2.3 Compositional simplicity: a summary

Figure 2: Four-level composition of gestures.

The preceding analysies indicated that the complexity of CoGesT 1.0 can be decreased by careful attention to formal properties of the conventions. The tentative result suggests that it makes sense to distinguish between the following gesture types in the formalisation:

Microgestures:

The components of an iterative simplex gesture, each described in terms of a feature vector embedded into the trajectory position of a simplex gesture. There are constraints between the Source and Target positions of Simplex Gestures and their component Microgestures: The Source of the matrix gesture is identical to the Source of the first in a sequence of Microgestures, and the Target of the matrix gesture is identical to the Target of the last in a sequence of Microgestures; any intervening Target/Source positions require interpolation assumptions (e.g. equal length or other properties of Microgestures).

Simplex Gesture:

The ``normal'' gesture level, described in terms of a 9 position vector; metaphorically the Simplex gesture level represents the intuitively prototypical gesture level.

Gesture Pair:

A pair of gestures performed by both members of a limb pair such as the hands additionally requires specification of relations between the members of the pair:

1. ``same'';
2. ``symmetrical'';
3. ``spatially offset in direction $<a,b,c>$,'' from a ``stretched'' position through a ``normal shoulder-wide separation'' to a ``contact'' position and finally to ``folded'' position (for example intertwined fingers);
4. ``temporally offset in direction $\Delta t$'', if gestures in a pair are not entirely synchronous.

This would entail the specification of the gesture pair as a gesture triple $<relation, member_1, member_2>$.

Complex Gesture:

A concatenation of gesture pairs with cohesive function in dialogue. Concatenation is represented by `` $\hat{ }$ ''.

The grouping discussed here is visualised in Figure 1, and the four-level compositional structure of the representation proposed here is shown in Figure 2.

2.3.1 Analogy with segmental speech structures

An analogy can be drawn with the lower levels of analysis in contemporary approaches to phonetic analysis, in which the components of speech sounds are taken to be gestures (though generally in shorter temporal windows) Browman and Goldstein (1986):

Syllable:

complex gesture.

Dual articulation:

gesture pair (cf. labiovelars, injectives, implosives, ejectives, clicks).

Phonetic constituents:

simplex gesture.

Phase of a (consonant) phone:

Microgesture (cf. plosive and affricate phases, diphthongs, trill elements).

It may be speculated that Simplex Gesture level is, in semiotic terms, the fundamental gestural lexical item, perhaps most plausibly in the case of the emblematic type of gesture. This speculation needs further further evidence from the semantics of gesture before it can be firmed up into a working hypothesis.