He [Finazzi] denies his [Bach's] music the effect of pleasure for the listener who would not savor such difficult harmony. Yet, assuming the harmonies [that is, musical structure] of this great man were so complex that they would not always achieve the intended result, they nevertheless serve for the connoisseur's genuine delight. Not all learned people are able to understand a Newton, but those who have progressed far enough in the profound science so they can understand him will find the greatest gratification and real benefit in reading his work."("Prologue: Bach and the Notion of 'Musical Science'". Johann Sebastian Bach: The Learned Musician. Christoph Wolff.)
Several interesting things here - the fact that Wolff feels that "harmonies" and "musical structure" are somewhat synonymous in the 18th century German tongue (drawing any tempting conclusions would be unsubstantiated conjecture on my part) and also the fact that, were it a reference to what we'd actually call Bach's harmonies, we wouldn't think of them as complex at all. We've been acclimatized in that respect. If it's a broader "musical structure" that's so complex (such as setting a poetic German text to 9 part choral harmony and two separate orchestras), I see what they mean. Finazzi's still wrong, though. Bach's congregations didn't seem to mind his "difficult harmony" so much, after all.