Site red 5: Fènghuà 奉化 (sub-subgroup: Taihu-Yongjiang)

The panel below shows the mean acoustics of a large corpus of citation tones from an old male speaker recorded by Phil Rose in 1977. There are some interesting features. Click on buttons next to the panel to hear a small selection of the tokens.

• This speaker distinguishes reflexes of all eight Middle Chinese tones.
• The two MC shang tones IIa & IIb both have two clear allotones: one with rising pitch (IIa.mh, IIb.lh) and one with rise-falling pitch (IIa.mhl, IIb.lhl). The rising-pitch allotones are plotted with dotted lines. It can be seen that the F0 of the rising IIb.lh tone is very similar to the that of the IIb.lhl rise-fall tone without the falling component. The F0 of the rising IIa tone is higher than that of its rise-fall counterpart, but this is probably due to a larger number of high vowel tokens in the former. There is no obvious phonological conditioning of the rising and rise-fall shapes, and so they must be provisionally treated as free variants. However, the IIa and IIb tones in the speaker's verb-object constructions, which can be listened to here, are all simple rises and contain no examples of a falling component. I suspect therfore that the falling component conveys in addition something like "this is the way we say this character". As such it could be treated as a L intonational tonal morpheme.Tokens with the final falling pitch component were more frequent in the corpus.
• These data also provide two clear examples of tonological depression whereby the onset of a tonal pitch is lowered, rather than all of it. Thus the rise-level reflex of MC Ib (blue) is clearly a depressed version of the the upper-mid reflex of MC Ia (red), and can be represented, using D for depression, as [D, H]. The reflex of MC IIIb (green) is a depressed version of the high falling MC IIIa (cyan), and can be represented as [D, HL]. The other tones of course also represent depressed~non-depressed pairs: but the relationship is less obvious because there is less contour shared between them. It is important to note that all depressed~non-depressed pairings are buttressed by tonological behaviour, for example the upper-mid level and rise-level tones constitue a natural class in tone sandhi behaviour, as do the high fall and depressed high fall; the mid rise(-fall) and low rise(-fall); and the short stopped high and low rise. This system can therefore be described as a set of depressed and non-depressed versions of level, rise(-fall), fall and short level contours.
• Note that this speaker's configuration contains two contrasting tones which both have rise-fall pitch:- IIb.lhl and IIIb. Minimal pairs are 怒 anger [nu depressed fall] vs. 努 strive [nu low rise-fall]; 旧 old [ʥ̥ɪu depressed fall] vs. 舅 uncle [ʥ̥ɪu low rise-fall]. They differ in the timing of the rise, which occurs earlier in the depressed falling IIIb. It can be appreciated that this particular contrast presents some problems for transcription/representation unless one identifies the depression component. Thus the depressed fall is [D, HL], whereas the low rise-fall would be [D, LHL].
• Fenghua tones were also described in 1985 by Fu et al. (浙江吴语分区 p.109). They indicate a system with upper and lower versions of level, dipping falling and short tones: [44], [33], [545], [324], [53], [31], [5], [2]. This thus corresponds fairly well to this speaker's system of depressed and non-depressed versions of level, rise(-fall), fall, and short-level.