Dual Nature of Soyaltepec Mazatec Tones

89 48 Pike’s Regressive Perturbation actually Register Lowering? Pizer 1994:115 Underlying form Surface form H L M 1 L a. ɲu˧me˥ + -ʔna˩ → ɲu˧me˦-ʔna˩ ‘bumblebee’ ‘my’ ‘my bumblebee’ H M 2 M 1 M 2 b. ts︢u˥ + ja˧ → ts︢u˦ja˧ ‘he says’ ‘inside’ ‘he explains’ 48a shows a High tone being lowered to a Mid-high tone when it is followed by a Low tone and 48b shows the same result when the high-toned syllable is followed by a Mid tone. The phenomenon of register change offers a phonological explanation for what may appear to be a phonetic effect. These two examples of register lowering thus far have no other phonological explanation. The representations employed to this point offer no motivation for the register effects. The mechanism and structures involved will be discussed in §2.4.5. It is possible that the H-Spread rule discussed above in §2.4.2 in which a Mid-high tone followed by a High tone across a morpheme boundary is raised to a High tone would also be an example of register raising. The changes involved in neutralization appear to indicate that Soyaltepec contours are acting as units, so, according to the traditional typology they should be analyzed as Asian type tones; however, how then should the process of M 1 -Shift be handled? There is an apparent discrepancy in the behavior of Soyaltepec tones.

2.4.4 Dual Nature of Soyaltepec Mazatec Tones

Pizer demonstrated a tone process in Soyaltepec, M 2 -Shift, which requires that the members of a contour tone exist and behave independently, but she also demonstrated the process of neutralization which seems to require that the members of a contour act as a unit. First, I summarize the effects that Pizer refers to as register effects, and then I discuss how these effects relate to the traditional Asian contour tones. The processes that Pizer describes as register effects or potential register effects can be summarized as followes: first, a High tone raises the entire M 2 -M 1 contour to M 1 -H 47; second, both Mid 90 and Low tones lower the entire M 1 -H contour to M 2 -M 1 46; third, both Mid and Low tones lower a H to M 1 48; fourth, a High tone may raise M 1 to H which is a potential alternate explanation for the process in §2.4.2. In order to process these facts, it is necessary to place them within the context of the greater picture of the Soyaltepec Mazatec tonal system. First, notice that there is a lower contour in Soyaltepec Mazatec, the Low to Mid ˩˧, which is not involved in the alternations that occur, so it is clear the ‘register lowering’ does not lower the contour in question all the way to the lowest contour possible. Second, M 2 and L are not raised; only the M 1 is raised if H-Spread is understood as a raising process. Furthermore, when the H is lowered, it is only lowered by one tonal level to the M 1 . Finally, there appear to be three classes of tones: L and M 2 which cause lowering, H which causes raising and M 1 which apparently does not have either a raising or lowering effect, making it a neutral tone. The traditional Asian type register lowering or raising involves two tones with the same shape that can either be expressed in the higher register or the lower register, depending on the context. The gradient process that Pizer is discussing is something else entirely. There is further support for the notion that the contours of Soyaltepec are different than those in Asian languages. Pizer points out that the five-way contrast of rising tones in Soyaltepec, 49, is impossible to represent in terms of two registers. 49 The five rising tones of Soyaltepec Pizer 1994 :116 a. M 1 -H b. M 2 -M 1 c.L-M 2 d. M 2 -H e. L-M 1 the˦˥ the˧˦ tõ˩˧ nta˧˥ nt ʃ︢e˩˦ ‘rubbish’ ‘itch’ ‘money’ ‘good’ ‘cooked corn’ Most linguists studying Asian tone specify two registers [ ±upper] and divide each into two tone levels [ ±raised]. Furthermore, they restrict contours to occurring within a given register. 56 In other words, a rising contour would move from [-raised] to [+raised] on either the [+upper] or [-upper] register. Thus defining two rising contours. Any attempt to define these five rising tones within the two traditional 56 See Yip 1989 and Bao 1990 for discussions of Asian contours. 91 registers fails. Only 49a and c align well with traditional Asian tone theories as can be seen below in 50a and b. Pizer offers the following structures for consideration. 50 Rising tones diagramed per YipBao Pizer 1994:116 a. M 1 -H b. L-M 2 c. M 2 -H d. M 2 -M 1 X H l h X L l h X ? l h X M l h First, it is impossible to define M 2 -H 49d and L-M 2 49e within one register as they each span three of the four tone levels. Pizer never tries to explain the structure she gives for M 2 -H in 50c, she simply notes that there is no way to limit the M 2 -H to one of the registers, thus she places a question mark in the register position. Second, it is unclear how to handle the M 2 -M 1 49b which would rise from the upper level of the lower register to the lower level of the upper register, something impossible to represent if both tones of a contour need to have the same register. Pizer suggests the possibility of adding a third distinction to the register mix, namely an M as in 50d but notes that there is still no way to make a distinction between rising tones that begin at the same point but rise different amounts such as the L-M 2 versus the L-M 1 . Since the traditional Asiatic approach fails, perhaps the Africanist approach which considers these tones as concatenations, i.e., completely separate but linked to the same TBU, would be more appropriate. However, Pizer’s analysis indicates that there are examples where a unitary approach, impossible to express with concatenations, would resolve behavioral oddities that otherwise appear arbitrary. Alternatively, it may be suggested that some of Soyaltepec’s contour tones are actually unitary tones, while others are concatenations. One problem with this approach that is pointed out by Pizer 57 is that the same contour appears to behave sometimes as a unit and other times as a sequence. For example, the Mid to Mid- high contour tone is the trigger in M 1 -shift, a process which separates the contour over two syllables i.e., tones in sequence, and the target of neutralization, a process which requires the contour to operate as a unit. 57 This dilemma pointed out by Pizer is solved in my dissertation with fresh data which behaves differently, thus making the resolution of the dilemma a mute point. 92 Besides, traditional tonal typology calls for an eitheror approach: contour tones either operate as units Asian or they behave as sequences African.

2.4.5 Pizer’s Resolution