A Grammar of Sign Language of the Netherlands (NGT)

1.1.1. Contrastive handshapes

The handshape inventory of NGT consists of 31 handshapes that function as phonemes within the active articulator. All of these have (gradual) allophonic variants, and it is useful to first look at the following selection of handshapes that were identified in the pioneering KOMVA project (see also SOCIO-HISTORICAL BACKGROUND, Chapter 4). All these handshapes are indeed used in NGT, but they are not all phonologically distinctive. Still, obviously, there are many more possible phonetic handshapes; Table 2.1 below is thus not exhaustive.

The phonetic handshapes are categorized into the groups listed below, based on the classification of the Dutch Sign Centre.

i. Handshapes with all fingers closed to a fist;

ii. Handshapes with all fingers extended;

iii. Handshapes with all fingers curved or clawed;

iv. Handshapes with one (selected) finger extended;

v. Handshapes with one (selected) finger curved or clawed;

vi. Handshapes with two (selected) fingers extended;

vii. Handshapes with two (selected) fingers curved or clawed;

viii. Handshapes with three (selected) fingers extended.

Two things are important here: firstly, note that the thumb is not considered a finger in these categories and is treated differently (see also PHONOLOGY 1.1.1.1). Secondly, we use the term ‘selected finger(s)’ here, which will be explained in more detail in PHONOLOGY 1.1.1.1, but which makes a handshape such as r end up in a different group than l. The naïve observer may see two handshapes in which three fingers are extended, but it is important to consider which fingers are phonemically relevant, i.e., which fingers are part of the active articulator. Since in the former handshape, the middle finger is the selected finger, but in the latter handshape the three extended fingers are selected, these handshapes are categorized differently. More explanations will follow.

Table 2.1. The seventy phonetic handshapes that were identified in the KOMVA project (handshape images © Dutch Sign Centre).

1. Handshapes with all fingers closed to a fist
2. Handshapes with all fingers extended
3. Handshapes with all fingers curved or clawed
4. Handshapes with one (selected) finger extended
5. Handshapes with one (selected) finger curved or clawed
6. Handshapes with two (selected) fingers extended
7. Handshapes with two (selected) fingers curved or clawed
8. Handshapes with three or more (selected) fingers extended

Since not all handshapes in Table 2.1 are phonologically distinctive, some signs can be articulated with multiple of these handshapes. An example is the sign morning, which can be articulated with extended fingers and juxtaposed thumb without any space in between (the ]-handshape, Figure 2.1a), or with an almost identical handshape in which the thumb is extended (the x-handshape, Figure 2.1b), or with a thumb position that lies anywhere between these two positions, for example (Figures 2.1c and 2.1d). Still, the meaning remains the same; in other words, the signs shown in Figure 2.1 do not form a minimal pair:


a. morning	b. morning

c. morning	d. morning

Figure 2.1. The sign morning articulated with four different phonetic handshapes.

Van der Kooij (2002) developed a phonological model according to which these handshapes can be categorized to represent a phonemic active articulator. According to this categorization, NGT has 31 combinations of distinctive features. These combinations may be articulated in very different ways, of which some are predictable based on the phonetic or semantic context. This variation was accounted for by defining sets of so-called phonetic and semantic implementation rules. An example of a phonetic implementation rule is the following, which accounts for thumb position in fist-like handshapes: “in a fist, if the point of contact is [palm] the thumb is adducted (…)”. In practice, the default articulation of the combination of no selected fingers together with the feature [closed] results in a fist, i.e., a 6-handshape. However, whenever this fist makes contact with the location of the sign and the orientation is specified for [palm] (see PHONOLOGY 1.1.2), the articulation will result in a handshape in which the thumb is positioned at the side of the index-finger, as in this handshape: . Consider the sign for coffee (Figure 2.2), where, instead of rubbing awkwardly on the radial side of the non-dominant hand, the thumb is specified to be positioned next to the index-finger, for ease of articulation. Notice how the handshape of the non-dominant hand is still the default 6-handshape.

Figure 2.2. The sign coffee (Crasborn et al. 2020).

We consider it outside the scope of this grammar to provide and explain all of the implementation rules, but refer the interested reader to van der Kooij’s work. Keep in mind, however, that it is this set of rules which, according to this model, further differentiates and predicts some of the articulations of the 31 combinations.

This categorization of the 31 combinations is represented in the five tables below: the first one covers combinations of phonological features in which no fingers are selected, the second one shows combinations with all fingers selected, the third table displays combinations with one finger selected, the fourth table provides combinations with two fingers selected, and the last one shows combinations with three fingers selected. The numbers in the left column of each table represent the phonemic category which is characterized by the phonological features in the second column. Explanation of and support for the phonological status of these features is given in subsequent sections. The third column shows phonetic handshapes which act as a “stand-in” for one articulation of this set of features. Handshapes in the same row thus share the same phonological features, but note that many more phonetic variants are possible. Some handshapes (e.g. the 6-hand) appear multiple times; this means that this handshape is an allophone within multiple phonemic groups (in certain contexts). In Table 2.5, some possible handshapes are marked by a square; these articulations were not attested, and are further discussed below Table 2.6.

Table 2.2. Phonological combinations of features in which no fingers are selected, exemplified by possible articulations (© Dutch Sign Centre).

#	Phonological features	Possible handshapes (derived from the KOMVA overview)
1	[thumb]
2	[close]

Table 2.3. Phonological combinations of features in which all fingers are elected, exemplified by possible articulations (© Dutch Sign Centre).

	Phonological features	Possible handshapes (derived from the KOMVA overview)
3	[all]
4	[all], [open]
5	[all], [close]
6	[all], [curve]
7	[all], [wide]
8	[all], [wide], [curve]
9	[all], [open], [curve]
10	[all], [close], [curve]

Table 2.4. Phonological combinations of features in which one finger is selected, exemplified by possible articulations (© Dutch Sign Centre).

#	Phonological features	Possible handshapes (derived from the KOMVA overview)
11	[one]
12	[one], [ulnar]
13	[one],
14	[one], [ulnar],
15	[one], or	(respectively)
16	[one], or	(respectively)
17	[one], [curve]
18	[one], [open]
19	[one], [close]
20	[one], [open], [curve]
21	[one] [close], [curve]

Table 2.5. Phonological combinations of features in which two fingers are selected, exemplified by possible articulations (© Dutch Sign Centre). The handshapes marked by a square were not attested by van der Kooij (see further below).

#	Phonological features	Possible handshapes (derived from the KOMVA overview)
22
23	, [open]
24	, [close]
25	, [open], [curve]
26	, [curve]
26	, [curve], [wide]
27	, [close], [curve]
28	, [wide]
29	, [ulnar]
30	, [thumb], [wide]

Table 2.6. Phonological combinations of features in which three fingers are selected, exemplified by possible articulations (© Dutch Sign Centre).

Table 2.5 includes ten handshapes which were not attested. We consulted the online dictionary of the Dutch Sign Centre to check how many signs were categorized under these handshapes. For handshapes with very limited results (e.g. the two framed handshapes in group number 25), we checked whether other articulations with the same phonological features yielded any results, to verify whether this particular phonological combination would still be relevant in the above overview. We concluded that every combination of phonological features indeed occurs in NGT – see Information on Data and Consultants for further methodological information.

Van der Kooij also investigated the absolute and relative frequency of the feature combinations in her dataset. These data offer an insight into the distribution of combinations when articulated by the dominant hand. In Table 2.7, we provide an overview of the ten most frequent feature combinations in her data (of about 3,000 signs), supplemented by an overview of the ten most frequent combinations found in the NGT Signbank. The first column indicates the ranking. The following two columns show van der Kooij’s data, where the number of the phonological group is mentioned (see Tables 2.2-2.6 above), together with a handshape to represent this group visually, and where the absolute and relative frequency of this group within her dataset is given. The last two columns show the same for the NGT Signbank data (see also Information on Data and Consultants).

Both datasets show the same six most frequent handshapes, although in different order, but the handshapes taking positions 7-10 are slightly different: whereas does not appear in the top 10 from the NGT Signbank, and does not make an appearance in van der Kooij’s first ten handshapes, groups 1, 8 and 28 are still represented in both datasets, but in different order. The dataset from the NGT Signbank is based on more signers, and on signers from different regions, and, thus, more representative of the whole sign language community.

There are indications that the distribution of phonemic handshapes is slightly different per region (see also SOCIO-HISTORICAL BACKGROUND 4.4). For example, as will become clear in PHONOLOGY 1.1.3, the -handshape appears to be quite frequent in signs coming from the Sint-Michielsgestel region, while it is infrequent in signs from other regions. Additionally, the f-handshape appears to be more frequently used in Groningen. Unfortunately, systematic research into these regional differences is still lacking.

Table 2.7. The ten most frequent combinations of phonological handshape features on the dominant hand, based on van der Kooij and NGT Signbank (© Dutch Sign Centre).

Order

Van der Kooij (2002)

NGT Signbank (2020)

Combination of phonological features

Absolute (relative)

frequency

Combination of phonological features

Absolute (relative)

frequency

Group 3

( )

500 (18%)

Group 3

( )

585

(15%)

Group 11

( )

433

(15%)

Group 11

( )

419

(11%)

Group 7

( )

358

(13%)

Group 2

( )

295

(8%)

Group 2

( )

278

(10%)

Group 7

( )

285

(8%)

Group 19

( )

153

(5%)

Group 21

( )

182

(5%)

Group 21

( )

135

(5%)

Group 19

( )

173

(5%)

Group 3

( )

108

(4%)

Group 8

( )

164

(4%)

Group 1

( )

104

(4%)

Group 28

( )

145

(4%)

Group 28

( )

(3%)

Group 1

( )

133

(4%)

Group 8

( )

(3%)

Group 9

( )

132

(3%)

Total

2250 (80%)

2513

(67%)

We now turn to the phonemic status of the phonological features. To illustrate the phonemic status of sublexical elements, it is common to use (near-)minimal pairs. One of such pairs consists of the signs grey (Figure 2.3a)and green (Figure 2.3b), which are two signs that differ only in selected fingers: grey is signed with a B-hand, whereas green is articulated with a T-hand. The location and movement of the signs are identical.


a. grey	b. green

Figure 2.3. The minimal pair grey (a) and green (b), differing only in finger selection.

In the next example, the signs brother and also, also differ only in handshape. Whereas brother is signed with a Y-hand, also is signed with a w-hand. This time, however, it is not only the selection of fingers that differs (see PHONOLOGY 1.1.1.1) but also the finger configuration (PHONOLOGY 1.1.1.2), since brother is articulated with spread fingers, while alsois signed with adjoined fingers. Finger spreading thus is a distinctive feature in NGT.


a. brother	b. also

Figure 2.4. The minimal pair brother (a) and also (b), differing in finger selection and finger configuration (Crasborn et al. 2020, symbols added).