Development of coherence and cohesion
Between the ages of 2 and 5 children learn to produce several narrative genres, whereby the ability to retell past events in dialogue develops first, several years before the production of fantasy types of narrative is mastered (for a detailed discussion see Morgan 2000: 281). It is important to note that although the linguistic devices used in the different genres may be similar, the pragmatic functions they fulfil vary across narrative types. Children acquire the necessary pragmatic skills at different ages (Berman & Slobin 1994; Morgan 2000).
Among the few studies that have looked at the inter-relation of available grammatical devices and narrative skills in sign language acquisition are those that were undertaken by Bellugi and colleagues (Bellugi et al. 1990). The narratives elicited in this research were analysed with respect to verb agreement, pronouns and cross-sentential cohesion (Bellugi et al. 1990). Four developmental stages were distinguished (cf. Lillo-Martin 1999: 551):
Stage I. At around age 2, children’s productions consist of short isolated sentences, verbs appear in their uninflected form. Spatial syntax, at this stage, is absent. According to Lillo-Martin (1999: 551) word order is used to convey grammatical relations. Utterances produced at this stage are often incomplete, fragmentary (95), and elements (subjects or objects) might be missing (96).
Stage II. At age 2;6-3;6 verb agreement occurs with present referents. Pictures in the story books (that is, the characters depicted on the pages) are sometimes used as “present referents” (children are assumed to need a crutch for non-present referents). At this stage, children are able to describe isolated events, but their stories don’t cohere (Bellugi et al. 1990: 20). Children produce verbs with spatial location (e.g. spill-on-head) but fail to indicate the referential shift.
Stage III. Coherence is still absent at this stage. It is only toward the end of this period lasting from age 3;6 to 5 years that verb agreement with non-present referents becomes productive. However, children may stack locations (cf. example (97) which shows the use of the same locations for two different objects) or use them indiscriminately without maintaining consistent associations of loci with referents. Hence, sentences are grammatical but reference maintenance is not observed at the discourse level. Rather than using pronominals, children tend to repeat the name of the referent each time it is used (Bellugi et al. 1990: 21). Children do not use the shifted referential framework at this stage.
Stage IV. Co-reference is achieved at this stage (age 5-6). The children’s cross-sentential use of verb agreement and anaphoric pronouns is accurate. Verbs that need to be marked for agreement appear in their inflected form. The example of a story provided in (98), produced at age 6;2, illustrates the availability of the relevant mechanisms. Notice that verbs are correctly marked for agreement, and that the referential loci for mother and girl are established through verb agreement. The construction with the verb SPILL correctly involves a referential shift (showing body location for the water spilling).
We close this section with a note on the caution imposed on the interpretation of data when it comes to the properties involving the syntax-discourse interface. As pointed out by Schick (2006: 119) “children do not have difficulty in understanding the concepts that underlie the abstract use of space to represent people and events, especially when these spatial maps are richly grounded in reality.” However, as this author remarks (2006: 119), “early evidence of the use of frames of reference and role shift during narration does not translate into early mastery”. Apart from the complexity of the task of integrating information from different levels of linguistic analysis for narrative purposes, Schick also remarks on the potential relevance of cognitive factors when it comes to the representation and production of narratives.