lingo.lol is one of the many independent Mastodon servers you can use to participate in the fediverse.
A place for linguists, philologists, and other lovers of languages.

Server stats:

66
active users

#LanguageVariation

0 posts0 participants0 posts today

Strange to not have any #Labov quotes to share that I find poignant despite being the giant of #sociolinguistics and despite how much of his work I've read. He always struck me as a workhorse with a lot of interesting method ideas but not someone terribly concerned with #socialtheory where one's writing might become more touching. His legacy and influence will undoubtedly persist for a very long time still.

Spotted on #sociology reddit: an "is regarded period" construction. The absence of "well" feels incredibly awkward to me, but apparently the version without "well" was not only more common in older literature but is still more common today (books.google.com/ngrams/graph?).

(Also, I have no idea what the answer to this question is.)

To "throw a rock in a haystack". This is a new #saying for me, and I can't find any references to it in Google books or on the web in general. #Crawford is from #Omaha, so maybe it's a saying in #Nebraska considering the state's #agriculture activities. It implies to me that one is making a commotion to get whatever it is one wants to come out and talk.

youtu.be/0at0kzeiPKA

Spotted in #Xmen c1966: "shall waken". We can suppose this in variation with "awaken" and take into account #style to an extent if we compare only when preceded by "shall". Unsurprisingly, "awaken" has always been more common in American #literature when #intransitive (books.google.com/ngrams/graph?), although these sort of pairs where the standard #transitive version is used intransitively could be interesting to look at further

Replied in thread

@grvsmth @maitxinha @linguistics It wouldn't be hard to quantify for the features that have been studied -- just count how many times the conclusion has been that it's a change from below vs not -- but it would be difficult to generalize to all #languagechange from there since #languagevariation/#sociolonguistic studies pretty much never use random sampling. But when the conclusion is the same for feature after feature in community after community, it becomes easier to say the claim is accurate