Solution for May 2002


"It comes in a huge bottle."

To properly view the phonetic symbols in the text below, you must have installed either SILDoulos IPA93 or Lucida Sans Unicode. In the description that follows, the first line gives the name of the symbol (from Pullum & Ladusaw, 1996) followed by the IPA reference number for the symbol named. The second line provides the symbol as a typed symbol in SILDoulos IPA93, followed by the same symbol as a Lucida Sans Unicode symbol. One or the other of these may not appear as the appropriate symbol, depending on which font(s) you have installed.


Small Capital I, IPA 319
[I], [ɪ]
This may start with a glottal stop, but I just couldn't cope with trying to explain why or why not, so we're starting with the vowel. The F1 is low, The F2 is quite high (1900? 1950? Hz). If you look a little further down in the spectrogram (between 900 and 1100 msec) you'll see the F2 getting higher still, suggesting that this isn't the most front vowel available. What's incredibly front but not as front as /i/?

Lower-case T, IPA 103
[t], [t]
There's not much in the way of useful transition information. But in the absence of any indication of velar pinch, and any indication of labial transitions, alveolar is the best guess. Note the apparent double burst, possibly triple (quadruple?), depending on what you count as what. I think the first might be the actual alveolar closure, the second the alveolar release, and the third (when the aspiration starts), the release of the following stop. See below.

Lower-case K + Right Superscript H*, IPA 109 + 404
[kH], []
*Technically, P&L don't name the right-superscript letter diacritics for aspiration, palatalization and so forth. This should probably be called 'Right Superscript Lower-case H', and I'd really prefer 'aspiration mark', but it's not always up to me.
There's not a lot of information here either. The lag between the big burst at about 250 msec and the beginning of the aspiration at 300 msec or so is a little long for your typical double-burst situation. And I think there might be a true second burst (of a velar release) about 20 msec after the aspiration begins (there's certainly something odd about the noise between 500 and 1700 Hz just surrounding the 300 msec mark). But there is some evidence of velar pinch in the transitions into and through the following vowel. The F3 is weak and difficult to distinguish, but there is some indication that both F3 and F2 (for instance in the following vowel are pointing to about 1500 Hz in the release burst/early aspiration moments, where the noise is definitely concentrated around 1500 Hz. This is the middle of the F2 range, and noise centered here is typical of dorso-velar type approximations (hence velar pinch). Definitely voiceless and aspirated at least.

Turned V, IPA 314
[], [ʌ]
If you transcribed this as Script A (IPA 305, SIL [A], LSU [ɑ]), you'd be justified. This has the high F1/Low F2/straddling 1000 Hz configuration of that vowel. But comparing it to the similar vowel later (1400 - 1600 msec), it's not quite as high an F1, not quite as low an F2, and it's quite short. Could be prosodic, although the amplitude is a little high and the pitch periods a little close together. Could be random, no two vowels are ever the same. But it might be the vowel it turns out to be. The official description of Turned V in the IPA is as a back, unround vowel, which is true of this vowel. But so often this vowel is central, it should probably be transcribed with another symbol. Especially in my California data, where this vowel is clearly more fronter-than-central. But in this token, it's quite back.

Lower-case M, IPA 114
[m], [m]
The abrupt change in amplitude indicates something is going on, This isn't a prototypical nasal--there's no clear zero, there's energy in the higher frequencies. But this doesn't have the structure of any of the usual approximants either. No swooping transitions into it or out of it. So this is probably nasal. If you know my voice, you could probably guess bilabial, especially given the transitions into it, but knowing it's nasal is probably enough for now.

Lower-case Z, IPA 133
[z], [z]
Just vaguely voiced (I can convince myself there are some striations at the bottom, although this might not be obvious on the web. Definitely fricative, spanning the spectrum. But the noise gets slightly stronger as you go up in frequency, suggesting /s/ noise. So [z].

Barred I, IPA 317
[], [ɨ]
Very short vowel, so probably reduced. Usually transcribe reduced vowels as schwa, since they don't carry enough information to make it worth doing more. Following Keating et al. (1994), if the F2 is closer to the F3 than the F1, I use barred I rather than schwa.

Lower-case N, IPA 116
[n], [n]
Another nasal, again without swooping transitions or a clear zero. On the other hand, this one is definitely different from the preceding one, and the pole, such as it is, is higher, closer to 1500. This is typical of my nasal pole in /n/. Notice here the other argument that this isn't an approximant--the transitions are not really continuous with the spectral pattern in the closure.

Barred I, IPA 317
[], [ɨ]
See above. But notice this one is transitioning in F2.

C Cedilla, IPA 138
[C], [ç]
Ah, controversy. All those who transcribed this [s], raise your hand. Esh? [h]? Okay good. This is the famous example of a true palatal (as opposed to post-alveolar/palatoalveolar) fricative. But, you cry, English doesn't have a palatal fricative. Well, yes it does. This is an /h/. /h/ is voiceless (usually) with glottal or epiglottal noise. This is source and it resonates through the vocal tract the way voiced source does. The result is formants, excited by noise. Here, it's combined with a /j/ supralaryngeal articulation, owing to this word being what it is. So imagine a palatal approximant [j], and now make it voiceless. But, you cry, the IPA doesn't have a voiceless palatal approximant symbol, it has a diacritic. True. But this isn't approximant. It's a fricative. Look at it. Now, this /h/ isn't absolutely voiceless. There's (all right, there are a lot of striations here. So I probably should have used the voiced palatal fricative symbol here, but no one would have recognized it. So I didn't. But it is. And for those of you who care, it looks like this: SIL [], [ɟ].

Lower-case J, IPA 153
[j], [ɨ]
There's just enough here that is clearly voiced and not fricative to transcribe, so that's what I've done. Also, leaving it out would just be confusing given my voice of vowel following.

Turned M, IPA 316
[], [ɯ]
The F2 indicates this vowel is back, but it never quite gets as far back and round (i.e. low F2) as a good back, round [u]. In my dialect this vowel (/u/) is almost never round, and never fully round, or if it is round, it's definitely never full back. So I've transcribed it as unround and back, because that's what it seems to be. Central and round is another choice, but given that it's me, and I know my rounding/backing situation pretty well, I pick the other one. BTW,this sounds that people transcribe conventionally/phonologically rather than phonetically. This is the famous /ju/, which has its origins as a reflex OE or ME /y/ (front and round), which 'decomposes' to /ju/ in those "new/few/due" words that Wells (1984) lumps into the GOOSE set. Compare this with a /du/ sequence sometime. You may learn something.

Lower-case D, IPA 104
[d], [d]
This one is going to get confusing, because this is the stop portion of an affricate. But I couldn't talk about them as a unit without making it more confusing. So here goes. This is a gap. The F2 rises into it, and the F3 doesn't really fall. So this is probably alveolar. It's got a fair amount of (mostly perseverative) voicing, so I'd transcribe it as voiced. If it were underlying voiceless, it might still have some perseverative voicing, but not that much. So it's probably underlyingly voiced. Keep that in mind and move on to the fricative.

Esh, IPA 134
[S], [ʃ]
Fricative, quite strong for its duration, broad band, but this time not strongest in the higher frequencies. Concnetranted in the range of F2 and falling off sharply in the low frequencies. Pretty typical of Esh. And it's completely voiceless. But, you cry, how can you have a d-esh cluster, especially followed by another stop. Well, you can. But underlyingly, this is a d-yogh affricate, i.e. a <j> in English. Actually a 'soft' <g>. Moving on quickly.

Lower-case P, IPA 101
[p], [p]
Long gap here, indicating a stop. The noise is drifting down in the preceding fricative, and the transitions, sort of, are rising in the following vowel, indicating labiality. Totally voiceless throughout, but with very (very, very) short VOT. So this is voiceless, unaspirated [p], which I have transcribed accordingly. So, given that it's English, it's /b/.

Fish-hook R, IPA 124
[S], [ɾ]
Very short, and actually this one is laterally released, but I'm not sure how you'd tell that from the spectrogram. Teeny short things like this are best transcribed as a flap and forgotten before your brain explodes.

Lower-case L + Mid Tilde + Syllabicity Mark, IPA 155 + 428 + 431
[l`], [l̴̩]
The F3 rises sharply before dying off completely, suggesting a high F3. F3 is difinitely up out of normal range. F2 is definitely very low. It would be easy to just say this was more Script A, but the upper formants do not permit this interpretation. The high F3/F4 indicates lateral. I'll leave it to someone else to describe why. If it were /r/ the F3 would be low. If it were /w/ the F2 would be lower. If it were any other vowel, the F3 and F4 wouldn't be the way they are.

Support Free Speech

Robert Hagiwara, Ph.D.

Linguistics Department
University of Manitoba
Winnipeg, Manitoba
CANADA  R3T 5V5