Men have bigger noses than women. Let's stipulate this is because a big nose is sexually selected--women think it's manly. 

How does this work out over the long term? Let's say the big noses spread through a population. Since the bigness of the nose is relative to other noses, wouldn't the advantage be cancelled out quickly?

Also, won't all the women end up having big noses, too? So wouldn't the manliness of big noses be diluted as well? This would not be a big problem in really, really sexually dimorphic species like peacocks, becase pen hens don't grow huge tail feathers. But in humans, with less sexual dimorphism, where both woman and men have noses, wouldn't the sexual signal be weakened? Or should we imagine that sexual signals don't depend on man-woman contrast as much as the man-man contrast?

Hi Matt,
So basically, if we use your example of bigger noses being sexually selected, then yes, noses will get bigger throughout the population. However there are two traits being selected for here, those males with bigger noses and also a female preference for males with bigger noses. Those females which are able to utilise the fact that bigger noses on males represent greater virility also gain a fitness benefit. As such while male noses get bigger, female preference for big noses also gets stronger. This is a situation known as runaway sexual selection, where the preference and trait both get more pronounced like the situation of the peacocks tail where male's have a highly exaggerated sexual ornamement.
Of course this situation relies on the trait being able to become so exaggerated, it would be unlikely that a nose would ever grow to elaborate proportions, given the natural selection disadvantages it would accrue which would far outstrip any sexual selection advanatges, so every trait will have some inbuilt limit at which it cannot become more exaggerated. At this limit, there will be a considerable energy cost to development of this ornament, and only the healthiest and fittest males will be able to produce it. Studies have shown that many bird's sexual ornaments are reliable indicators of their parasite load etc. allowing females to discriminate between males with good health and strong genes etc.
In humans the there is less sexual dimorphism, and this is a sign that sexual selection is weaker, it is a sign that our mating strategies probably rarely involved one alpha male with access to all the females, and may have been more likely monagamous pairings or perhaps more polyamarous grouping like the bonobo chimps. In these situations, sexual dimorphism is usually weaker as there is less intense competition between males for mates.