No role for a feature [MCP] in the description of

The data used to investigate the potential contrast between signs with MCP joint flexion and signs without MCP joint flexion came from different sources. The main source is SignPhon (see §3.2.3 and Appendix B). Dictionary CD-ROMs made for educational purposes (NSDSK 1996, 1997ab) were also used. Both sources offered the opportunity to compare most of the signs in their citation form to a version of the same sign in context. we found a great deal of variation in MCP joint position when comparing these different contexts, that is, citation forms vs. forms in phrasal contexts. Because in most cases we could only compare the citation form to one instantiation of the sign in (a randomly occurring) context, we cannot make any significant generalizations as to the nature of the influence of context. However, the frequent occurrence of variation in MCP joint position across the lexicon, for both flexed and unflexed citation forms, demands an explanation.

91 _{There is another lexical item that is more commonly used as the noun 'car'. The sign I refer to here is} actually a classifier predicate with the meaning ‘a car moves forward’. For sake of ease I gloss it as CAR-

Both for signs with extended fingers (that is, without MCP joint flexion) and for signs with (full) MCP joint flexion in the citation forms, we find variation in the actual amount of MCP joint flexion in different realizations of the sign. In fact, the whole range of possible amount of flexion occurs, from –30 degrees (hyper- extended), as in PENGUIN, to almost 90 degrees (flexed), as in CALL. This contrasts with claims about other sign languages that there are two allophones of some handshapes, one with 0 and one with 90 degree flexion (e.g. Wallin 1996 for Swedish Sign Language).

Not only did we find variation in MCP joint position between signs in citation form vs. sentence context, but also variation in the realization of the MCP joint in different morphological contexts. The SLN compound sign PARENTS is composed of the signs FATHER and MOTHER. In citation form, sometimes both members of the compound are articulated with an extended index finger, as in Figure 3.39. In the citation form of the compound, however, the sign MOTHER is found with the index finger flexed at the MCP joint, whereas the part FATHER has the MCP joint of the index finger extended (see the illustrations in 3.40ii).

i. in isolation

ii. in the compound PARENTS (FATHER+MOTHER) Figure 3.40

MOTHER

Different morphosyntactic contexts can also give rise to different MCP joint positions. For instance, in the verb sign VISIT (Figure 3.41), a B-hand moves from a location near the locus of the semantic source to a location near the locus of the semantic goal, the fingertips pointing in the direction of the goal’s locus. If the goal

of VISIT is the first person, we always find MCP joint flexion, despite the fact that it is possible to touch the chest while bending the wrist, thus leaving the MCP joint position of the citation form unaltered.

i. 1-VISIT-2 ‘I visit you’

ii. 3-VISIT-1 ‘s/he visits me’ Figure 3.41 Two forms of the verb VISIT

In order to find out what might determine this variation and the specific MCP joint state in the citation form we examined signs that contain handshapes with MCP joint flexion in their citation form. we assumed that these signs in particular would be good candidates for an underlying specification of [base]. In total we found 225 different signs that occurred with bent handshapes, in either the SignPhon database or the Dutch dictionary CD-ROMs, or in both. (We did not consider handshapes

containing an aperture specification, although below I will discuss the influence of aperture specification on MCP joint flexion.)

Examining these citation forms, We found that other formational aspects of the sign seemed to determine the actual state of the MCP joints: in none of them is a phonological specification of MCP joint flexion needed to determine the phonetic surface form. we distinguish three factors, listed in (3.4).

(3.4) Factors explaining MCP joint flexion in citation forms 1. Aperture specification

2. Relative orientation and location specifications

3. The presence of semantic motivation for the articulator shape

I will discuss these three factors in §3.4.4.4. we consider these to be the three main factors responsible for the occurrence of MCP joint flexion in the 225 citation forms that we found. In establishing this set, we left out handshapes involving aperture specifications, because the opposed position of the thumb clearly sets them apart from the remaining cases. The distribution of these signs over the two other factors is given in Table 3.5. The distinctions within the categories will be discussed below.

selected fingers

space body relative orientation

total

shape delimiter semantic total rest 1 (n=95) 47 48 95 (100%) – – 2 (n=15) 10 5 15 (100%) – – 4 (n=115) 37 28 65 (57%) 18 18 36 (31%) 14 (12%) Table 3.5

Frequency of the different factors

From this table, we can conclude that relative orientation is an important factor for all sets of selected fingers, whereas semantic motivation only seems to play a role in signs with 4 fingers selected. In the category of 4-hands, 14 signs (12%) are put in the category ‘rest’. All of these signs containing a B-bent were either a variant of the C-hand, or the source of the variation could not be established (i.e. ‘free variation’).

Of the two main factors in Table 3.5, the role of relative orientation is most frequent. Moreover, it is also important in that it plays a role in two other situations. Firstly, I will argue that the apparent minimal pairs illustrated in Figures 3.37 and 3.38 above can be analyzed as having the same articulator specification, but different orientation and/or location specifications. Secondly, I hypothesize that the same line of reasoning that I use to explain the bent citation forms can be applied to predict the MCP position state of any sign in different contexts, whether sentence context (coarticulation) or sociolinguistic context.

3.4.4.3.2 Signs with MCP joint movement in their citation form

A few signs have a change in MCP joint position in their citation forms, while the other finger joints do not change state. At first sight, these seem to resemble signs with movement at the PIP/DIP joints, such as DEPENDENT, TOILET, or PHOTO, which have a contour for the feature [curved]. I propose that contrary to these changes in the PIP/DIP joint, the changes in MCP joint position are not phonological contours in feature values, but are in fact realizations of a path movement (that is, a change in location), following the hypothesis in van der Hulst (1995).92 _{An example of such a} change is the sign OFFER.

Figure 3.42

OFFER

Other examples are LICK, SEPTEMBER, AUGUST, LATE, WARM-FOOD, WARM- WEATHER, PAST.

Uyechi (1996) argues that in ASL, the sign PASTASL has three different forms: one with MCP joint flexion, one with wrist flexion, and one with elbow flexion. They seem to be in free variation. The same sign in SLN (also glossed as PAST) does have different meanings signaled by the different articulations, referring to the recent, intermediate, and more distant past, respectively. I do not consider these data to contradict our proposal. In many constructions in SLN, the size of the movement can modify the lexical meaning of the morphemes involved. In so-called classifier predicates, changing the size of the movement is used to express changes in size of the objects, movements, etc. that the signs refer to.

92 _{A similar proposal is made in Brentari (1998) to cover free variation of sign movements in general.} However, Brentari does not specifically propose that all base joint movements should be considered as reduced path movements, as van der Hulst (1995) does. She seems to imply that all handshape changes can be ‘phonetically enhanced’ by articulating them with joints proximal to the MCP joints, but in general she is not very explicit about what the different (reduced and enhanced) forms of a given phonological parameter look like.

Uyechi suggests that the three forms of PASTASL have three different phonological representations. They are, respectively, a handshape change, an orientation change, and a location change. My proposal makes it possible to see all of these movements as location changes, with constant location, relative orientation and finger selection specifications. What differs is not the phonological specification of the movement, but rather the size of the phonetic realization of this movement. Actually, this can be expressed in different ways, as the size of the moving articulator varies, as well as the size of the path that the endpoint of the articulator traces through space. The size of the movement is not only influenced by morphosyntactic operations, but also by extra-linguistic factors (e.g. Koenen et al. 1993). We can therefore formulate the hypothesis that different kinds of movements occur in different situations. This hypothesis will be further discussed at the end of this chapter, and tested in Chapter 4.

In Brentari (1998), a feature [base] is proposed for a full phonological analysis of handshape changes ASL. This analysis includes considerations of redundancy and predictability that are not taken into account here. Without a feature [base], but with the curving and aperture features proposed here, all handshape changes in SLN can be described. However, van der Kooij (in prep.) argues on phonological grounds that the feature [base] is not necessary for SLN at all.