Date: Wed, 1 Dec 1993 13:42:08 +22305606 From: "Ellen Johnson Faq. Filosofia y Hdes." Subject: dialect diversity I'm back with the statistics I promised last week. The following is a list of the percentages of words/variants linked to each of the non-linguistic variables I studied in both the LAMSAS data and my own. These are just the words that were analyzed statistically: 488 of them in the first sample and 484 in the second one. 1930s 1990 region 6.35 1.45 rurality 4.51 2.89 education 4.10 2.89 race 2.66 2.89 age 2.05 2.69 sex 1.43 1.65 total assoc. w/any above 19% 13% If you look at all of the words collected in response to 150 questions to 39 informants in each set, the vocabulary increased by 40%, though more words in the new corpus were too infrequent to be tested statistically. 1930s 1007 words, 1990 1402 words As noted previously, Lowman and I collected about the same average number of responses per informant, so this is not more words in each person's vocabulary or differences in methodology, but a diff. in the total number of words in use. Although TV is probably a factor in some of these new words (the most likely prospect is giddy-up to get a horse to go), it is difficult to know what changes to attribute to the media and which to education and wider reading, for example. Also, it is not likely to lead to homogeneity, if only due to different audiences for different shows. There are bound to be generational differences in those influenced by MTV and Lawrence Welk reruns (rock n roll!), and those who might pick up terms from certain sports shows (e.g. stock car races) would remain limited to some extent by gender and social class. I thought the following quote might be appropriate to this discussion: "Mass communications technology is a powerful cultural agent. Linguists cannot afford to disregard it simply because of misguided statements by popular writers that exaggerate its importance as a linguistic influence." Ellen Johnson ejohnson[AT SYMBOL GOES HERE]