Here's a nifty article on human pheromones (or more accurately, how they don't exist in humans, or in mammals). One bit that makes perfect sense to me is:
Not surprisingly, scientists do not agree on what defines a pheromone, and attempts by chemists to identify such putative agents have failed. Among the many reasons for this failure is that the basic tenet of the concept - that one or a few hormone-like chemicals, specific to each species, are triggers of social behaviour - is wrong. In mammals, chemically mediated behaviours are rarely hard-wired, and most biochemicals involved in communication between members of the same species are not specific to that species and comprise many compounds, some of them affected by diet, stress and other factors.What I've seen in discussions of pheromones is what I've seen in most evo psych, a desire for some way to reduce much of human behavior into simple mechanistic formulas. This is to some degree possible with hive insects, both because arthropods are very different from vertebrates, and because hive insects are exceptionally specialized creatures. In any case, I find the fact that the various theories surrounding humane pheromones are based on this desire for simple mechanistic explanations to say well more about the psychology of the people who promote them than they do about the mechanisms of human behavior.
The two main classes of pheromones said to exist in mammals are "releaser" (biochemicals that elicit particular behavioural responses in others) and "primer" (biochemicals that alter endocrine function in others). In fact, nearly all phenomena attributed to releasers in mammals turn out to depend on learning, context, or novelty. Take mating preferences. When mice of strain A are fostered as babies with mice of strain B, they tend to prefer to mate with B mice rather than with A mice, their own genetic strain. Pheromones need not be invoked to account for this behaviour as it can be explained by the smell of the foster nest: the fostered mice mate mainly with mice that smell the same. The learning of smells can occur before birth, with adult offspring of a number of species, including humans, showing a stronger preference for foods and smells to which their pregnant mothers were exposed.
As for primer pheromones, in mammals most phenomena attributed to them turn out to reflect physiological and psychological responses to abnormal changes in the social and physical environment, such as stress. And when it comes to hormones and the endocrine system, those influences come from many sources.