In late March, my daughter came home. She was born west of Glasgow and lived her first year and a half of life there with a wonderful foster family, also from area. Since late March she’s spent the vast majority of her days speaking only to me and my husband, who are American, living in Edinburgh. This post is about what’s happened to her accent.
When a child moves from one family to another the most important thing to help with the transition is to minimise differences between the two places as much as possible. We switched soaps and laundry detergents to the same ones that they used; we bought and served the same foods, in the same way; we changed our hourly schedule to match hers as best as possible. We learned the songs she knew and liked, and which of her toys and books were the favourites. Linguistically, my husband and I have done our best to use the same terminology that her foster family was using for the objects familiar to her: our default family terms now include nappy, pram, and dummy, to name just a few. But our accents are still distinctly American. Like my other American colleagues at the university, after years in Edinburgh we haven’t acquired any Scottish accent features except for the occasional unflapped /t/. I was worried, at first, that the difference in accents might be disconcerting to her, but happily we haven’t noticed any problems. Unsurprisingly, what we (and our friends and family) have noticed has been a shift in her own accent over the past few months. So I decided to measure it.
Although I’m a sociophonetician, I never planned to collect data for this comparison. Although I wrote down some of the words she knew, and their pronunciations, it was really more for the sake of proud parental documentation than for the sake of studying her accent shift. Our daughter is very verbal. She had about 100 words in spontaneous (initiative) production at the time she moved in (1;7), and then at least 250 words less than a month later (after which, I stopped counting). It was really this incidental fact that made it possible to even measure her production patterns, since all of the audio that these data come from family videos that were taken for the purpose of capturing a cute or funny moment and not for the purpose of linguistic analysis.
In the following plot, each dot represents the mean of all measured instances of a particular vowel class, and the ellipses are one standard deviation from that mean. The vowels are represented by keywords which are taken from Jane Stuart-Smith’s (2003) chapter, “The Phonology of Modern Urban Scots” from The Edinburgh Companion to Scots (rather than, e.g., Wells’s (1982) lexical set terms, which refer to vowel classes in RP). The exception is the BALL class which I’ve labeled that way simply because every instance of that class is the word “ball” (her favourite toy)! I only included vowel classes that have representative tokens in both the March and May recordings.
If you’ve never seen a vowel plot before, just know that the location of the vowel roughly corresponds to where that sound is produced in the mouth, and that vowels that are closer together on the plot are more similar to one another. The lines jutting out from the dots represent how the pronunciation of the vowel changes from the middle of the sound to the end of the sound. These plots were generated using the handy web-based suite, NORM (Kendall & Thomas 2010).
What stands out to me the most in this plot is her vowel in BAIT. While this is raised and fronted in Glaswegian, it’s lower and backer in American Englishes, and you can see this in the plot. I should say that there’s a lexical bias here: in March, all of the tokens were from the word “baby,” while in May, the tokens came from the words “blanket” and “daisy.” But I tried to control for coarticulatory effects, and I’m also confident about this result because it’s something that I’ve noticed a lot, anecdotally. I was telling a colleague yesterday that the strongest holdout to her BAIT-retraction was in the song “Wheels on the bus,” which she came to us already knowing very well. The word “day” in the final line “all day long” has consistently had an /i/-like quality until just yesterday morning when we were walking through town and I heard her say /deɪ/ for the first time.
In the following clip you can hear this change in action. She and I are repeating the word “oops-a-daisy” over and over while playing with her toys. This is a word that she learned before coming to live with us on the 25th of March, and the recording was made on the 21st of May. When she initiates the first instance, the first vowel in “daisy” sounds distinctly /i/-like. But after I offer my own pronunciation, by the third instance that vowel sounds more like /eɪ/.
The other thing you can see suggested in the plot is that her BIT vowel seems to be raising — the Glaswegian one is lower than the American one, so this makes sense. Her May vowel space also seems more compact than her March one, but I would guess that that’s either an effect of language acquisition (her pronunciations are much clearer in May) or the stylistic context (the March recordings were taken from more high-energy, excited speech).
There are other changes that I’ve noticed but which haven’t been audio recorded or which didn’t make it into these data. The one that stands out to me the most is her pronunciation of the word hoover, which had a distinct /i/-like vowel in late March and is now pronounced with an /u/. Unfortunately I don’t have any recordings of the earlier pronunciation (in fact, no recordings of the /u/ vowel all from March except for one exclamative “ooh!” which was too loud to measure).
Anyway, she starts nursery (AmEng: daycare) in September, so we’ll see what happens!