Another amazing resource:
Welcome to the UCLA Phonetics Lab Archive. For over half a century, the UCLA Phonetics Laboratory has collected recordings of hundreds of languages from around the world, providing source materials for phonetic and phonological research, of value to scholars, speakers of the languages, and language learners alike. The materials on this site comprise audio recordings illustrating phonetic structures from over 200 languages with phonetic transcriptions, plus scans of original field notes where relevant.
It was a project of the late Peter Ladefoged, a great phonetician and teacher about whom you can read some reminiscences (including a mention of his work as linguistic consultant on the film My Fair Lady) at the MetaFilter post from which I got the link.
I suppose this includes this collection of Ladefoged soundfiles which I have used to find my epiglottis, what a word-final [h] sounds like, and to figure out what it is that I recognize as /h/ (it’s not simply [h])?
Oh my, I’ve had a look at the lists for Polish, and they’re riddled with all kinds of mistakes. The most hilarious one I’ve noticed is listing ‘byj’ as the imperative form of ‘być’ (it’s ‘bądź’). Someone was trying too hard for a minimal pair, it seems, but why didn’t they go with the infinitives ‘być’ and ‘bić’?
On David’s link, above, I clicked to hear BBC English vowels. I was shocked and horrified. This is someone pretending to be a BBC announcer, but someone who is actually of some sort of foreign persuasion. Either that, or he’s got a cold. Or he’s holding his nose. The Ds at the end of the words are all wrong, for example, (tongue’s in the wrong place).
This is someone pretending to be a BBC announcer
I assume “BBC” means “southern British standard” or “RP” rather than “actual BBC employee”. Certainly the speaker does not sound like a trained announcer; probably one doesn’t used trained announcers for this purpose. It sounds to me like a southern British standard speaker whose word-list pronunciation is very different from his connected-speech pronunciation (artificial/unnatural/stilted). Possibly one shouldn’t use such people for this purpose either.
I checked the Slovak sample. “The speaker left Bratislava when he was 2 years old and has lived in the United States since then”. WTF? He definitely sounds like it – “brať” sounds like [bɹac] instead of the actual Bratislava [brac]. “Každá” sounds more like [‘gaʒdaː] than [‘kaʒdaː] which is also strange. “Bočit”, should be “bočiť” [‘bɔt͡ʃic] in both Slovak orthography and standard pronunciation, though [‘bɔt͡ʃit] is not that uncommon in Bratislava and the [c]-[t] shift shows up all across the list. But the one thing that betrays the informant as someone who is not profficient in Slovak are the liquids in “mlkvy” and “mrtvy”. Here is how they’re really sound when spoken by a male, 29, originally from Eastern Slovakia and a female, 22-ish, born in Bratislava.
And just like in Polish, there are some mistranslations here: “vedie” means “he leads”, not “he knows” (that would be “vie”); “na” means “on”, not “into” and “vzchopiť sa” definitely doesn’t mean “hat” (that would be “klobúk”), but rather “pull oneself together”.
No, of course it wasn’t really a BBC announcer. It was possibly Ladefoged himself, in which case… he shouldn’t have been doing it. It isn’t a southern England accent, or even a British accent. It is a Euro accent of some kind, check the Ds.
On mature reflection: the recordings are from the chapter “Vowel contrasts” in Ladefoged’s book “Vowels and Consonants”; so I think the speaker is deliberately exaggerating the vowels and underenunciating the surrounding segments. It’s not meant to be a realistic overall recording. I don’t think the Ds are relevant, although partial devoicing of final voiced obstruents is normal in English.
It sounds like the Ladefoged recordings were made using the speakers that were available, such as foreign students, or indeed himself. And could it be that he had (a trace of) a Welsh accent?
It is very difficult to record authentic speech when the person being recorded knows that the focus is on their pronunciation. I tried once to have a single French text (which had been designed as a pronunciation test) recorded by several French colleagues, originating from different areas, whose pronunciation had some subtle but marked differences if you listened carefully. It was a disappointment as all of them made efforts to sound really standard and also to read the text as if they were dictating it, not as they would say the sentences in normal conversation, in spite of my recommendations to them to sound natural.
I don’t think the Ds are relevant
Yeah, they are. They’re evidence that this wasn’t spoken by anyone who comes from southern England. I don’t point the tip of my tongue towards the roof of my mouth when I say D, I lay the top front nearly flat against it. So if the UCLA Phonetics Laboratory can’t even be bothered to find someone in LA who speaks BBC English (like there aren’t hundreds of British actors there, just waiting for the chance to do a little job like this) then why should I believe anything else they’ve got? It’s sloppy.
I used to partially devoice my final voiced obstruents, but now I think it’s really childish.
The British actors would want to be paid for what they would consider a job. In a university setting it is likely that the persons recorded would be paid a token sum if anything.
Yeah, I can understand why amateurs and non-natives were used, but it’s still disappointing that the results are so unreliable.
If UCLA Phonetics Laboratory has a project like this to do and they don’t have enough money, then they probably just apply for funding before they start making recordings. They don’t just say ‘Oh we couldn’t afford real people, but Peter’s just incredible; you should hear his Xhosa clicking, it sure fooled me. Saved us a bunch of money’.
This izzzzzzzzzzzzzzzzzzzzz….. A TOMATO !!!
(anone hear dat one…?)
Hold on, now. The “BBC English” stuff is from UCLA Phonetics Lab Data which explicitly says “This is NOT the UCLA Phonetics Archive, completed in Dec. 2008 with NSF funding. This page (Phonetics Lab Data) is the phonetics teaching materials compiled by Peter and Jenny Ladefoged” If it’s meant to teach undergraduates about vowels, then the Ds are irrelevant.
If you want the Phonetics Archive English, it’s here. All the recordings are American, apparently.
I studied with professor who had been a student/postdoctoral researcher with Dr. Ladefoged at the UCLA Phonetics Lab. It’s impossible to overstate the love and respect students and colleagues have for the man. Having never met him myself, I was still moved when I heard of his death, as the courses I was taught in phonological science were taught with his methods and his materials and to no small extent in his (recorded) voice. While not impossible, it’s also quite difficult to overstate his contributions to science in general and my understanding of a science I’m very passionate about. I rejoice in his acclaim!
Well, if he’s ok with you David. I would have done it for nothing. He only had to ask.
The Ds aren’t irrelevant. As a consequence of not using someone who speaks the dialect the vowels are not authentically pronounced. Maybe you’d like a recording of me doing an Irishman? Very cheap, I’ll do it for nothing. Good enough?
My favorite so far is the “speaker of Lithuanian Yiddish” living in New York who is clearly heard saying [meshige] and then (presumably self-correcting) says [meshuge]. Therein hangs a tale, I bet.
Could be a natively English-speaking linguist who tried too hard to not aspirate the /k/. I’ll have to listen to that soundfile (…sometime).
@AJP
*sigh* His reputation as a researcher has no bearing on the facts that science is always a work in progress, the man is deceased, and what you would have done had he asked is an entirely moot point. My suggestion would be to record your own samples and then offer them for study.
Thanks. Ok, I’ve recorded them. Now what?
“Thanks. Ok, I’ve recorded them. Now what?”
Upload them to YouTube and post the link here.
Thanks. Ok, I’ve recorded them. Now what?
Don’t forget to run them through Speech Analyzer and add the screenshots.
Upload them to YouTube…and add the screenshots.
You want to see pictures of my vowel movements?
My favorite so far is the “speaker of Lithuanian Yiddish” living in New York who is clearly heard saying [meshige] and then (presumably self-correcting) says [meshuge]. Therein hangs a tale, I bet.
Posted by Zackary Sholem Berger
I don’t quite understand how you, Zackary Sholem Berger, simultaneously: write a blog, translate it into Yiddish, write a novel, teach Yiddish at Columbia, have two small children and translate and publish children’s books — this while you’re also doing a medical residency at Bellevue — how you also have time to read other blogs and comment here. I guess I’m just lazy.
Sorry, I forgot your Ph.D. work. (Luckily, you didn’t.)