Brett Reynolds writes:
Over on English Language & Usage, the following question appeared:
Many Japanese textbooks of English mention the "feminine 'so'": the use of "so" for "very" is more typical of a feminine speaker. I don't think this is true in the US (I learned English living in Southern California and have now lived in the US for 10 years), but is it at all true in the UK? In other parts of the world?
I don't have access to a male/female tagged corpus. Would you be interested in following this up?
I don't have time this morning for a very elaborate investigation, but a small (one cup of coffee) Breakfast Experiment™ suggests that the use of so as an intensifier is indeed (statistically, not categorically) sex-associated. Looking at so with a half a dozen different following adjectives, in the LDC's conversational telephone speech transcripts, yields this:
|WORD||F #||F "so _" #||F "so _" %||M #||M "so _" #"||M "so _" %|
Thus those specific following adjectives are intensified with so 4.4% of the time by female speakers and 2.0% percent of the time by male speakers.
In comparison, very yields this:
|WORD||F #||F "very _" #||F "very _" %||M #||M "very _" #"||M "very _" %|
Female speakers precede these adjectives with very 4.0% of the time, compared with 3.9% of the time for male speakers.
Thus a first crude guess would be that female speakers use "so" as an intensifier about twice as often as male speakers do. But there are serious differences among adjectives, so that a much larger sample would be appropriate; we should look at the effects of age; there's potentially a difference between "so X that Y" and plain "so X"; geography may play a role; there's the question of whether the "so" is stressed or otherwise given a marked pronunciation; and so on.
Update — there are of course several other ways in which the use of so is developing: as a modifier of verbs, in phrases of the form "so not X", etc. There's a worthwhile survey in Mai Kuha, "Investigating the Spread of “so” as an Intensifier: Social and Structural Factors", TLF 48, 2004.