stress from the Latin humanist axis. Does that even make sense?
Or to put it another way, if I wanted my 'w' to appear completely symmetric, would I make it mathematically symmetric or adjust for this possible optical illusion?