1. Obviously, typologies may also be based on phonological or semantic/conceptual categories, but we will not discuss these aspects here.
2. Languages can also be partly flexible, for example, flexible at some level (e.g. clause level) but have strict order within some other domain (e.g. the noun phrase).
3. See Neidle et al. (2000) for discussion.
4. Notational conventions: As is common practice in the sign language literature, signs are glossed in English small caps. A hyphen is used when a single sign gloss consists of more than one English word (e.g. LONG-AGO). Subscript numbers represent points in the signing space used in verbal agreement and pronominalization (see Figure 3); ‘++’ indicates reduplication of a sign (e.g. in pluralisation). The sign glossed as INDEX is a pointing sign towards the signer or a location in the signing space. Depending on the context, it can fulfil the function of a locative adverbial (‘there’) or of a pronoun; in the former case, we gloss it as INDEX-LOC in the IUR examples. CL refers to classifier, and the subscript indicates which object is classified. A line above a gloss indicates the scope (i.e. onset and offset) of a particular non-manual marker (e.g. a negative headshake – glossed as hs).
5. Supalla (1986) refers to these as “semantic classifiers”.
6. Padden (2010) suggests that in different sign languages nouns referring to the same object may be lexicalized in different ways, based on their iconic properties. The phonological form of nouns referring to objects held by hand (e.g. a toothbrush or a comb) can be motivated either by how the object is handled or by shape properties of the object (the instrument). Six sign languages showed a clear preference for either the handling or the instrument pattern, but the preference was not as strong in all languages (ranging from 60-85%).