Read on Twitter
My paper with colleagues here at KU and Jeju National University has finally been published 
In this study we explored the question of how the age of a speaker can affect where it sounds like they're from. https://journals.sagepub.com/doi/10.1177/0023830919861666
In this study we explored the question of how the age of a speaker can affect where it sounds like they're from. https://journals.sagepub.com/doi/10.1177/0023830919861666
Korea has really wide dialectal variation, but a lot of the variation is realized through verb endings and vocabulary. So when you ask people to read text, a lot of the clues to their regional origin are lost. But not all! Intonation and vowel quality also play a role.
But just based on read speech alone, the only region that listeners can usually pick out is Gyeongsang (= southeast). Yes, e.g. Chungcheong dialect is also different from Seoul, but when they're just reading text they sound *really* similar to Seoul speakers. Gyeongsang, less so.
In this study we had speakers from six different regions in their 20s or 50s read some sentences. We then played them for listeners and asked them to guess where they thought the speaker was from.
In general, when the speaker was from Gyeongsang, listeners could often (but not always) tell. But the listeners had no idea what to do with speakers from Gangwon, Chungcheong, or Jeju. They're just weren't enough clues in the speech, since it was read text, not spontaneous.
The pattern we observed was that listeners' responses to these non-obvious speakers depended on the speaker's age, e.g. a young Gangwon speaker sounds like they're from Seoul, whereas an old Gangwon speaker would be perceived as anything but Seoul (and not necessarily Gangwon!)
What was more interesting was that even older Seoul speakers don't sound like they're from Seoul (even though they were born/lived their whole lives there). Listeners would think they must be from Chungcheong, or Gangwon, or Jeju (but not Gyeongsang, bc that would be obvious!)
There are some alternative explanations for this result. So we ran a second experiment in which one group of listeners only heard the younger speakers, and another group heard only the older speakers.
What we found was that when you're only listening to older voices, you're more willing to categorize them as sounding Seoul-like. Presumably because they sound Seoul-like "for an older person". But that same voice would sound e.g. Gangwon-like when compared with a younger voice.
Likewise, when you're only comparing younger voices to each other, some of the ones that sounded very Seoul-like when compared to older voices began to sound more like they could be from somewhere else, like Gangwon or Jeju (but again, not Gyeongsang, bc that would be obvious!)
So, I think the study actually raises more questions than it answers, but to our knowledge it is the first large-scale study of Korean dialect perception. It also shows pretty definitively that regional origin is not always so easily detectable in Korean speech.
I was personally glad to show this because I feel like in Korea I've often encountered the claim that non-Seoul folks can't ever speak "true Seoul speech". It's like, fine, whatever that means, but when push comes to shove you often literally can't hear a difference, so 
This thread unavoidably glossed over lots of small details, so please read the paper if you're interested in learning more! It's paywalled, so if you need a pdf just shoot me an email ([email protected]) with the subject line "Korean dialect paper plz".
