Reference forms and functions

The command of the skills involved in discourse organisation, including the appropriate choice of form-function combinations to designate referents, has been found to be the result of a protracted development (cf. Hickmann 2003, Karmi- loff-Smith 1983, Wigglesworth 1990, Morgan 2000). As we remarked in section 3.1.4.8 choice of appropriate reference forms involves not only syntactic knowledge but also pragmatic skills needed to determine the information status forms fulfil in a particular discourse context. As outlined above the functions served by reference forms correspond with a distinction of narrative contexts involving the same agent (maintenance), an agent that has been referred to before but who was not the agent of the event previously described (reintroduction), and reference of a new agent (introduction). The analysis of form-function relations in the data collected allows for the following observations.

Linguistic forms used. Our analysis of the reference forms used and the functions they serve from a narrative perspective reveals that the relative frequency of reference forms is strikingly similar across participants and files (cf. Table 3.49): subject drop occurs most frequently (between 49.1 and 82.1%), followed by NPs (with a frequency between 14.3 to 37.7%), with the use of article determiners (det_art) or pron typically making up the lowest rate (between 1.8 and 13.2%).

Table 3.49: Relative frequency of reference forms in the participants’ narratives.

Table 3.49: continued

However, the functions served by NPs, determiners and subject drop differ. While the distribution of forms corresponds roughly with the functions of introduction, reintroduction and maintenance respectively, participants’ usage differs markedly, and it changes over time (for an overview of the results obtained for each participant cf. Table 3.50 [first recording], Table 3.51 [third recording], Figure 3.16 and Figure 3.17).

Muhammed, for example, demonstrates a clear preference for overt reference forms in contexts that require an unambiguous identification, as it is the case in the reintroduction of protagonists (with the relative proportion of NPs serving this function amounting to 66.7% in file 1 and 78.9% in file 3). The other participants, by contrast, make a rather frequent use of subject drop in reintroduction contexts. As we can glean from Table 3.50, at the onset of the recording the proportion of subject-drop in reintroduction contexts ranges between 30.8% and 35.7% in the narratives of Maria, Fuad and Hamida, amounting to about 60% in the recounts of Simon and Christa. Interestingly, Table 3.51 makes apparent that the frequency of subject drop serving this function drops to 18.2% in Christa’s file 3 (with a percentage of NPs serving this function rising to 77.3). By contrast, the percentage of subject drop in Simon’s file 3 narrative remains relatively high (46.2%) (with the same relative percentage of NPs serving this function).

As indicated previously, the choice of subject-drop in reintroduction contexts is problematic where referential loci have not been established previously. It has to be noted in this context that the greater part of the problematic sequences identified in the data pertain to those that involve the main protagonist of the story (the boy),

which we might take as an indication that choice of reference form is also bound to choice of thematic perspective. We will take up this issue in the next sub-section.

Table 3.50: Reference forms and their referential functions in the participants’ file 1.*

* Expressed as a percentage of the total number of reference forms (proportions of forms used for respective function in brackets).

Table 3.51: Reference forms and their referential functions in the participants’ file 3.*

Table 3.51: continued

* Expressed as a percentage of the total number of reference forms (proportions of forms used for respective function in brackets).

Figure 3.16: Proportion of reference forms and functions in files 1 and 3 of Muhammed, Simon, and Maria.

Figure 3.17: Proportion of reference forms and functions in files 1 and 3 of Fuad, Hamida, and Christa.

Further to the quantitative measures undertaken about the distribution of reference forms and their functions, it is useful to consider the insights obtained in the qualitative analysis of the data about the participants’ orchestration of linguistic devices for the purpose of creating cohesion and coherence.

Consistent non-manual marking of POVs. With respect to the consistency in non-manual marking of POVs, Maria’s narratives stand out against the productions of other participants, in particular in her recount of narrative episodes in which referential shifts succeed each other. Recall, for example, her description of the boy and the dog looking at the frog, who, in turn looks at them. Perspective

changes are marked through a change in eye gaze direction as well as through changes in the modulation of the sign look. Referential identity is unambiguous because Maria associates referents with contrastive loci on the vertical axis, so that body lean forward and eye gaze directed toward the bottom mark reference to the boy (looking at the frog) and eye gaze directed toward the top of the sign space marks reference to the frog (looking up to the boy and the dog).

Sequential use of FRF-SRF. In our discussion we remarked that some participants adopt a neutral perspective before they shift reference to adopt the perspective of another character. For example, Fuad uses this strategy in his file 3. The alternate use of FRF and SRF corresponds with the “sequential structuring mechanism” described by Perniss (2007: 1324) whereby a “linear sequence of prototypically aligned representations” is used to map observer and character perspectives (which corresponds with our distinction of FRF and SRF):

Location and orientation information, represented in observer perspective with entity classifier forms, is followed by constructed action representations using the hands, face, and body in character perspective. In the character perspective representation, the depicted location and orientation information remains valid. A change in location and orientation must be indicated by a return to entity classifier usage in observer perspective. Achieving mapping in this way structures discourse sequentially with respect to the use of perspectives.^[1]

Reintroduction of protagonists via NPs (or other overt lexical means). The

comparison of early and later narratives of the participants in our study with respect to reference forms used shows that the use of full NPs clearly has the effect of reducing ambiguity, particularly, in narratives in which referential shifts are common and loci are not always established contrastively, as it was found to be the case in Hamida’s narratives. A preference for full NPs in the introduction and reintroduction of referents was also observed in other studies on sign language acquisition (cf. Morgan 2000) and in the narrations of oral language learners. The following excerpt is an example of a 5-year old hearing child’s narration of the frog story in English (Berman & Slobin 1994: 65):

(368) When the boy and the dog were asleep the frog jumped out of the jar. And then the boy and the dog woke up. The frog was gone. Then the boy got dressed, and the dog stuck his head in the jar. And then the boy opened up his window and called out for his frog, and the dog still had the jar on his head. Then the dog fell, and the boy was scared.

In his study on child and adult narrative productions in BSL, Morgan (2006: 325) (cf. section 3.2.3.4) remarks on the early frequent use of NPs by 4-6 year olds that reflects a focus on reference at the sentence level. While the choice of NPs by 7-10 year olds indicates a progress toward the adult usage, choice of NPs in 11-13 year olds still differs from the adult usage. The following two excerpts of narratives from 9-year olds make apparent how reference maintenance represents a challenge, in sign language and in spoken language, in the narration of a picture story in which several events occur simultaneously. Example (369) (from Berman & Slobin 1994: 70, [for further convenience we added the referents intended in brackets]) illustrates the type of ambiguity that emerges where pronouns are used without further indication on the changes of the agents involved. Interestingly, in (370) we find a sequence with a repair indicating the participant’s awareness about the failure to mark the change of perspective by using the same pronoun.

• (369) ... So they [they boy and the dog] went off to find his [the boy’s] pet frog. And he [the boy] looked in a hole, and the dog was chasing the beehive. It [the hole] was a home to a ground squirrel. And he [the boy] got his nose scratched. And the dog was still over playing with the bees.
• (370) . And then they start calling after the frog, and the dog’s looking at this beehive. Then some little gopher comes up and then the dog’s still looking at the beehive. So then the beehive falls and then the dog’s still looking at that beehive. So then the beehive falls and all the bees are - they start chasing after him, and he um - the little boy climbs up a tree and looks into a hole, and an owl flies out, and he falls off the tree.

Choice of a thematic subject perspective (the boy’s). The analysis of the narratives with respect to the reference forms used reveals that choice of a dominant or thematic perspective affects choice of reference forms and the functions they might serve.

Recall that in the frog story the boy and the dog are the two main protagonists. During their search they come across several other characters, such as bees, a mole, an owl, a deer, and the members of the frog family. Typically, these other characters are involved in the story for a limited narrative episode only (they are not reintroduced as characters at a later point in the narrative). By contrast, the boy and the dog are not only introduced and involved in a series of events; they are also reintroduced as protagonists after the description of events involving other characters. Hence, switches in perspective occur either from the boy’s or the dog’s perspective to the perspective of another character, or back

from another character’s perspective to the perspective of the boy and the dog. It is interesting to note in this respect that it is the latter type of switch that most often involves subject drop and, hence, referential ambiguity unless other means are used to secure the identification of the referent involved.

The choice of the boy’s perspective as the dominant perspective is a recurrent phenomenon in this corpus. This finding patterns with the results obtained in other studies on ASL and BSL discussed in section 3.2.3.4. Recall that the choice of a thematic perspective (Morgan 1999: 52) is defined by Morgan as the one that refers “to the signer’s use of a main or dominant perspective to report events, as contrasted with a secondary perspective.” Notice that this reference strategy differs from a “parallel representation” (Berman 2004: 269), which would involve a balanced account of the activities of the characters involved. In previous studies using the frog story as elicitation material adults were found to preferably choose one perspective (the boy’s) vis-a-vis 7-year old signers who chose the dog’s perspective, and 5-year olds who tend to evenly distribute the narrative focus on different characters (Emmorey & Reilly 1998; Morgan 1999). Morgan (1999: 51) also reports on the choice of boy’s perspective as the dominant perspective, leaving the dog as a secondary character. Typically, this phenomenon is reflected in the choice of subject-drop rather than the use of full NPs or pronouns to refer to the boy, also in those contexts where he is reintroduced as a character. Maria, for example, only introduces the boy once in her file 1 narrative via an NP. All subsequent activities involving this protagonist involve subject drop. Note that, at times, as is documented in Hamida’s file 3 example in (371) signers resort to a repetition of the proposition to disambiguate reference that might not be clear in a reintroduction sequence with subject drop, even if it refers to the main protagonist of the story.

While the perspective kept in discourse focus is not labelled overtly, overt reference forms are used to refer to the secondary perspective (Morgan’s 1999: 52). This strategy, used also by the participants in our study, can be considered to be an effective way of using reference forms “contrastively”, avoiding unnecessary explicitness where referential identity can be conveyed through other means.

On a more general level the preceding observations make apparent that the participants’ choice of reference forms significantly affects the cohesion and coherence of their signed narratives, which patterns with the findings obtained for oral language production. Crucially, what we can glean from the preceding observations is that the disambiguation of reference forms is equally a task in

signed and spoken narratives. This is an important observation given the persistent myths that continue to abound about the impact of iconicity on sign language discourse and the use of space in this modality of expression. Clearly, mastery of narrative skills, irrespective of the modality of expression chosen, involves not only full competence of linguistic devices available in a language but also knowledge of the functions they might serve in different communication situations. Once again, it becomes apparent that the orchestration of linguistic devices in discourse involves the integration of different types of knowledge.

• [1] However, Perniss (2007: 1327) also remarks that while this type of mapping is easy to understand and “highly informative with respect to the expression of spatial relationships” it is notthe most efficient strategy, when compared with a “non-prototypically-aligned construction thatmakes unique use of the possibility of simultaneity (...) as a discourse-structuring mechanism”(see ibid.: 1327f. for further details).