Lunchtime experiment: the FOOT/STRUT split

Someone on Language Log does breakfast experiments, but I did a lunchtime experiment today.

Necessary terminology: This is about vowels in English. People tend to discuss these referring to standard lexical sets, keywords that represent all the words with the same vowel. J.C. Wells is often credited with the introduction of standard lexical sets (in his Accents of English from 1980), but I've read something by J. Catford from 1957 that used the same idea already. I don't know if Catford stole it from someone else. Anyway... The standard lexical sets relevant for today are STRUT, FOOT and GOOSE.

Background for the experiment: Apparently I sound Northern English sometimes. This is because I use the [ʊ] vowel in STRUT words where Southern English, Scots, and Americans use the [ʌ] vowel. Southern English and Americans also use the [ʊ] vowel, but they do that in FOOT words. Northern English also uses [ʊ] in FOOT words, so they lack a distinction that Southern English and Americans do make. (Historically, the North is right, and the others introduced the distinction, which is therefore called a FOOT-STRUT split.)

Scots, as said, use [ʌ] in STRUT words, but they don't use [ʊ] in FOOT words. Instead, they use [ʉ], which is the sound they also use for GOOSE words. (English and Americans use [u]. The Scottish [ʉ] is further front, a mixture between [u] and [i] almost.) I think that historically FOOT and GOOSE were distinct, so that the Scottish system represents a FOOT-GOOSE merger, but I'm not 100% sure on this. I'm sure Wikipedia will know. Look there.

So, in summary: Southern English and American have a three-way distinction STRUT - FOOT - GOOSE. Northern English has a two-way distinction STRUT=FOOT - GOOSE, and Scottish English has a two-way distinction as well, but it goes STRUT - FOOT=GOOSE.

What do I do?

The HTML for that last sentence: What do <I>I</I> do?

The experiment: Simple. Record myself speaking, measure the acoustic properties (first and second formant frequencies) of the vowels with Praat, and plot in a graph to see what they're doing. The words were luck, butt, buck, pun, shut (STRUT), look, book, cook, soot, foot (FOOT, obviously), and Luke, chute, lute, fool, rule (GOOSE).

The result:


In this gorgeous graph (F1 on the inverted Y axis, F2 on the inverted X axis, both in Hz), there's the five STRUT words in green, the FOOT words in blue, and the GOOSE words in pink. I've also added five standard vowels – which I must confess I didn't get the formants for from my own recording, but from Wikipedia. They may not accurately represent where these vowels are in my system, but you get the general idea.

So it's obvious I have a Scottish-type system for these vowels: STRUT - FOOT=GOOSE. The realization for FOOT=GOOSE is also pretty Scots, because it's the [ʉ]. But whether I have the Northern English realization for the STRUT vowel (rather than the Northern English system where STRUT=FOOT) is not as clear. The STRUT vowels are all over the place. There's two that are where [ʊ] should be, more or less: luck and pun. Then there's shut and butt which look like they are [ɵ] (I think this is a very cute vowel) or maybe just boring old [ə]. I have no idea what buck is doing. It looks like it is where [ɛ] should be (or [œ] 'cause it's rounded), but that's way too far front (I probably measured wrong). Somewhere at the same height, but further back (say around 1250 Hz for F2) is [ʌ].

So? Well, yes, no one cares. But it's got numbers and graphs so surely this is evidence that linguistics is actually a science? Surely?