Solution for February 2004


"I hang those keys on a hook."

Lower-case A + Small Capital I
[aɪ], IPA 304 + 319
Well, that's how I transcribed it. I've been having fights with various bits of literature having to do with the nucleus of this (and similar) diphthong(s). But I've noticed two things looking at this again. One, there's clearly a glottal 'attack' to this vowel/utterance, as evidenced by the first two pulses being out of sync with the others, so I probably should have put a glottal stop at the beginning. B (I just saw a rerun of Mad About You and Paul does this), this vowel clearly starts out more like a classic 'script a' The F1 couldn't be higher, so this vowel is as low as it gets--the F2 couldn't be lower, so thsi vowel is as back as it gets. The two formants, straddling 1000 Hz like that are classic for 'script a', rather than [a] as traditionally transcribed in this diphthong. Okay, so anyway, this is a diphthong, the F1 drops (the vowel rises) and the F2 rises (the vowel fronts). So this is [aɪ].

Hook-top H
[ɦ], IPA 147
Well, if I'm not trying to be literal about transcribing my diphthongs, I am trying to be literal about voicing, and the continuous striations here indicate that this is voiced. On the other hand, it's very definitely fricative, and its spectrum is very cleary shaped by the surrounding resonances. Which is pretty classic for [h], except this is voiced. In spite of the usual description, it's not unusually to a voiced [h], especially between vowels. For trivia value, I seem to be one of those speakers who allows [h] word-medially only if it is initial to the stressed-syllable (as in pro[h]ibit, but pro[*h]ibition). Mostly. So this either has to be word initial, pre-stress, or both. As it turns out ...

Ash
[ɒ], IPA 325
Well, once the frication subsides, we've once again got something quite low (the F1, though weak, is just a little lower than it was in the preceding vowel nucleus), but very definitely front. Frankly, fronter (higher F2) than I think I usually produce this vowel. Maybe I'm reacting against the general trend centralizing this vowel, which I found (as others have reported) in California, and has been demonstrated for some Canadian speakers.

Eng
[ɲ], IPA 119
So here we have a segment that's fully voiced, and apparently sonorant, but of greatly reduced amplitude compared to a real vowel. It's got a nice little resonance just below 1500 Hz, which would make you think of [n], if you were use to my voice, except that doesn't jive with the serous velar pinch going on in the preceding transition. So this most likely is velar.

Eth
[ð], IPA 131
There's a sudden change at about 625 msec, suggesting the end of the preceding eng and the onset of something else. It's still fully voiced, but is again of reduced amplitude compared to the nasal. So this might be some kind of voiced obstruent. And given the slushiness of my stop closures, you might be tempted to suggest this is a plosive. But there's that funny F2 thing going on, and the noise at the top of the spectrogram. So maybe this isn't plosive, but fricative. So it's a voiced fricative. There's not a lot of useful transition information, I don't think, to tell us which. Let's rule out the sibilants due to weakness. [h] is possible, since there seems to be some formant-like organization to the noise, but this looks nothing like the preceding [h]. Which leaves us with [v] and Eth. Probably that's the best we can do at this point.

Lower-case O + Upsilon
[oʊ], IPA 307 + 321
Well, it looks like there's a serious discontinuity around 775 msec, but since there's vowel on either side, I don't know what it could possibly be, unless it has something to do with the sudden pitch change in this vowel. But the good thing is that the discontinuity, whatever it is, clearly tells us that the vowel goes from basically mid, 500 Hz, to just a little higher (lower F1). The F2 starts sort of neutral-to-lower than neutral, and gets lower. So this long vowel goes from mid and sort of back to higher than mid and definitely back.

Lower-case S
[s], IPA 132
Now this is what a sibilant should look like, except it's a little short. Spectrally, this is a classic [s], with very broad band energy, quite high in amplitude, and concentrated well above the 4000 Hz range, with very little formant-like organization.

Lower-case K + Right Superscript H
[kʰ], IPA 109 + 404
Okay, I know there was an aspiration mark in my transcription. I must have dropped it somewhere. Oops. Well, here's what I meant by mushy. I don't think that moving noise around 1000 msec is really air moving anywhere. Maybe it's reverberation. But if you ignore it, you can explain the double burst (approaching 1050 msec) as the relase of a velar. A pretty front velar, juding from the concentration of ernergy, not to mention the formants of the followign vowel. So if this is a frontish velar plosive. it is defintely voiceless and probably aspirated, having a VOT of something like, oh, 75 msec (max). This isn't the longest VOT one might see for a velar apsirate in English, but it'll do. Don't ask me what happened the the apiration mark.

Lower-case I
[i], IPA 302
Well, F1 as low as it ever seems to get for me, maybe 300 Hz or so, with an F2 well above 2000 Hz. That's about as high as my F2 can go. So this is the highest, frontest vowel I can produce.

Lower-case Z
[z], IPA 133
Another sibilant, spectrally very similar to the previous [s], except if anything the noise is slightly broader band, extending all the way down to the low frequencies. But this one is voiced. Barely. But there you go.

Script A
[ɑ], IPA 305
Well, look at that F1 and F2, as I said before, straddling about 1000 Hz.

Lower-case N
[n], IPA 116
This is the duration of a flap, and I'm not 100% sure why I decided it wasn't a nasal flap rather than the straight nasal I transcribed. But it's definitely not a straight flap--it's got very clear resonant energy (formants) all the way up. But it's weak compared to the flanking vowels, so it must be a nasal. Its shortness may tell us it's flappy, which makes it mostly likely underlyingly alveolar.

Schwa
[ə], IPA 322
Short vowel. Actually, it looks just like the previous [ɑ] but it's short, and it's F1 is more ambiguous. TMSAISTI. I'm a little concerned that the F3 seems to split into the F3 and F4 of the following vowel (the intervening noise notwithstanding), but, as I'm so fond of saying, oh well.

Hook-top H
[ɦ], IPA 147
Well, here's naother one of these. Definitely noisy, though apparently voiced, and with energy cleary organized in the formant pattern of the surrounding vowels.

Upsilon + Right Superscript Glottal Stop
[ʊ + ???], IPA 321 + ???
Okay, there's no such diacritic, and if there were, it should probably go on the following plosive. But here goes. We've got a vowel somewhere in the mid-to-high range, with a very, very low F2, suggesting both back and round. It also gets creaky towards the end, probably due to glottalization of the following stop--it's a little rough to just be the F0 drop.

Lower-case K
[k], IPA 109
Well, there's not a lot of velar pinch, although the F3 is definitely falling, and the F2 is definitely rising. But the rising F2 rules out bilabial, and the falling F3 probably rules out alveolar. Which only leaves one option. And that would explain the apparenty double burst at about 1825 msec as well.