Referential shift

Finally, we turn our attention to the acquisition of referential shift which, as we learned in section 3.1.4.3 is determined by grammatical and discourse constraints. We also learned in that section that referential shifts represent a characteristic of sign language discourse and that they serve various pragmatic functions. Unfortunately, the development of this linguistic means remains largely unexplored in sign language acquisition research. First insights were obtained in a study of ASL narratives produced by deaf children aged 3-7 (cf. Emmorey & Reilly 1998). This study examined the use of reported quotation and reported action. The results obtained are summarised in the following (cf. also Table 3.10).

Table 3.10: Direct quotation and reported action in children’s productions (based on Emmorey & Reilly 1998).

Table 3.10: continued

Direct quotation. What can be gleaned from Emmorey and Reilly’s study is that 3-year olds (with the exception of one child) do not use direct quotation in their story retelling. While only one child produced a character’s facial expression, 5-year olds made no or an inconsistent use of shifted facial expressions. 7-year olds were found to use facial expressions much like adults. With respect to the linguistic means used to introduce direct quotation, Emmorey and Reilly (1998: 86) remark on the 5-year olds’ use of lexical means (that is, performative verbs such as the verb SAY) which is uncommon in adult narratives. As 7-year olds correctly use non-manual markers, the developmental pattern observed previ- ously^[1] becomes apparent here again: early lexical signalling is replaced later by non-manual marking, which is indicative of “a re-organisation and transition to incorporating affective facial expression into the linguistic system” (Emmorey & Reilly 1998: 86).

Reported action. 3-year old children were found to use reported action, however, the perspective adopted remains unclear which contrasts with the adults’ use. These children use affective facial expression but fail to identify the subject through lexical or pronominal marking. 5-year olds’ use of reported action, in turn, is characterised by an even distribution of the perspectives reported between the different characters (according to the authors, adults, in contrast, show a preference toward the use of one perspective; in the case of the picture story used to elicit the signed narratives, the so-called frog story, the boy’s

perspective was chosen). Finally, Emmorey and Reilly report that when 7-year olds’ used reported action, which they seldom did, their perspective choice differed from that of the adults (they chose the dog’s perspective). Emmorey and Reilly (1998: 87) remark on the discrepancy between the narratives of 7-year olds and those produced by adults and 5-year olds regarding the relative frequency of reported action predicates. According to these authors (1998: 87), the higher incidence of reported action predicates in the narratives of adults and 5-year olds (about 82% in the adult narratives) compared to the predominance of the use of the narrator perspective in 7-year olds’ productions patterns with the findings obtained in studies on narrative development in English and Hebrew showing that “7 and 8 year olds tell structurally more complex, but affectively more bland narratives than younger children.”

Crucially, Emmorey and Reilly (1998: 90) conclude that children master direct quotation before reported action. They assume that this difference is related to two factors. First, direct quotation is assumed to present “a single and coherent perspective” (that of the quoted character) whereas reported action involves a “dual perspective” (manual means reflecting the narrator’s and non-manual the character’s perspective): “That is, the signer as narrator chooses the verbs that describe the action. However, the facial expression is not that of the signer, but of the character whose actions are described” (Emmorey & Reilly 1998: 90). Secondly, direct quotation and reported action are assumed to differ in the possibility to use a lexical sign to introduce the referential shift to the extent that “the lexical sign say can be used to introduce quotation (in fact, five year olds use this mechanism to introduce a quote), but no lexical marker signals referential shift for reported action” (Emmorey & Reilly 1998: 90). Indeed, Emmorey and Reilly (1998: 90) speculate on the possibility that “[t]he availability of the lexical sign say may aid children in acquiring the use of referential shift for direct quotation.” Notice that the assumption patterns well with the observation that learners express several grammatical phenomena manually first before they use the target non-manual means.

• [1] In addition to the phenomena we noted previously, Emmorey and Reilly (1998) also mentionnegation, adverbial modification, and conditionals.