3.7. Diseño estadístico
3.7.2. Prueba de hipótesis para el objetivo específico
If dentals are represented as in (15) above, how do we represent palato-alveolars?^"^ The question arises because I, which has the effect of fronting in consonants, is
uncontroversially associated with palatalization and so dentals must be distinguished from palato-alveolars and also from palatalised alveolars and yet all three forms will constitute some arrangement of A and I. What we have to decide here is the form that these arrangements should take.
Palato-alveolars in English are an interesting class. They have two main sources:
coalescence, whereby /t, d, s, z/ merged with /j/ as in /tisju:/ [tiju:], for example, and so-called 'velar fronting' (which will be considered in some detail in chapter 6). It is
worth quoting Lass (1976) at length here, because his comments on palato-alveolars have been accepted by a number of phonologists, including Anderson and Ewen (1987).
"To take the phonetic side first: [s J] for instance vary considerably from dialect to dialect, and even from speaker to speaker, in the backness of the coronal occlusion. Thus for some [J] is indeed quite posterior to [s], while for others it is only marginally so (if at all). W hat is however distinctive in all these cases is tongue SHAPE, not position, in two ways. Thus in the first place [J] always has the blade of the tongue domed, whereas it is hollowed for [s]...And in the second place, they are strongly PALATALIZED: there is a definite raising of the tongue toward the hard palate...The palato-alveolars are hyphenated not because of an 'intermediate' position, but -properly - because of their SECONDARY
ARTICULATION." (Lass 1976: 189)
If we accept that palato-alveolars are alveolars with secondary articulation, then the question of how to represent palato-alveolars is also the question of how precisely to represent secondary articulation?
Ladefoged (1982: 210) defines secondary articulation as "an articulation with a lesser degree of closure occurring at the same time as another (primary) articulation." Catford (1988: 103) describes two types of co-articulation: double articulation and secondary articulation. He claims that the main difference between the two is the rank of the articulation.^^ What he means by this is that in the case of a double articulation, such as
[£p] for example, both the labial and velar instructions have equal status i.e. they are both stops. However in secondary articulation the two places of articulation do not have equal rank: the secondary articulation is usually an approximant. Sagey (1988; see also Kenstowicz 1994) argues that double articulations and secondary articulations have the
same representation and that it is languages that differ in the way in which they interpret such representations in the phonetic implementation.
Sagey (1988) assumes that in the phonetic implementation the segments manner features are copied under the articulator node responsible for the primary articulation. The
manner features of the secondary node are "supplied" and the manner features for approximant are the default settings. The major articulator is indicated by an arrow from the root to that node. This is illustrated in (16).
(16) [ + c o n t ] PLACE Coronal Labial [s"l [+round]
An alternative approach might be to assume that when a 'double' articulation is produced, e.g. [kp], both labial and velar are indexed to the manner feature [-continuant]. However, when the correspondence between manner and place is strictly one to one the primary articulation will be read with the manner specification and the remaining articulation will be read as a glide, since it is vocalic information in an onset position.
Dependency Phonologists draw a representational distinction between secondary articulations and double articulations as follows: "Consonants showing secondary articulation are, at least phonologically, single segments in which the components characterising the place of secondary articulation are subjoined to those characterising the place of the primary articulation." (Anderson and Ewen 1987: 247) This is illustrated in (17)
(17) 1
1
l,u
palatalised alveolar velarised alveolar 16
(Anderson and Ewen 1987: 247)
Double articulations are represented in DP simply as ‘double’articulations:
(18) {ICI} {ICI}
{111,lull
[kp] [tp]
(Anderson and Ewen 1987: 250)
Notice that on a DP approach secondary articulation is not viewed as the superimposition of vocalic features since the linguality component {111} is non-vocalic.
But since a combination of A and I has been used to represent a simple articulation: dental, we do not want to say that this kind of representation can be used to represent a 'fronted' segment, a palatalized segment and a doubly articulated segment. W hat we have seen so far supports the adoption of the structure in (19) for palato-alveolars, a structure used by Smith (1988) to represent palato-alveolar.^^
(19)
We can further assume (following Sagey and Kenstowicz) that this structure can be used to represent [t-*] or [ ^ . Presumably the difference between the two is that for [t^], the I is read as vocalic information in an onset i.e. as a glide, whereas in it is often assumed that the manner features equivalent to stop, continuant are ordered on the same tier and result in the stop fricative interpretation characteristic of an affricate.
In the remainder of this thesis, I will assume that dentals and palato-alveolars are represented as follows:
17
Although Smith's interpretation, as already pointed out, interprets A and I differently.
This suggests a one-to-one reading of manner to place. I shall return to consider the derivation of in more detail in chapter 6.
(20) dental C palato-alveolar C A I I
If we adopt this representation of dentals and palato-alveolars, an interesting
consequence ensues. According to Lahiri and Evers (1991: 95) "Palatalised dentals often undergo a change of place and become palato-alveolars." Notice that a palatalised dental (given in (21) below) bears considerable resemblance to a palato-alveolar.^^
(2 1)
A I
In coastal dialects of kiSwahili, palato-alveolars become dentals as in * [n ^ ] —> [nti]
20 country.
19
Hume (1994: 145) provides the interesting example of Ngiyambaa in which dentaiised alveolars become palatal stops (described as postalveolar) [j ji] when immediately preceded by a front vowel.
20
4.2. Palatals.
Palatals in languages which exhibit them are uncontroversially represented by the I
particle, a representation which aligns them, equally uncontroversially, with front
vowels. But what is not so clear is why palatal place is highly marked and why languages prefer palato-alveolar to palatal. In languages with a four stop series the preference is for labial, alveolar, palato-alveolar and velar, rather then labial, alveolar, palatal and velar (see for e.g. Chomsky and Halle 1968).