Battison (1978), without doubt one of the most-cited works in the field of sign language linguistics, proposes two conditions which together constrain structural complexity in two-handed signs. These dual constraints, the Dominance Condition and the Symmetry Condition, are typically assumed to be phonological in nature and are at work in multiple sign languages. However, few studies have sought to determine whether these conditions are indeed phonological or rather consequences of other more general and non-linguistic constraints.A notable exception is Kita, van Gijn & van der Hulst (1998), who compare sign language with gesture in order to determine whether the Symmetry Condition is specific to sign language, and therefore phonological, or common to sign and gesture, and therefore by definition not uniquely linguistic, after all. The data that they use to answer this question, balanced two-handed signs and co-speech gestures collected from retellings of animated cartoon clips, come from two populations: signers using Sign Language of the Netherlands (NGT) and hearing non-signers speaking Dutch. Kita, van Gijn & van der Hulst find that signers and co-speech gesturers do not differ in adhering to the Symmetry Condition. The authors take this as evidence that the Symmetry Condition can be considered a universal and non-linguistic constraint, rather than a phonological one. However, the authors also investigate whether signers and co-speech gesturers differ concerning the articulatory distinctness of their handshapes (i.e. handshapes being categorical rather than gradient) and their use of body contact as a place of articulation, and find that the signers were less likely to make use of sloppily articulated handshapes, but more likely to make use of body contact, than the gesturers. Unlike the Symmetry Condition, then, the constraints that yield these regularities in sign language can be considered phonological. One of Kita, van Gijn & van der Hulst's (1998) contributions is to argue that some regularities in sign language structure are the result of abstract linguistic constraints, and others are the result of non-linguistic constraints.