3.4 A proposal for a feature set for SLN
3.4.3 Orientation
3.4.3.1 Introduction
In the first linguistic analysis of sign language phonotactics in 1960, Stokoe did not consider orientation to be a separate parameter of signs: it seemed at the time that the way the hand was oriented in space did not systematically distinguish pairs of signs (Stokoe 1960). To cover the few signs that were distinguished only by palm orientation, such as CHILDASLand THINGASL (Klima & Bellugi 1979: 48), Stokoe used a diacritic on the symbol for hand configuration in his notation system. In the literature since 1960, orientation has been systematically attributed the status of ‘parameter’ (Battison et al. 1975, Battison 1978, Sandler 1989).
Since the hand (or the fingers) are three-dimensional objects, specifying two of the three degrees of freedom fully specifies its orientation. This is illustrated in Figure 3.10.
that such an analysis by extension could be valid for other sign languages as well. Miller (1997) takes the opposite view from Brentari, making the strongest claims about the decomposability of movement in LSQ. Miller argues that all movement features, including shape and [alternating], can be represented as sequences of static values and/or are derived from prosodic structure in terms of feet, syllables and moras. In the absence of a comparable prosodic analysis of SLN, I will make no specific claims about the nature of movement features.
66 This section is partly based on work done together with Els van der Kooij; this was published as Crasborn & van der Kooij (1997). The two core claims of that paper, viz. that only one degree of freedom of orientation needs to be specified and that this specification is relative instead of absolute, are maintained in the present analysis. For sake of readability, I will use ‘I’ instead of ‘we’ in the discussion of those ideas below; see van der Kooij (in prep.) for similar discussion.
i. One degree of
orientation is fixed; the object can still rotate about one axis
ii. Both degrees of orientation are fixed
Figure 3.10
Orientation of three-dimensional objects
If the orientation of one axis is vertical, as in Figure 3.10i, the orientation of the top and bottom of the object is fixed. They will be facing up and down, respectively, no matter how the object is rotated about the axis. Rotation about this axis still causes four other sides of the object to change orientation, or in other words, rotation around two other axes. If the orientation of a second axis orthogonal to the first one is specified, as in Figure 3.10ii, than the orientation of the whole object is fully specified. The same applies to the orientation of the hand in signs: there are two degrees of freedom, and if both are specified, the whole orientation of the hand is known.
Orientation has been described in three different ways, which I will refer to as absolute, relative, and articulatory. Absolute orientation refers to the orientation of the hand in space, or, more informally, the direction that the sides of the hands are facing (Sandler 1989, Brentari 1990). These directions in space can be described in terms of up-down, left-right, and forward-backward. As the hand is a three- dimensional object, specifying the orientation of two sides of the hand describes its full orientation. As far as I know, without exception, and without explicit motivation, the palm side and the finger side have been chosen in sign language literature. Intuitively, these seem to be the two sides of the hand that we use most in daily life tasks as well as in many gestures of hearing people.
To give an example, the sign HOLIDAY in SLN would include in its lexical specification that the palm faces contralaterally and the fingers are pointing
upwards. This is illustrated in Figure 3.11. The sign is made with a B hand (all fingers extended and adducted) which touches the ipsilateral side of the chin twice.
Figure 3.11
Absolute palm and finger orientation in HOLIDAY
To facilitate the description for signs with different finger configurations, finger orientation has frequently been interpreted as ‘extended finger orientation’, describing the direction in which the fingers point in a sign if the fingers were extended (Greftegreff 1992). This is particularly useful for handshapes in which the selected fingers are curved in such a way that the fingertips point in the same direction as the palm (as in Figure 3.12ii and iii). Describing the orientation of the palm and fingertips would then not capture the two degrees of freedom needed to fully specify the orientation of the hand. Another way of characterizing this view of finger orientation is by saying that it is not the direction in which the fingers point that is relevant, but rather the direction in which (the distal end of) the metacarpal bones in the hand point. It is really the orientation of the hand that is described in those cases. In the illustrations in Figure 3.12 one can see that the extended finger orientation indicated by the arrow does not always point in the direction that the finger tips do: it does so only in 3.12i, where the finger is fully extended.
i. 1 ii. 1-curved iii 1-bent iv. 1-hook Figure 3.12
‘Extended finger orientation’ in different handshapes
The above way of describing orientation is termed ‘absolute’, because it refers to the ‘absolute’ three-dimensional space around us, without reference to any linguistic construct or property of the sign.67
Alternatively, orientation can be described by specifying how the hand relates to the place of articulation (Friedman 1976, Mandel 1981); this I will call relative orientation. For the sign HOLIDAY, one would then say that it is the radial (or thumb) side of the hand that is related to the place of articulation, namely the (left side of) the chin.
Surveying the literature, confusion is likely to arise: various terms have been used for relative orientation. Among these terms are focus, facing, and point of contact. Focus often refers to the part of the hand that is pointing in the direction of the movement (Mandel 1981). Facing often means the part of the hand that is oriented towards the location (e.g. Liddell & Johnson 1986), and point of contact is the part of the hand that touches the place of articulation (e.g. Mandel 1982, Liddell & Johnson 1986). These different concepts are partly overlapping in content. In the present model, the three aspects are all subsumed under relative orientation. As will become clear later on, the interpretation of relative orientation depends on the specified movement and location aspects of the sign.
A third way to describe orientation is in articulatory terms, by referring to the state of one or more joints of the arm. For example, Stokoe (1960) distinguished ‘prone’ from ‘supine’ states of the forearm, and these were the only distinctions in orientation that were included in his model. In HOLIDAY, the orientation could be
67 It appears to me that some authors implicitly refer to the orientation in ‘signing space’, rather than in the whole space surrounding the signer: it is unlikely that anyone would claim that the orientation stays constant when the signer shifts his upper body. Rather, the signs ‘shift along’ when the signer is lying down, for example, or when the upper body rotates left or right to take on different roles in discourse. Although this signing space is a linguistic construct, it remains an unstructured three-dimensional entity. The orientation values that are used are still ‘up’ or ‘forward’, and they do not refer to a specified part of the sign, such as a body location.
described as ‘neutral’, that is in between prone and supine. Nagahara (1988) drives this approach to its extremes, using articulatory parameters for all phonological features, including location. Apart from these two authors, articulatory descriptions of orientation have received little attention. More frequently, articulatory terminology has been employed in the description of repeated changes in orientation, such as wrist flexion (‘nodding’, e.g. Liddell 1990) and forearm rotation (‘twisting’, e.g. Liddell 1990).
In the phonological models proposed in the literature, absolute orientation has been the dominant perspective. In the few models where relative orientation has been proposed, this has always been in addition to a full absolute orientation specification, e.g. Liddell & Johnson (1989). The only place where relative orientation has been given emphasis is in Greftegreff (1992), in her discussion of indexical signs. These are signs in which the hand (generally, index finger extended only) points to a location in space which has been assigned a grammatical role. Greftegreff argues that variation in handshape of these signs, from variants where the index finger is fully extended to variants where it is bent to various degrees at all joints, makes untenable the position that it is the (palm and finger) orientation of the hand in space which is crucial in these signs. Rather, the pointing of the fingertip in the direction of the (grammatical) locus is what matters, allowing the actual orientation of the hand to vary from utterance to utterance.