The use of intensifiers in the television series Friends between 1994 and 2002 provides a unique opportunity to (1) study linguistic innovation in real time and (2) test the viability of media-based data as a surrogate to“real-world” data in sociolinguistic research. TheFriends data exhibit almost the same overall rate of intensification as similar studies of contemporary English, and the same intensifiers occur most frequently: really, very, and so. Frequency of intensifier correlates with its time origin, reflecting the typical layering of forms in language. Moreover, in Friends the once primary intensifier in North America, really, is being usurped byso, which is used more often by the female characters than by the males. Taken together, these findings support the claim that media language does reflect what is going on in language and may even pave the way for innovation. Television data can provide interesting and informative sociolinguistic data for study.