In the talk on Wednesday, we contrasted two different patterns
for neutral-tone (unstressed) syllables in Mandarin:  the Guoyu
pattern in the utterance fig01a_rh1b.wav and the Wu-influence 
Putongua pattern in th1b.wav.  In the Guoyu utterance, the F0 
falls gradually across the two neutral-toned syllables in 
hai2-zi-men, as if interpolating from the H target in the 
rising lexical tone on hai2 to a low or mid boundary tone at
the end of the word.  Also only the bu4 in yao4-bu4-yao4 is 
neutral(ized) tone, so that there is a downstepping sequence 
of falls in the reduplicated verb stem.  In the Rugao-accented 
PTH utterance, by contrast, the F0 is high and level across 
the tw o neutral-tone syllables in hai2, suggesting a Wu-type 
tone spreading across them.  Also, the second verb stem in 
yao4-bu-yao is neutral(ized) tone, and there is low level 
level F0 across bu4 and this second verb, again suggesting a 
Wu-type tone spreading.  Fangfang has recorded her husband 
and mother saying this same sentence, and their patterns are
different from both the Guoyu speaker and the Rugao speaker.
See all four utterances in ./typology/haizimen.4speakers


notes on PRACTICE THREE INTERMEDIATE
-------------------------------------------------------------
<<friend2>>

The "Guidelines ..."  say of this file "[Compare to <<friend1>>.]"
These utterances are from a paper by Ayers (1994), in which she 
looked at the differences between spontaneous and read speech in
the ways in which discourse structure is marked.  The utterance
<<friend1>> is the original spontaneous version that this speaker
produced in the elicited narrative, and <<friend2>> is from his 
later reading of the transcript of the narrative.  Comparing the
two gives a good feeling for the meaning of the "canonical 
downstepping contour"  H* !H* (!H*)+ L-L% which Hirschberg 
describes in her paper at LabPhon9.

This utterance also illustrates the difficulty of picking out 
some of the accents in the contour -- in a situation that is 
somewhat reminiscent of the ambiguity in the hat pattern.


-------------------------------------------------------------
<<romanelli>>

The tune on the first part is just the same as in the second
utterance in <<theresa>>.  See also <<mile>>.  The last clause
here also illustrates H+!H*, which is the ToBI label for what 
Pierrehumbert & Hirschberg (1990) called H+L*.  

-------------------------------------------------------------
<<mile>>

You give him an inch,   he takes a mile.
     |          |   ]               |  ]
     H*         H*  H-L%         H+!H* L-L%

This is almost the same intonation pattern as in utterance two
in <<theresa>>.  Kate Welker pointed out in 1987 that this tune
seems typical of constructions like this, and that it is a good 
illustration both of the meaning of the H-H% boundary sequence
and of the H+!H* pitch accent -- which is the ToBI label for 
what Pierrehumbert and Hirschberg (1990) called H+L*.  See the 
timing of the fall on the word "mile"; inflection point is in 
the initial [m], not in the vowel.  Also compare the <<romanelli>>
utterance.


-------------------------------------------------------------
<<thatone>>

Do you really think it's that one?
        |                  |     ]
      L+H*                !H*    H-H%
      L+H*                 L*    H-H%

This pair is meant to give you a clear contrast between a 
downstepped !H* and a L* pitch accent.

Remember that in Pierrehumbert (1980) and later work, downstep
is not a choice on the downstepped accent, but a compression
triggered by the *previous* accent.  Specifically, in her
theory, downstep is triggered by a bitonal accent, so here
the downstep on "that" is triggered by the L+H*

If you buy this theory, what do we do with:

H* !H* L-L% 

In Pierrehumbert (1980), this pattern would be analyzed as:

H*+L H*+L L-L%

Pierrehumbert (1980) also analyzes the accent that ToBI calls
H+!H* as H+L*, and these are related in meaning -- see
Pierrehumbert & Hirschberg (1990), section 5.4.

<<flour1>> last clause:

and ... oh,   I   don't  know.
         |    |    |       |    ]
         H*  !H*  !H*     !H*   L-L%   ToBI 
         H*+L H*+L H*+L    H*+L L-L%   Pierrehumbert (1980)
Beckman & Pierrehumberg (1986), Pierrehumbert & Hirschberg (1990)

-------------------------------------------------------------
<<word>>

Your word is your word.
      |            |  ]
    L+H*          !H* L-L%

This one shows the delayed peak marker for the L+H*, and
also contrasts downstep to the non-stepped H* (or L+H*?)
in <<word1>>:

Your word is your word.
      |            |  ]
    L+H*           H* L-L%