One of the mainstream positions in general personality theory (i.e. approaches that try to characterize all of personality) is the lexical hypothesis. This hypothesis has been stated e.g. by John, Angleitner and Ostendorf (1987; subscription necessary) as follows:

Those individual differences that are most salient and socially relevant in people’s lives will eventually become encoded into their language; the more important such a difference, the more likely it is to become expressed as a single word. (p. 175)

This hypothesis is the origin of the five-factor model of personality; I won’t review the complicated history, but basically one could say that Allport and Odbert, adopting the lexical hypothesis, started with a huge collection of adjectives and somewhat arbitrarily reduced that list, which then became the starting point for Cattell’s research on general personality traits which is also characterized somewhat by arbitrariness and somewhat lacking reproducibility. Nevertheless, all those joint efforts at some time led to a gargantuan body of papers claiming that the number of personality traits is five, and only five, and thou shalt not doubt Costa and McCrae, and they are the prophets of the BIG FIVE MODEL.

Ok, I got carried away a little bit; Costa and McCrae actually have started to claim that their model isn’t really based on the lexical hypothesis anymore, and they are not using adjective scales, but still their basic hypothesis of five factors has been derived from the lexical hypothesis. And what’s more, Lewis Goldberg, probably the most important figure in mainstream personality psychology apart from C&McC, actually developed adjective marker scales for the Big Five. (I realize that nowadays Goldberg and others think there are not only five, but six general personality traits, e.g. here, but the point I want to make applies nevertheless).

Here’s a simple rebuttal of the lexical hypothesis, taken from Cervone and Lott (2007, probably a subscription necessary): The lexical hypothesis states that people invent a word for things (specifically personality characteristics) that are important for them. This hypothesis has been based e.g. on the myth that there are many more terms for snow in Inuit language than e.g. in English. But this latter assumption is wrong! Cervone and Lott cite sources showing that there are actually “Some Eskimo-Aleut languages [that] have fewer such words than does English” (p. 431). So if there is something wrong with the foundation of the lexical hypothesis, what does that tell us about the models that are based on this approach?