Abstract
Whitely, C., & Colozzo, P. (2013). Who’s who? Memory updating and character reference in children’s narratives. Journal of Speech, Language, and Hearing Research, 56, 1625–1636.
In Word of Mouth 26:4 (March/April 2015), I reviewed an article by Barnes, Kim, and Philips (2014) on the ways young children, ages 4 to 6 years, introduced characters in the narratives they told and their listening comprehension for narratives. Proper character introduction was related to all outcome measures of narrative comprehension. Introduction of characters was considered proper when characters were introduced using an indefinite article (“a boy”), a proper name (“John”), or via a reference to a previously introduced character (“a boy and his mother”). Introductions with definite articles or pronouns (i.e., “the boy” or “he”) were not considered proper.
This study by Whitely and Colozzo (2013) investigated how typically developing children in kindergarten through second grade introduced characters in a story, maintained reference to characters, and reintroduced characters after the narrative focus had been on other characters. Furthermore, the research investigated the role of executive function/working memory (WM) in adequately referring to story characters. The authors expected individual differences in the context of continued development for both the WM updating process and referential abilities. The children completed two narrative production tasks and multiple updating tasks, with each updating task making more or fewer demands on the auditory–verbal and the visual components of WM.
Whitely and Colozzo explain that WM must be regularly updated and monitored when persons are engaged in language processing. They suggest that this updating is particularly relevant for narrative production, particularly the demands of clearly referring to story characters. At a given point, only some story characters are at the center of the narrative and are, thus, the focus of attention of both the speaker and the listener. Both parties must track the various characters throughout the story and presumably do so by creating and modifying a discourse model—a specific mental representation of an individual’s knowledge of the discourse experience (Brown & Yule, 1983). The discourse model allows the speaker to establish and sustain a clear reference to story characters throughout the narrative. How successful a narrator will be at clearly referring to story characters will likely depend on individual-level variables (i.e., task-related knowledge, processing resources) as well as characteristics of the narrative production task itself (e.g., length, number of characters) that could make it more or less demanding. Processing load influences which linguistic forms (i.e., pronouns vs. noun phrases) narrators use to refer to story characters. Adult participants were less likely to use pronouns in their oral narratives when they were also required to complete attention- or memory-demanding secondary tasks (Arnold, 2010). Adolescents with and without autism were less likely to use pronouns in longer compared with shorter clauses within their stories (Arnold, Bennetto, & Diehl, 2009). Hence, processing constraints will influence how a speaker refers to story characters. This could also affect whether a reference is clear.
Referential functions would likely be taxing of memory resources, particularly for children, who are continuing to consolidate and refine their narrative abilities alongside developing executive resources. The selection of an appropriate linguistic form (i.e., proper name; indefinite/definite determiner + noun; pronoun) to refer to a character will be constrained by whether or not it is currently or has previously been the focus of attention, which corresponds to different referential functions: introduction of a new character, maintenance of reference to a recently mentioned character, or reintroduction of a character who was no longer the center of attention (Bamberg, 1987). In each instance, the speaker must modify his or her own focus of attention, evaluate the familiarity of the intended referent for the listener, and then select an adequate referential form, all of which would draw upon the memory updating process.
The various referential functions could be differentially demanding of the memory updating process at a given point in development or when comparing individuals of different ages. Children ages 3 to 10 years have more difficulty accurately introducing and reintroducing characters than maintaining reference to the same character across successive utterances (Bamberg, 1987; Wong & Johnston, 2004). Maintaining reference would presumably be the least demanding referential function in terms of memory updating resources, as it involves continuous reference to a character that remains the focus of WM. In contrast, introduction and reintroduction require the speaker to evaluate the listener’s familiarity with the to-be-mentioned character and then select the appropriate linguistic form. Reintroduction may place particularly high demands on the updating process as a result of the greater distance between the referent and its identifier. In addition, in young children, difficulty with introduction is often associated with an undifferentiated use of the indefinite and definite determiners (“a” vs. “the”), a skill that may still be developing up to 10 years of age (Hickmann, 1980). Hence, linguistic and processing factors likely interact to make the various referential functions more or less taxing in terms of updating resources, particularly for children whose language abilities are continuing to develop.
The Study
Participants
A total of 68 typically developing, monolingual, English-speaking children from kindergarten through second grade participated. They completed eight tasks: four that assess the ability to update WM, two that tap Short Term Memory (STM) storage, and two narrative productions from wordless picture books.
Updating Tasks
N-back: 1-back and 2-back
The children completed two n-back tasks. The children watched a sequence of images that appeared one at a time for 2,000 ms in the middle of a computer screen, followed by 500 ms of blank screen. Each stimulus image consisted of one of 10 possible configurations of three white dots on a black background. The children were instructed to identify when the stimulus was identical to the target—that is, the image that had immediately preceded it in 1-back or that had appeared two images previously in 2-back. The children were required to respond by pressing the spacebar only when the image matched the target. The children completed the two tasks on different days (with 1-back preceding 2-back).
Sound monitoring
In the sound monitoring task, the children listened to sequences of nine sounds in random order: three birds, three bells, and three car horns. Each sound lasted 500 ms, followed by 1,500 ms of silence. The children had to monitor all three sounds at once and were instructed to respond by pressing the spacebar when they detected the third presentation of each sound.
Keep track
The children watched as a sequence of pictures appeared in the middle of a computer screen. Each picture belonged to one of six categories (i.e., body parts, toys, food, clothing, animals, and transportation) and was presented for 2,000 ms, followed by 1,000 ms of blank screen. Pictograms representing the current two (or three) target categories remained on the bottom of the screen throughout each trial to serve as mnemonic cues for the relevant categories. Once the sequence of pictures was complete, the child was asked to name the last picture presented in each of the two (or three) target categories. Each sequence included some pictures from the target categories and some foils. The stimuli consisted of 48 color pictures with one- or two-syllable early acquired labels spoken by 30 months of age by a majority of children based on (a) the MacArthur–Bates Communicative Development Inventories American English norms (Dale & Fenson, 1996) and/or (b) Morrison, Chappell, and Ellis (1997).
STM Tasks
Visual STM
The children watched as a matrix filled with equal numbers of red and white squares appeared on a computer screen for 2 s, followed by 2 s of blank screen and then a response matrix with one fewer red squares. The children had to point to the square that changed from red to white. The task began with one practice trial consisting of a 2 × 2 matrix and three test trials at this level. If the participant answered correctly on at least two of the three trials at any level, the size of the matrix increased by two squares (one red, one white), to a maximum of 30 squares.
Verbal STM
Verbal STM capacity was assessed with the Number Repetition–Forward task from the Clinical Evaluation of Language Fundamentals—Fourth Edition (Semel, Wiig, & Secord, 2003).
Narrative Productions
The authors used the wordless picture books Frog on His Own by Mercer Mayer (1973) and April Fools by Fernando Krahn (1974) to elicit narrative productions. The experimenter explained that the story would be recorded for a friend who would listen to the story later but would not be able to see the pictures. The physical setup and the instructions were designed to reduce the shared contextual information between the participant and the listener, and, thus, increase the likelihood that the children would be more precise when referring to the story characters.
Coding the Narratives
The authors identified and coded each reference to a story character on two dimensions: function and adequacy. They classified each mention of a story character into one of the following three categories: (a) introduction: the reference to a story character corresponds to its first mention by the speaker (e.g., “One day a boy went to the park”); (b) maintenance: the speaker continues to refer to the same character within an utterance, in successive utterances, or across an utterance (e.g., a descriptive statement) that does not advance the thematic progress of the story (e.g., “His name was Ryan. He had a dog, a frog, and a turtle”); or (c) reintroduction: the speaker refers back to a character who was previously introduced after having focused on another character in the intervening utterances (e.g., “The frog jumped out of the bucket and waved goodbye. But the boy did not notice”).
Each reference to a story character was judged for referential adequacy. Adequate reference indicates that the child used a linguistically appropriate device that would enable a listener to ascertain the character’s identity easily and without ambiguity. Inadequate references occur when a character’s identity is unrecoverable from the text. Examples include a character being introduced using a definite article, a pronoun that can refer to more than one character, and a child pointing to the book to specify the identity of a character.
Results
Conclusions
The study demonstrated that establishing and sustaining reference in narratives does rely on WM updating. There were moderate correlations between referential adequacy and three of the four measures of memory updating taken separately. Hence, children who performed better on all but the most difficult updating tasks were also more successful at keeping track of story characters and, thus, tended to refer to them more clearly.
Given the demands of the specific referential functions, the authors hypothesized that maintenance and switching (introducing and reintroducing characters) might differ in terms of the strength of their relationships with the updating process. The updating tasks (particularly sound monitoring and 1-back) significantly predicted accuracy of maintenance referencing but not switching. Switching was, however, predicted by verbal STM.
Both the maintenance and switching functions were challenging for the children. Both required that the children update the discourse model as the stories progressed. In this study, the referencing would have been challenging because of the length of the stories and the numbers of characters the children had to track. In addition, there were many instances in which multiple characters of the same gender were the focus of attention. This would have increased both the linguistic and executive demands for the narrator to disambiguate them for the listener. The children exhibited more difficulty with the switching functions, particularly with reintroduction of characters. The authors suggest that basic linguistic capacity and STM memory played an important role. Conversely, for maintenance, WM updating exerted a more important influence.
Considerable attention has been given to how children develop basic narrative skills. Numerous research articles and clinical material describe the development of basic narrative macrostructure, or what is termed story grammar structure. Many programs are available to teach children to identify and include macrostructure elements in their stories—settings, characters, initiating events, emotional responses, plans, attempts, consequences, and resolutions. Generally, the focus has been on single episode stories with few characters. This article addresses the cognitive and linguistic demands that children encounter when dealing with more complex narratives with multiple episodes and numerous characters. Speech-language pathologists will want to consider how to promote the linguistic and cognitive skills students need for building cohesive and coherent complex stories.