Abstract
This study examines the expression of simultaneity in the film-based oral narratives of 100 English monolinguals in the following three age groups: preschoolers (4–6 years), school-aged children (7–10 years), and adults (19–48 years). Participants told a story of what happened in the film, in an off-line task, to an interviewer who had not seen the film. The film was rich in simultaneous events at various sites through the episodic structure. Focus was on quantitative and qualitative aspects of simultaneity, from the perspective of forms and functions. Quantitative results showed very little simultaneity at preschool and almost similar expression at school age and adulthood. Qualitative analyses revealed that perceptual, semantic, and discourse factors affected the profiles of development. Preschool children expressed local simultaneity between situations in adjacent clauses, more frequently between unbounded situations that are implicitly simultaneous. Besides, they tended to express more simultaneity in scenes that were perceived in a single screen shot. From age 7, children became more able to express simultaneity across larger stretches of the text, covering a wider scope on situations on parallel timelines. Top-down knowledge of narrative organization guided older narrators to take temporal perspectives that go beyond the semantic properties of events, giving way to discourse-motivated simultaneity where causality plays a substantial role. Language forms to express simultaneity showed a long developmental route – through verb semantic and tense-aspect alternations as the widest, basic usage to specific lexical forms like conjunctions (e.g. while), adverbs (e.g. meanwhile), and more sophisticated syntactic configurations. The form-function analyses enabled an exploration of the cognitive and language abilities in the production of simultaneous relations under the constraints of narrative discourse organization. The study reinforced the results of previous developmental studies of temporality, shedding further light on the relatively unexplored topic of simultaneity expression.
Introduction
Narratives are linguistic reconstructions of events as they unfold in a real or imaginary world, structured in temporal and causal ways (Bruner & Lucariello, 1989; Mandler & Johnson, 1977). Narrators provide meaning to these events by adopting a perspective that assembles action and the minds of the characters together in a coherent whole (Bamberg, 1991). Narratives are organized as unified episodes that combine sequential, plot-advancing events with non-sequential descriptive and evaluative material, yielding various temporal textures (Hopper & Thompson, 1980; Labov & Waletzky, 1967; Reinhart, 1984). For example, a narrator might comment on how much she disagreed with a character’s motivation in the middle of narrating the story, thus suspending the storyline. This article examines the ability of narrators to take temporal perspectives on events from their subjective point of view – from which the events become related with each other in specific ways to form a coherent whole (Bamberg, 1994). Specifically, we focus on the expression of simultaneous situations, namely those that are visually perceived or can be inferred as temporally simultaneous, in a film-based narrative. 1
Expressing simultaneity across sentences in discourse
The temporal framework adopted here for representing temporal simultaneity is based on the notion of ‘reference time’ as it has been proposed by Reichenbach (1947), and later extended to account for the interpretation of temporal relations across utterances in extended discourse (Klein, 1994; Smith, 1991, 2004, 2007). According to Smith (2007), ‘in narrative, events and states are related to each other temporally. The first sentence establishes a time, actual or fictional; after that, events and states are given in sequence or they are simultaneous with previous times in the text’ (p. 425).
2
This relational system operates on two levels: on one level, a temporal relation is established between Speech Time (SpT) and a reference time (RT) from which a certain situation is presented. This relation is mainly – but not exclusively – represented by the linguistic category of ‘tense’ which locates a situation with respect to a certain reference point (Comrie, 1985). Thus, speakers can locate situations on intervals of time that coincide with the time of speech or they can anchor situations on intervals of time that precede or are posterior to speech time. On another level, a temporal relation is established between RT and the time span that a certain situation occupies – referred to as ‘situation time’ (SitT). This relation reflects various ‘aspects’ or ways of looking at situations, which, in turn, has implications for how situations are related to each other in the thread of discourse, as further elaborated below. The short narrative excerpt in (1) – where S holds for ‘situation’ illustrates how this relational system works: (1)S1 Mary
The situations in (1) occur in a past time, as typical of narrative discourse and events are assumed to be reported in the order in which they occurred, according to the Principle of Natural Order (PNO; Labov, 1972). S1 is anchored in SpT and provides the reference point for the next situations. S3, S4, S5, and S6 – like S1 – are included in the time span provided by RT and each introduces a new RT that advances the narrative time. S2 and S7 include the RTs of the surrounding situations, thus, they surround the preceding (and subsequent in the case of S2) situations and do not advance the narrative. Within this theory of temporal interpretation, the aspectual properties of situations and their realization through language play a major role. Following Smith (1991, 1999), aspect includes two components – situation type and viewpoint aspect – both assigned at the clause level. Situation types are conveyed by the verb and its constellation in the sentence and they can be classified as events or states according to the temporal features of dynamicity, boundedness, and duration (see also Dowty, 1979, 1982; Vendler, 1967). Events are dynamic and include the following three types: activities (e.g. paddling along the river, dancing) which are durative and unbounded since they have no intrinsic endpoint and do not imply a change-of-state; accomplishments (e.g. paddling to shore, building a house) which are durative and bounded since they have a terminal point provided by the verbal argument or any other contextual clue; and achievements (e.g. reach the top of a mountain, fall asleep, open the door) which are non-durative, bounded changes of states. States are not dynamic; they are unbounded and durative. In (1), S1, S3, S4, S5 and S6 are bounded while S2 and S7 are unbounded and thus simultaneous with the preceding or following RTs.
The second component of aspect – viewpoint – refers to morphological presentation of the verb as perfective or imperfective, depending on the available means in a certain language. In English – the language studied here – the simple form presents a perfective viewpoint, which focuses on the endpoint of the situation, while the progressive form implies an unbounded situation viewed from an internal perspective that focuses on the time interval rather than on the initial or final points of the situation. 3 The interaction between situation type and viewpoint aspect determines the full aspectual constituency of the situation expressed in the sentence (Smith, 1991; Vendler, 1967) and affects its relational location along the timeline. This can be illustrated by comparing S2 with S7. S2 is a state interpreted as simultaneous by default (unless explicitly marked otherwise) even if marked in the past simple. 4 In narratives, states are simultaneous to all or part of the events in the storyline (Comrie, 1986; Kupersmitt, 2015; Labov, 1997), and they generally occur without an overt grammatical or lexical marker of simultaneity (Schmiedtová, 2004). S7, in contrast, is an achievement that would acquire a sequential interpretation by default (Borik, 2006) if not grammatically marked in the progressive aspect. As shown in (1), S7 overlaps with a narrower scope of situations, most likely with S5 and S6. The aspectual constituency of situations has a critical role in the distinction between narrative foreground and background (Berman & Slobin, 1994; Hopper & Thompson, 1980; Reinhart, 1984). The foreground stands for sequential, plot-advancing material, encoded by bounded situations in the perfective aspect. The background, in contrast, suspends the narrative time – as in the case of simultaneity – and is typically encoded by unbounded situations in the imperfective aspect.
Beyond the clause level, the temporal interpretation of situations depends on the arrangement of predicates in syntactic configurations of isolated, coordinated (with same or different subject), or subordinated clauses (Berman, 2008; Kupersmitt, 2015; Matthiessen & Thompson, 1988). Thus, the temporal relations between the situations ‘S3’ and ‘S4’ can be manipulated by adding different subordinating conjunctions. Adding ‘when’ before S3 (e.g. When she walked into the kitchen she drank a glass of wine) will enhance a sequential or nearly simultaneous interpretation (Hinrichs, 1986; Smith, 1999) but adding ‘before’ (e.g. Before she walked into the kitchen she drank a glass of wine) will change the order of the situations and RTs. Time adverbials can change the temporal interpretation of connected situations too. Adding the adverb ‘meanwhile’ before S7 reinforces the simultaneous interpretation but an adverbial like ‘five minutes later’ would imply that the time of that situation follows the RT of S6. Finally, world knowledge and contextual information provide additional implicit support for a certain temporal interpretation and play a significant role in the ability to understand and express temporal relations (Caenepeel & Moens, 1994). It is likely in (1) that S4 occurred after S3 unless specified otherwise, by means of viewpoint aspect or explicit lexical means, as shown earlier. 5 In sum, expressing simultaneity is about manipulating temporal points and intervals of time along an abstract timeline, by means of various linguistic systems supported by discourse-organization abilities and world knowledge.
The development of understanding and expressing simultaneity
The development of time concepts and the ability to understand and express temporal relations follow a long developmental route, from early to middle childhood, and rely on interacting cognitive and linguistic processes (Aksu-Koç, 1988; Berman & Slobin, 1994; Weist, 1986; Zhang & Hudson, 2018). Part of the reason for the long developmental trajectory is that time is an abstract domain represented in human cognition through experience and language – it cannot be sensed or perceived as compared to, say, space (Evans, 2004). Piaget (1946) argued that children’s conceptualization of time was abstracted from their understanding of space and was tightly related to the understanding of other event dimensions such as duration, speed, and distance (Friedman, 1978). Among temporal concepts, he claimed, simultaneity shows a particularly long developmental trajectory, probably because children’s understanding of simultaneity was based on their sensorimotor experience. For example, young children were likely to judge two situations as sequential in time, if they had to turn their heads to view them (Piaget, 1946). Only middle childhood or older children could abstract away from their own perspective and correctly judge two situations as simultaneous. Similar results came from studies in which children aged 3–7 were asked to act out the sentences read aloud by the experimenter using various toys. The results showed a slow progression in the correct acting out of simultaneity, which did not reach ceiling even at age 7 (Feagans, 1980; see also Natsopoulos & Abadzi, 1986, for similar results).
More recent language-oriented research has suggested that the developmental origins of simultaneity may start earlier than Piaget thought. In a number of studies performed by Weist (see review in Weist, 2014), he provides cross-linguistic evidence for a claim that children move from an event time temporal system at age 3 to a more abstract RT system at age 5, where reference can be detached from speech time. At this developmental phase, they start using aspectual contrasts and varied tense-aspect forms combined with lexical and syntactic strategies, showing a gradual move from a mono-referential to a bi-referential time system from age 4 and beyond. At this age, children have started to develop a decentered temporal system where events (as analogous to objects) can be located at various arbitrary points in time, and can be temporally related with respect to the other independently of the ‘here and now’ deictic perspective (McCormack & Hoerl, 1999). Taken together, these studies suggest that by age 5, children have acquired a mature temporal cognition that involves three types of knowledge: (1) an abstract temporal framework for event-independent thought along a time line, (2) a concept of duration so that time can be abstracted as a dimension, and (3) an understanding of RT, where the time of referring can be detached from the time of the event (Droit-Volet & Coull, 2016; McCormack & Hoerl, 2017; Pawlak et al., 2006; Weist, 2014). Once these notions have emerged, the road is paved for the understanding and production of simultaneity relations.
Children’s learning of the language expressing simultaneity also develops slowly across childhood, starting from not expressing simultaneity at all at 2–3 years of age to using aspectual morphology and/or verb semantics to even later using more complex explicit means available from the specific language (Aksu-Koç & von Stutterheim, 1994; Bronckart & Sinclair, 1973; Flecken & Schmiedtová, 2007; Hickmann, 2003; Morgan, 2002; Winskel, 2007). Children acquiring a language with obligatory aspectual distinctions (e.g. Polish, Spanish) use them earlier than in languages with more optional contrastive forms such as English (Weist et al., 2004). Winskel and Luksaneeyanawin (2009) found that English-speaking children aged 6 and 7 used more connectives to express simultaneity relative to their Thai-speaking peers. However, even after children start expressing simultaneity, their use of explicit linguistic means remains scarce up to at least the age of 10 years (Hickmann, 2003; Winskel and Luksaneeyanawin, 2009). Among connectives referring to simultaneity, English-speaking children first produce the less restricted when, using it with both punctual and durative predicates at age 4, whereas while appears by school age and the more contextually constrained as occurs only among adults (Silva, 1991; Winskel, 2004).
Expressing simultaneity in narrative discourse
The forms and functions of simultaneity continue to develop from preschool through school-age to adulthood, as these become integrated with semantic and syntactic systems in extended discourse. In their pioneer cross-linguistic study on narrative temporality, Aksu-Koç and von Stutterheim (1994) found that 3-year-olds express simultaneity deictically from the frame of speech time. In languages with rich aspectual distinctions like English, Spanish, and Turkish (Aksu-Koç, 1988; Kupersmitt, 2015; Palmer, 1974), young children used aspectual shifts in juxtaposed constructions with and to contrast between durative states or activities with more punctual events, at the local level of the narrative. Overall, younger children tended not to mark simultaneity relations explicitly, leaving the reader to infer these relations from the temporal properties of situations and world knowledge. From preschool imperfective and progressive forms served for backgrounding purposes in subordinated constructions to express simultaneity in the temporal axis of the discourse, but only after age 9, children mark simultaneity by means of specific conjunctions and adverbials. At this age, the function of simultaneity serves more global purposes of temporal organization as it occurs in complex packages of events and states that represent a unified scene (Aksu-Koç & von Stutterheim, 1994; Berman, 2014; Hickmann, 2003; Morgan, 2002; Winskel, 2007).
In a study performed with film-based narratives produced by children aged 8–12, Bocaz (1987) corroborated the protracted route to expressing simultaneity, claiming that only after age 10 children could express simultaneous relations in more complex structures that combined various explicit language forms. More recently, in a study of narrative temporality in personal narratives produced by speakers of Spanish, English, and Hebrew from age 9 to adulthood, Kupersmitt (2015) found a preference to express unmarked (i.e. implicit) simultaneity with states at all ages and languages. Although both implicit and explicit simultaneity increased with age, only the latter was more frequent in languages with morphological aspect and with available lexical and syntactic means to express this function.
In sum, understanding and expressing simultaneity show a long developmental trajectory through childhood. To express simultaneity linguistically, children must understand it conceptually. Most studies have focused on the development of simultaneity in the context of sentences (e.g. Weist et al., 1999; Winskel & Luksaneeyanawin, 2009) or in conversational contexts (e.g. Uccelli, 2009; Weist, 1986). In this study, we focus on the context of narratives, which offer the child a stage to take temporal perspectives that ‘involves mentalizing abilities’ (McCormack & Hoerl, 1999, p. 174). In this sense, the child must learn to deviate from the canonical sequence of events in the narrative to present simultaneity relations that serve various discourse functions not only at the local level of the narrative (e.g. contrasting a durative and a punctual event), but also at the global level of the episode and beyond its boundaries – by merging the action and mental levels of the story. This is a demanding task that relies on cognitive, pragmatic, and linguistic abilities together (Berman, 2014).
The present study
This study compares the expression of simultaneity in narratives produced by monolingual speakers of English in preschool and school-aged children and adults, using a short film as stimuli for narrative production. It is assumed that the preschool participants have an emergent RT system at the sentence level (Weist, 1986) and they have mastered the acquisition of tense-aspect forms (Berman & Slobin, 1994), as well as less constrained simultaneity conjunctions such as when (Winskel, 2004), complex syntax such as complement and adverbial clauses (Aksu-Koç & von Stutterheim, 1994; Hickmann, 2003), and temporal adverbials (Weist & Zevenbergen, 2008). They also have emergent abilities of arranging a sequence of events with causal relations and at least simple episodes (Berman & Slobin, 1994; Botvin & Sutton-Smith, 1977; Kupersmitt, 2016). Thus, the study focuses on mastering already acquired forms for various temporal and narrative functions and on diversification of forms for existing functions.
Unlike previous studies that used picture books (Aksu-Koç & von Stutterheim, 1994; Winskel, 2007), or video clips in studies with adults (Carroll & Stutterheim, 2003; Schmiedtová, 2004), this study uses a film suitable for all ages depicting a structured story that is rich in simultaneous events. The context is genuine to examine the development of simultaneity for discourse and communicative purposes beyond the sentence level, and it should contribute to understanding various discourse factors that may be related to its expression. Rather than verbalizing or interpreting neatly sequential versus simultaneous events as in experimental settings, the task for narrators is to integrate sequential with simultaneous events to form a coherent whole (Bamberg, 1994).
The off-line retelling task is adequate to examine how narrators frame events in an abstract referential time without being bound to speech time reference – and is different from retelling tasks while watching a film or looking at pictures, as performed in previous studies (Berman & Slobin, 1994). The challenge that narrators face is how to move the frame of reference away from the ‘here and now’ (Weist & Zevenbergen, 2008), how to represent and arrange the situations into a relational temporal system where they are linked to moving RTs and space, and how to perspectivize them through language. Against this background, the study has the following three aims: (1) to investigate the developmental profiles in presenting simultaneous situations in quantitative and qualitative terms, (2) to evaluate the discourse factors associated with narrative-embedded simultaneity, and (3) to follow the trajectory of linguistic forms used to express simultaneity.
It is predicted that the development of simultaneity will reflect the ability of children to manipulate various temporal perspectives from a RT that is different from speech time. At preschool, simultaneity will be implied by means of unbounded states and activities that acquire a simultaneous interpretation by default. Older children and adults will show an increasing ability to express simultaneity explicitly with various situation types including accomplishments, which duration needs to be estimated for the purpose of establishing a temporal relation of simultaneity.
The properties of discourse construction will play a role in simultaneity expression – thus, younger children will express simultaneity at a local level between two adjacent clauses, while older children will become gradually able to perspectivize various situations as simultaneous across larger stretches of the text (Aksu-Koç & von Stutterheim, 1994). Besides, older children are expected to express simultaneity along a larger number of episodic components for backgrounding purposes within and across episodes, or for advancing the plot from the dual perspective of both protagonists, as reflecting the discursive pattern of the story. Given that younger children are expected to experience some difficulty interpreting and estimating the duration of certain situations, they will tend to focus on their initial or terminal points rather than on processes occurring simultaneously, particularly for situations presented in non-continuous scenes. Older children and adults, instead, will be able to present more specific simultaneity in more complex scenes that demand a higher level of abstraction to infer and represent multiple referential anchors.
In terms of linguistic profiles, simultaneity will be first expressed implicitly by means of situation type, thus, its interpretation will rely on contextual cues and world knowledge. Since young children use tense and aspect shifts to indicate boundedness versus unboundedness early on in development (Bronckart & Sinclair, 1973; Winskel, 2007), these strategies are expected to appear before more complex lexical and syntactic means to express simultaneity (e.g. while), which are more demanding both cognitively and linguistically. Older narrators will combine various linguistic strategies in more complex configurations for the purpose of expressing simultaneous relations.
Methods
Participants
In total, 100 English monolinguals participated in this study in three age groups: preschoolers (4–6 years), school-aged children (7–10 years), and adults (19–48 years). All participants were living in a majority-English-speaking part of Canada. The children had minimal knowledge of any other language than English (e.g. could count to 10 in Spanish). Some of the adults had taken foreign language courses in university but could not speak any language other than English well enough to carry on a conversation. Table 1 summarizes some of their background characteristics.
Background characteristics of participants.
Materials and procedures
All participants watched two video clips, each lasting 4 minutes, and each from a different Pink Panther cartoon, consecutively and without a break. In this study, we focus on the first clip that they watched, a clip from In the Pink of the Night. The story in the film is structured around a canonical plot that includes an initiating event (IE) or problem (PR) that triggers a series of episodes comprised of a Goal-Attempt-Outcome (GAO) structure followed by a resolution (R) (Berman, 1988; Stein & Glenn, 1979; Trabasso & Rodkin, 1994). The characters react in emotional ways leading to various internal responses (IR) that motivate their subsequent actions. These components are described below, from the perspective of the panther:
Setting (Set): Pink Panther is sleeping and the cuckoo clock he had installed on the wall of his bedroom keeps ticking.
IE: the bird starts cuckooing to wake the Pink Panther up. Pink Panther is annoyed (IR).
GAO I: a superordinate structure that includes various unsuccessful attempts (A) of the panther motivated by his goal (G) to shut the cuckoo up and go back to sleep. Each negative outcome (O–) brings about an emotional reaction. As the panther is furious (IR), he finally goes to the bridge, throws the cuckoo clock into the river, returns to his house, and goes back to sleep (O). Meanwhile, the cuckoo floats up, saves himself, and paddles back to shore.
GAO II: Pink Panther cannot sleep. He thinks the bird is drowning (PR) and he feels remorseful and guilty (IR). Following his superordinate goal to save the bird (G), he tries a series of attempts – goes back to the river, looks over the bridge, dives around searching for the bird (A), but the bird has dragged the cuckoo to the panther’s home and settled back into its clock on the wall. In sorrow (IR), and having not found the cuckoo (O), the panther throws flowers into the river to mourn for it (R).
GAO III: Pink Panther falls asleep (Set) and the cuckoo starts cuckooing to wake him up (IE). Pink Panther is overjoyed to see the cuckoo (IR). While the two sleep together (Set), a mechanical alarm clock rings (IE) and the cuckoo bird smashes the clock with a wrench (GAO). Pink Panther pats the cuckoo on its head, and they happily go back to sleep (R).
The participants were asked to tell a native English-speaking interviewer what they had seen in the clip. The interviewer explained that she or he had not seen the cartoon. The narrations were videotaped for later transcription.
Coding
The participants’ narratives were transcribed in orthographic words and the stories were divided into clauses (Berman & Slobin, 1994, pp. 660–661) as a basis for the analysis of simultaneity, which included various types of overlap between situations (Schmiedtová, 2004). In total overlap, two situations completely co-occur along a time axis (e.g. all this time he’s looking around the ocean the bird paddles his way back in [HA, adult]). In partial overlap or inclusion, one or more situations are included or contained within the time axis of the other and their temporal boundaries do not coincide (e.g. while he’s at the river the bird got out of the river and went out to his house [CAM, adult]). Total overlap and inclusion (with durative or punctual situations) have been classified by some authors as ‘prototypical simultaneity’ (Broccias, 2006; Declerck, 1997). Coding also included cases of partial overlap between two situations, where either initial or final boundaries coincide (e.g. the clock was ringing, and the bird grabbed a huge wrench and smashed the clock [SW007, age 10]) or where the overlapping points or intervals are open to interpretation (e.g. he’s throwing flowers into the river while the bird is back at home [ALM, adult]).
The discourse features and the linguistic expression of simultaneity were coded out of pairs of simultaneous situations, as illustrated in the examples below, where clause boundaries are marked by #: (2) He was sleeping # and the cuckoo clock went off. (ATM, adult) (3) All this time he’s looking around the ocean # the bird paddles his way back # and goes # and sets his house back up in the Pink Panther’s house. (HA, adult) (4) He’s feeling remorseful # and he imagines the cuckoo # drowning. (WBM, adult)
In example (2), there is one pair of simultaneous situations between two adjacent clauses. In (3) and (4), the pairing is more hierarchical, involving a larger number of clauses. In (3), the situation of ‘looking around the ocean’ is simultaneous with three sequential situations ‘paddling back, going and setting his house back’. In (4), there are two pairs of simultaneous situations, one in two adjacent clauses between ‘imagines’ and ‘drowning’ and the other between non-adjacent clauses. These relations are depicted in Figure 1.
Pairs of Simultaneous Situations in Adjacent and Non-Adjacent Clauses.
The analytical approach proposed here considers top-down and bottom-up levels of narrative production (Karmiloff-Smith, 1979). At the top-down level, the study focuses on the perspectivation of events as simultaneous and on the semantic and discourse factors that affect this specific temporal perspective according to the following criteria:
Type of situation. Simultaneity could occur between different types of situations classified according to their aspectual constituency determined by combined lexical and viewpoint aspect (Smith, 1991; Vendler, 1967), as described in the introduction, further illustrated in the next examples by a preschool and a school-age narrator:
(5) And then, and then he, and then, and then the cuckoo (6) And while he
In (5), the underlined situations involved in an implicit relation of simultaneity are unbounded. In (6), however, only the first situation is unbounded, while the last three – arranged sequentially but simultaneous to the first – are bounded and durative. Besides, for bounded situations, imperfective viewpoint aspect was coded to address the degree of flexibility in combining these two parameters in simultaneity expression, as occurring in (6), where the bounded situation ‘go to look for it’ – is marked in the progressive. Type of situation was coded at the clause level to understand the type of situations involved in simultaneity relations, and at the inter-clause level to find out the combination of situation types in pairs of simultaneous situations:
Scope of simultaneity. A temporal relation of simultaneity can occur between situations in two adjacent clauses or across larger chunks of clauses (Aksu-Koç & von Stutterheim, 1994; Berman, 2008; Kupersmitt, 2015), as illustrated in examples (3) and (4) earlier. Simultaneity that covers larger stretches of text should be inferred and interpreted across shots or scenarios, as compared to more local simultaneity perceived between situations within a single screen shot.
Simultaneity and episodic structuring. The story analyzed here has two main characters – the panther and the bird – and the episodes evolve from both perspectives. This creates a complex temporal texture, since in most cases, coherence depends on a simultaneous perspectivation. For this reason, simultaneous situations were coded according to the characters that were involved in the relation, the episodes where they occurred and the narrative functions performed: (a) moving the story timelines from the panther’s and bird’s perspectives (e.g. panther runs back to river and bird pulls the cuckoo clock back to shore); (Aksu-Koç & von Stutterheim, 1994), (b) backgrounding situations as against plot-line events (e.g. the bird was sleeping with the pink panther and in the morning the alarm clock went off [SM, age 10]), and (c) building the background of the narrative (e.g. the panther is very angry because he’s trying to sleep [FK, adult]; Berman, 2014; Hopper & Thompson, 1980; Kupersmitt, 2015; Reinhart, 1984). Thus, the scenes were selected as representing the main episodic components, where simultaneity played a central role in achieving coherence both at the local and global levels.
At the bottom-up level, the study considers the specific forms that narrators select from the lexical, morphological, and syntactic systems of their language for expressing simultaneity. English has a relatively rich system of aspect, marking perfect, and progressive aspect by the auxiliary verbs have/be combined with a past or present participle, respectively (Palmer, 1974). Besides, English widely uses the -ing present participle in various syntactic contexts, not only as verb complement but also in adverbial or relative clauses (Berman & Slobin, 1994). The linguistic strategies coded in the study are presented below (underlined for clarity) in a continuum from implicit to explicit (Flecken & Schmiedtová, 2007).
Situation aspect (Smith, 1991) refers to the semantic properties of the verb and other elements in the clause that create a certain type of situation, as described earlier. Two unbounded situations (as in (7)) or a durative, unbounded state juxtaposed to a bounded situation (as in (8)) imply a simultaneous temporal relation:
(7) He (8) The pink guy
The situations in (7) are unlikely to happen in sequence, that is, it is common that one dives and looks at the same time, or that one moves one’s head while swimming. In (8), the unbounded situation ‘being tired’ surrounds the RT of ‘smacking’ the bird. In both cases, contextual cues in the film and general knowledge enhance the simultaneous interpretation of the events (Schmiedtová, 2004). Situation aspect was coded only in the absence of more specific strategies to trace the development of more explicit forms.
Syntactic configurations are potential carriers of more explicit simultaneous relations, too. Three types of constructions were coded: clausal complements of perception and cognition verbs (Berman & Slobin, 1994) as occurring in example (4) ‘imagines the cuckoo
(9) And then (the bird) kept on saying cuckoo (10) And he was getting
In example (9), the bare participle in the adverbial clause carries a state that overlaps with the unbounded situation in the previous clause. Cases like (10) express causality between a framing state that temporally includes the resulting situation in the second clause. A particular configuration of juxtaposition in parallel constructions (e.g. ‘Every time he kept looking at him and the cuckoo clock kept moving time’ [SW007, age 10]) was coded separately.
Tense-aspect shifts are perhaps the most studied morphological strategies for temporality expression in developmental psycholinguistics (Bamberg, 1987; Berman & Slobin, 1994; Weist, 2014). In the present context, two types of shifts were considered: aspectual shifts between simple and progressive aspects (as occurring in examples (2), (3), (4), (5), and (10) above) and tense shifts between present and past (e.g. he
Lexical means included temporal adverbials such as meanwhile, in the meantime, prepositions like during or subordinating conjunctions such as while, when, as (Silva, 1991). When was coded separately since it denotes less specific or ‘potential’ simultaneity (Broccias & Smith, 2010) and it can express simultaneous as well as sequential relations depending on the aspect of the predicate and on word knowledge related to the event (Hinrichs, 1986). Other lexical items that enhance a simultaneous interpretation, such as aspectual adverbs (e.g. still, already), aspectual verbs (e.g. keep, start), and lexical repetition (e.g. swim, swim, swim, swim) were counted separately.
Analytic approach
To compare age groups, wherever possible, we used analyses of variance (ANOVAs). In some cases, the assumptions of ANOVAs were not met with the present data, often because the preschoolers produced few instances of the variable of interest and the assumption of normality was violated. In those cases, we compare age groups with non-parametric Kruskal–Wallis tests. To compare the relation between episodic structuring and the expression of simultaneity across ages we used descriptive statistics.
Results
The results presented below are organized according to the questions of the study. Sticking to the analytical approach of the study, the analyses start with top-down processes addressed in the first question – the profiles of development in presenting simultaneous situations from quantitative and qualitative perspectives. This is followed by a descriptive analysis of the discourse features related to simultaneity in the context of narrative production. Bottom-up, the last section shows the results for the development of language forms that serve to express simultaneous relations.
Developmental profiles in presenting simultaneous situations
To evaluate the quantitative profiles in reconstructing simultaneity, the study considered all the situations that were perspectivized as simultaneous. Table 2 shows the length of the texts in number of clauses, the total number of clauses that were coded for simultaneity and the pairs of simultaneous situations. There were significant differences by age group for the text legth in number of clauses, F (2, 97) = 46.25, p < .001,
Mean number (SD) of clauses, clauses involved in simultaneity (also in %), and pairs of simultaneous situations, by age.
As for the pairs of simultaneous situations, there were significant differences by age group, Kruskal–Wallis test statistic (df = 2) = 62.9, p < .001. Pairwise comparisons showed that the preschool children produced fewer than both the school-aged children (Std. test statistic = –4.9, p < .001) and the adults (Std. test statistic = –7.6, p < .001), though no significant difference was found between the school-aged children and the adults (Std. test statistic = –1.6, p = .10).
From a more qualitative view, we analyzed the types of situations that were involved in a relation of simultaneity, as shown in Table 3. Our predictions were confirmed – around three quarters of the simultaneous situations at preschool were unbounded states and activities, as compared to a half at school-age and less than a half among adults. For unbounded activities, there were significant differences by age group, F (2, 88) = 10.49, p < .001,
Mean % (SD) of situation types involved in a simultaneity relation by age group.
As predicted, Kruskal–Wallis test statistic, (df = 2) = 12.5, p = .002, confirmed an effect of age for bounded situations combined with imperfective viewpoint, going from only 9.9% at preschool to 15.7% and 21% at school-age and among adults, respectively. Pairwise comparisons showed that the preschool children tended to produce fewer combinations than the school-aged children (Std. test statistic = –1.8, p = .075) and the adults (Std. test statistic = –3.5, p = .001), but no significant difference was found for the latter (Std. test statistic = –1.3, p = .18).
Discourse features of simultaneity – quantitative and qualitative analyses
The second question of the study addressed the discourse features related to simultaneity in narrative discourse. Results are reported quantitatively and presented in Table 4 for each of the categories analyzed (see ‘Methods’ section), calculated out of pairs of simultaneous situations (see Table 2).
Average (SD) discourse features associated with simultaneity expression, in percentage by age, out of pairs of simultaneous situations.
Only three of these features showed significant age effects: adjacent clauses, F (2, 88) = 25.98, p < .001,
The qualitative analyses below present the distribution and the narrative functions of simultaneous situations within the core episodes in the film. Table 5 presents the episodic components (see the ‘Methods’ section) from the perspective of the panther and from the bird’s perspective (in small letters introduced by B). The situations in each of the episodes are described between two parallel lines // indicating that these occurred simultaneously.
Percentage of narrators expressing simultaneous situations within the core episodes, by age.
GAO: Goal-Attempt-Outcome; IE: Initiating event; IR: internal response.
GAO’s I and II represent superordinate episodes that include a repeated pattern of the panther’s Goal and Attempt with various failure Outcomes and reinstatement of subordinate GAO’s (Trabasso & Rodkin, 1994). For this reason, the episodes have a scope on various scenes. For a detailed description of the GAO’s see the Methods section.
The initiating event that triggers the plot was perspectivized as simultaneous to the setting, particularly at school-age and among adults. At preschool, simultaneity was expressed at the local level in a deictic fashion to describe two durative and recurring situations, but with no clear backgrounding function (e.g. the bird was waking the cat up, the cat was still sleeping [VW, age 5]). At school-age, simultaneity served for backgrounding the setting as against the initiating event that triggers the episode at a more global level of the episode (e.g. he was sleeping soundly and the cuckoo clock went off and the pink panther got really mad so he smashed something [SA, age 10]).
The attempts in the first GAO are perspectivized as simultaneous by almost half of the preschool children and more than two thirds of the older groups. These recurrent scenes, occurring mostly in a single screen shot, are encapsulated as simultaneous, as shown in examples (11) and (12): (11) That pink guy was trying to get rid of that bird that keeps trying to wake him up. (EM05, age 4) (12) And the cuckoo clock just kept waking him up and um he got so mad that he kept hitting everything around him. And then he
The examples earlier are similar in representing the recurrent attempts of both protagonists as simultaneous, though older children seem more able to create a hierarchical temporal structure of backgrounded and foregrounded situations. First, they include the internal responses of the characters within the scope of the simultaneous situations. Second, they delegate the superordinate GAO to the background against the characters’ attempts to reach their respective goals which are presented in a sequential order (as underlined in (12)) in the subsequent subordinate episodes.
The scenes in GAO II comprise a hierarchical temporal structure with several situations occurring simultaneously, each performed by a different protagonist and advancing the plot in two parallel temporal lines. When narrating these complex scenes, which are not perceived in the same screen shot, most of the preschool children presented simultaneity of two unbounded situations that represented the problem leading to the new goal (e.g. he was dreaming his bird was dead in the water [EM07 age 4]) or one of the panther’s attempts to find the bird (e.g. he swam around looking for him [EM104 age 4]). Only two narrators could express simultaneity between the situations in each of the story lines, as shown in the examples below and further represented in Figure 2: (13) S1 Then this guy it’s in the water, S2 but he wasn’t. S3 He was in the house. [SH, age 5] (14) S1 And then raced back S2 but he couldn’t find him S3 ôcuz he was getting … the birdy was going back home. [EM102, age 4]
Representation of Temporal Relations Between the Situations in Examples (13) and (14).
The three unbounded states in (13) are implicitly simultaneous. In (14), a more complex type of simultaneity is achieved by including S1 within the time span of S3, and S2 overlapping with both S1 and S3. Interestingly, some of the youngest narrators could locate the bird’s coming back as anterior to another situation in the timeline, which served as RT (e.g. So he jumped in. But the bird have [cf. has] already got to the other side. [EM03, age 5]).
At school-age, the ability to express simultaneous perspectives in the complex scenes that comprise GAO II increases abruptly to above 50%. The development occurs in quantity – since more episodic components are mentioned along both timelines as compared to the youngest group – and in quality. Situations are temporally arranged in complex ways across clauses, presented as sequential and simultaneous alternatively in hierarchical temporal chunks, within and between the parallel temporal lines as illustrated in example (15) from a story told by a 9-year-old girl and in (16) by an adult, further represented in Figure 3: (15) S1 He (panther) was having dreams S2 that he (bird) was drowning and stuff S3 so he went back and S4 meanwhile the cuckoo bird was like still uh still uh pad-paddling and S5 he got to the shore. (LAC, age 9) (16) S1 He was all sad S2 so he decided to go and look for the bird and S3 when he got to the bridge S4 he jumped off and S5 was swimming around and S6 he couldn’t find him. S7 So then he finally got on shore. S8 Meanwhile the bird had paddled its way to shore and S9 dragged the clock back to the Pink Panther’s house and S10 set it up again and S11 was just cleaning everything up and S12 getting everything back to normal. (AMS, adult)
Representation of Temporal Relations Between the Situations in Examples (15) and (16).
In (15), the unbounded situations S1 and S2 (which are implicitly simultaneous) and the time span in S3, all overlap with the span time of S4 – ‘paddling to shore’ and the punctual event that follows – ‘getting to shore’. In (16), the scope on simultaneous events is even larger. The first state S1 surrounds the RT’s of the plot advancing situations in the panther’s line (S2 to S7) which are themselves simultaneous to the bounded and unbounded situations in the bird’s timeline. Adding to this complexity, older children and adults showed an increasing ability to express simultaneous internal states for backgrounding purposes through the story line, as shown in example (17): (17) S1 Then he feels really guilty about it S2 so he goes back S3 and the cuckoo’s managed to get ashore and S4 gone into the house S5 while he’s searching S6 and feeling really guilty. S7 He gets flowers S8 and drops them into the water S9 hoping that his spirit will be forgiven. (MK, age 10)
In example (17), the panther’s internal response in S1 serves as background – and is simultaneous to the attempts in S2 and S3 which are themselves foregrounded against the unbounded states in S5 and S6. The internal state in S9 is simultaneous to the situations that comprise the resolution after the negative outcome.
The last GAO partially replicates the beginning of the story – while the panther and the bird are sleeping together, the bed clock rings but this time it is the bird who smashes it. In total, 22% of the preschool children could mention part of the scene as simultaneous (e.g. and then when the clock started ringing he smashed it [Andi, age 5]) focusing generally on two overlapping temporal points. After school-age, more than half of the narrators provided a fuller reconstruction of the events as simultaneous, where more salient, plot-line events are foregrounded against protracted situations in the background (e.g. and then he was sleeping with the bird and the clock was ringing and the bird grabbed like a huge wrench and smashed the clock [Sw007, age 10]).
Linguistic expression of simultaneity
For answering the third question of the study, each pair of simultaneous situations was coded for one or more of the linguistic strategies selected by narrators as described in the ‘Methods’ section. As predicted, Table 6 shows significant differences by age in use of most of the strategies. The use of situation aspect to express simultaneity is more frequent among children and significantly decreases among adults, as confirmed by Kruskal–Wallis test statistic, (df = 2) = 6.2, p = .04. Pairwise comparisons showed that the adults produced less than the preschool children (Std. test statistic = –2.2, p = .026) and the school-aged children (Std. test statistic = –2.2, p = .032). There was no difference between the preschool children and the school-aged children (Std. test statistic = 0.3, p = .76).
Language strategies used to express simultaneity, in average % out of pair of simultaneous situations by age (SDs between brackets).
p < .05; **p < .01.
Tense-aspect shifts, as expected, are the next common means for expressing simultaneity at all ages. There is, however, a qualitative difference by age since older children and adults showed more flexibility in use of aspectual varieties switching between simple, progressive, and perfect, as compared to shifts between simple and progressive aspect at preschool. Syntactic configurations occurred twice more often among adults (24.1%) than among preschool children (10.7%). These significant differences by age group were confirmed by Kruskal–Wallis test statistic, (df = 2) = 19.4, p < .001. Pairwise comparisons showed that the preschool children produced fewer than the school-aged children (Std. test statistic = –2.0, p = .045) and the adults (Std. test statistic = –4.4, p < .001). The difference between the school-aged children and the adults was close to significance (Std. test statistic = –1.9, p = .059). Complex syntactic configurations such as non-finite subordination were frequent among adults to express attendant circumstances and internal states related to plot-advancing events (e.g. the pink panther dived
The conjunctions ‘when’ and ‘while’ – as expected – show a different development. ‘When’ remains constant across ages, though ‘while’ and ‘meanwhile’ increase significantly with age, as corroborated by a Kruskal–Wallis test statistic, (df = 2) = 25.7, p < .001. Pairwise comparisons showed that the preschool children produced fewer than the school-aged children (Std. test statistic = –3.0, p = .003) and the adults (Std. test statistic = –4.9, p < .001), but no difference was found between the school-aged children and the adults (Std. test statistic = –1.4, p = .16). The bulk of examples presented earlier (see, for instance, examples (15), (16)) show that these explicit means occur in combination with other strategies such as aspectual shifts and other time words in sophisticated syntactic configurations after middle childhood. These other time words increased significantly by age group, too – Kruskal–Wallis test statistic, (df = 2) = 10.7, p = .005. Pairwise comparisons showed that the preschool children produced fewer than the school-aged children (Std. test statistic = –2.4, p = .016) and the adults (Std. test statistic = –2.9, p = .004), but there was no difference between the school-aged children and the adults (Std. test statistic = –0.3, p = .77).
Finally, parallel structures with -ing forms are scarce in percentage, but they occur significantly more after school-age, too – Kruskal–Wallis test statistic, (df = 2) = 13.3, p = .001. Pairwise comparisons showed that the preschool children produced fewer than the school-aged children (Std. test statistic = –2.3, p = .023) and the adults (Std. test statistic = –3.4, p = .001) with no difference between the school-aged children and the adults (Std. test statistic = –0.9, p = .39).
Discussion
At the outset of this study, we set out three specific aims: (1) to investigate the developmental profiles in presenting simultaneous situations; (2) to examine the discourse features associated with narrative-embedded simultaneity – quantitatively and qualitatively, by age; and (3) to follow the trajectory of language forms used by narrators to express simultaneity. The discussion is organized along these three aims allowing for interfaces when relevant.
The film used in this study provided a meaningful context for the expression of simultaneity, not only because it was rich in simultaneous situations but also because simultaneity relations played a major role in the organization of a coherent episode – at the level of action and mind (Bamberg, 1991; Bruner, 1986). Irrespectively of how situations are depicted in the film, how they are perceived by the narrator, and what their semantic shape is – it is only the narrator who presents them in a specific configuration that reflects various temporal perspectives. As predicted, the quantitative analyses showed an increasing ability to presenting simultaneous situations which is notable at age 7–10, the middle school-age group in this study, supporting the findings of previous studies with narratives (Aksu-Koç & von Stutterheim, 1994; Flecken & Schmiedtová, 2007; Winskel, 2007). Not only the number of clauses increased in length but also the proportion of clauses that were temporally simultaneous with others, meaning that school-age children – similarly to adults – are able to manipulate the temporal thread of the story from various temporal perspectives. Preschool children were able to express simultaneous relations in a much smaller proportion suggesting that an abstract RT system is already emerging around age 4 but has not yet become integrated with other top-down and bottom-up abilities of discourse production (Karmiloff-Smith, 1979, 1981) and has been probably challenged by memory constraints in coping with the off-line narrative production task.
Since we coded all types of simultaneity, we were interested to understand the connection between situation type and simultaneity (Winskel, 2007). As expected, at preschool simultaneity is more frequently expressed between unbounded states and activities than between bounded situations. While states continue to play a significant role in the expression of simultaneity later in development (Kupersmitt, 2015), results indicate that unbounded activities are gradually replaced by durative situations that imply an endpoint, as these have an important role in building the episodes, particularly in the story investigated here. This development does not occur with punctual bounded situations, which were less frequently involved in simultaneity and did not change with development. This is predictable since bounded, punctual events are commonly arranged in a sequential relation as they conform the backbone of the story (Berman, 2014; Hopper & Thompson, 1980; Kupersmitt, 2015). Thus, they are often contrasted with durative, ongoing situations already from an early age at a local level and gradually manipulated against the background at later ages (Berman, 2014; Berman & Slobin, 1994). The difference in semantic features of situations involved in simultaneity is further addressed below from a discourse perspective.
The second question addressed the relation between the knowledge of discourse organization and the expression of simultaneity expression. As expected, preschool children could express simultaneity at a local level between two adjacent clauses. This is not in line with experimental, sentence level studies that reported a later emergence and understanding of simultaneity (Feagans, 1980; Natsopoulos & Abadzi, 1986). From age 7–10, as discourse organization abilities improve children started expressing simultaneity across larger stretches of the text (Aksu-Koç & von Stutterheim, 1994) covering a wider scope on situations that are simultaneous in two parallel timelines. In these cases, the situations within an episode or across episodes were organized in hierarchical configurations of combined simultaneous and sequential relations, particularly among adults. The development occurs not only in the number of situations involved in simultaneity across clauses, but also in the arrangement of complex temporal structures that acquire various narrative functions (Berman, 2014; Berman & Slobin, 1994; Labov, 1972).
Unlike experimental studies, which examined simultaneity at the sentence level of two un/bounded dynamic situations (e.g. Bronckart & Sinclair, 1973; Winskel & Luksaneeyanawin, 2009), the narrative context provided here offered multiple opportunities to trace simultaneous relations between dynamic and non-dynamic situations within complex episodic structures where each of the characters is motivated to attain a goal – thus, feeling, thinking, and acting. From this perspective, we looked at the role of simultaneity in building the background and the foreground of the story (Hopper & Thompson, 1980; Reinhart, 1984) and in emphasizing a character’s perspective (Bamberg, 1994).
The results of the study showed that irrespective of the age, narrators expressed simultaneous situations corresponding to one or two of the characters in similar proportion. The difference, though, was significant in the functions of simultaneity across ages. Younger children seem to be guided by the semantic properties of situations identifying perceptually salient situations and emotional states (e.g. by facial gestures) which are implicitly simultaneous. This explains why preschool children frequently expressed simultaneity with internal states (e.g. and he hated the bird doing that so he wanted to sleep some more [DYL, age 5]) or with salient and protracted activities performed by each of the characters (e.g. he tried to sleep but some bird clock tried waking him up [Sophie, age 5]) for the purpose of building the story background. In the narratives of school-aged children and adults, in contrast, two of more simultaneous states served more mature backgrounding purposes within and across episodes, against which the story advanced.
In all groups, simultaneity was most frequently expressed between bounded and unbounded situations. Qualitative analyses revealed that at the earlier ages children are guided by perceptual and semantic cues to pay attention to local contrasts between punctual versus durative situations, marking these contrasts through language (Bamberg & Damrad-Frye, 1991; Winskel, 2007). This type of simultaneity gradually develops from local to global for the purpose of episodic structuring with combined backgrounded and foregrounded material (Aksu-Koç & von Stutterheim, 1994; Hickmann, 2003; Kupersmitt, 2015).
Results showed that bounded situations were less likely to be presented as simultaneous overall but showed a gradual development with age – scarcely emerging at preschool, and more frequent among adults. This may suggest that it takes much longer for children to understand that the sequences of bounded events in the story are thematically, temporally, and spatially related (e.g. the panther goes back to the river while the bird goes back home). Following the principle of chronological order in narrative discourse (Comrie, 1985; Wiemer, 1997), bounded situations capture a prospective movement along the narrative timeline. In this story, the main episode advances with simultaneous bounded situations that occupy a time span along two parallel timelines as illustrated in the examples earlier (see examples (15) and (16)). For younger children, though, it was demanding to estimate the duration of these events (Feagans, 1980; Piaget, 1946) so as to open a referential interval of time, and the tendency was to focus on the initial or terminal points of situations (Flecken & Schmiedtová, 2007; Hickmann, 2003). Accordingly, qualitative analyses showed that several children could present retrospective temporal perspectives using the past perfect, referring to initial or terminal points of situations that occurred in an anterior time (Irandoust, 1999). This may indicate, as noted, an emergence flexibility of preschool children for locating referential points and relating between them, corroborating the claim that the youngest children must have acquired a bi-referential system of temporal reference (Weist, 1986, 1989).
The challenge for the preschool children is to integrate bottom-up with top-down principles of discourse construction (Berman, 1988; Hickmann, 2003, 2004; Karmiloff-Smith, 1981), thus, they were not able to express simultaneity when the spatial scenario of non-continuous scenes did not depict the actual temporal order of situations. For doing this, children should keep track of events that have been completed against those ongoing, while manipulating two different representations that consider duration, speed, and temporal contour of situations. They should understand that the vantage point of the narrator is different from that of each protagonist – panther and bird – each of them maintaining a different representation of the same temporal reality (McComarck & Hoerl, 1999). Since the simultaneous situations in the most complex episode are shown in non-consecutive shots and cannot be perceived simultaneously, children have to be able to reason that some of the events as seen in the screen were ongoing from the perspective of another referential time – let us say the time when the panther rushed back to the river.
From school-age, top-down knowledge of narrative organization guides speakers to take temporal perspectives independently of the semantic properties of events, giving way to discourse-motivated simultaneity where principles of discourse organization and causality play a substantial role (Bamberg, 1987; Hickmann, 2003; Trabasso et al., 1995). Thus, at ages 7–10, children are already able to channel the attention to simultaneous situations that have a central role in coherence (Aksu-Koç & von Stutterheim, 1994; Berman & Slobin, 1994; Kupersmitt, 2015), represented in the film along several screen shots. In these particular cases, the relation of simultaneity was completely ‘out of sight’ (Weist, 1989, p. 69) and had to be inferred by narrators as an act of perspective taking that involved an abstract representation of two time axes with complex temporal relations within and between them.
Although simultaneous relations are already present in the narratives of the youngest children, it is only at school-age that they become fully and unambiguously represented through language. This was the focus of our third question. The stimuli chosen for the study provided a rich setting to follow the developmental trajectory of a varied range of language forms that serve for the purpose of simultaneity – from the most implicit strategies to the most explicit ones. As expected, implicit, unmarked simultaneity decreases with age, giving space to more explicit marking of this temporal relation after school-age.
Since the study was inspired in a form-function tradition (Bamberg, 1987; Berman, 1996; Slobin, 1997), both quantitative and qualitative analyses provided an ample source of information on the elaboration and diversification of forms for temporal and discourse purposes, in two directions. On one direction, forms to express simultaneity show a long developmental route (Berman, 2004) – through verb semantic and morphological alternations as the widest, basic usage to the use of specific lexical forms like conjunctions, adverbs and aspectual verbs, and more sophisticated syntactic configurations. Unlike previous studies (Aksu-Koç & von Stutterheim, 1994; Flecken & Schmiedtová, 2007; Winskel & Luksaneeyanawin, 2009), this study showed an earlier use of explicit lexical forms and syntactic constructions for simultaneity already at preschool – though far less frequently and less specific than at school-age. We attribute this to the stimuli – which was rich in simultaneous situations of different types – and the off-line task that probably triggered more explicit language forms. Replicating previous findings, around age 9, children rely less on implicit expression of simultaneity by verb semantics and start using more specific and varied conjunctions (e.g. while) and adverbs (e.g. meanwhile, still) for simultaneity (Aksu-Koç & von Stuterheim, 1994; Berman, 2014; Hickmann, 2003; Silva, 1991; Winskel, 2007). Unexpectedly, the results of the study revealed no quantitative differences between school-age children and adults in any of the language strategies analyzed, but qualitative differences were found in the variety of forms that occurred within complex temporal chunks of simultaneous and sequential situations. On a second direction, the study traced the ‘developmental history’ of forms (Berman & Slobin, 1994, p. 603) in the sense of how they are deployed in various linguistic configurations across the text to serve more global purposes, reflecting more mature strategies of text construction (Berman, 1996, 2008). Thus, the youngest children in the study are using aspectual shifts – progressive versus non-progressive aspect in juxtaposed clauses – to take perspectives on punctual as compared to durative situations (Aksu-Koç & von Stuterheim, 1994; Bronckart & Sinclair, 1973; Weist, 1989). In line with the conclusions stated in the study by Aksu-Koç and von Stuterheim (1994) ‘it could be that in English, it is the opposition between progressive and simple present which triggers the distinction between time/reference time and time of perception as the necessary basis for developing a hierarchical narrative structure’ (p. 450). However, it takes longer until children employ this same progressive form in other configurations (e.g. syntactic parallelism and adverbial clauses) and in combination with other means to serve more global purposes of temporal organization. An example of this flexibility is the increasing ability to mark bounded, plot advancing situations (e.g. running to the bridge) in the progressive for the purpose of opening a defined time span to serve as reference for other situations on a parallel temporal line.
Limitations and future studies
Unlike previous cross-linguistic studies of simultaneity as described earlier (e.g. Aksu-Koç & von Stutterheim, 1994), this study included English speakers only. Further studies should include typologically distinct languages to understand the effect of language-specific knowledge on the implicit and explicit expression of simultaneity.
The relatively early expression of local, implicit simultaneity among the youngest narrators may be pointing at the strength of eliciting a narrative which provides rich data showing off preschoolers’ strengths. However, the weakness of the methodology was noted in that participants could simply leave out scenes for reasons of memory constraints or could choose not to perspectivize certain scenes as simultaneous. Future studies could include both narration and sentence-based elicitation from the same participants to capitalize on the strength of both methods.
The findings clearly suggest that more flexible and multiple perspectives on simultaneity are depending on other systems of cognitive knowledge. Future studies should include a wider age range in early childhood and after age 10 until adulthood – as well as independent measures of participants’ perspective-taking ability, counterfactual thought, working memory, and linguistic abilities.
Conclusion
The production of simultaneity relations relies on various types of knowledge and requires the integration between them. Linguistic knowledge is required at various levels – the semantic level to contrast between bounded and unbounded situations including dynamic as well as stative predicates; the morphological level to mark tense and aspect on verbs to express contrastive viewpoints; the lexical level to mark temporal distinctions and relations; and the syntactic level to manipulate various simultaneous perspectives within larger chunks in the text (Berman, 2008, 2014; Berman & Slobin, 1994; Hickmann, 2003). Handling with these levels of linguistic knowledge to express simultaneity seems possible even for the youngest, though age plays a significant role in the activation and integration of all the language systems.
Beyond the linguistic domain children need to be equipped with cognitive abilities to manipulate an abstract story timeline, in which events can be located and related in complex ways beyond the anchor time at speech-time (Weist, 1989). The ability to consider multiple perspectives is related to the notion of decentering in the development of a mature temporal system (Cromer, 1971; Weist, 1989) as analogous to other cognitive abilities associated with language development such as the understanding of false belief (Wimmer & Perner, 1983), or explicitly referring to deception (Mascaro & Sperber, 2009) which show a protracted development across childhood. In narratives, aspects of theory of mind have been explored in relation to expressing evaluative content (Bamberg & Damrad-Frye, 1991; Veneziano et al., 2009) and reference maintenance (Hickmann, 2003; Kail & Hickmann, 1992). The findings of this study reveal a noticeable similarity between the developmental profiles of simultaneity and those of the latter domains – they all emerge at ages 4–5 but become well established by ages 7–10. From a more cognitive perspective, the findings of the study suggest that simultaneity moves from perceptually triggered, with clear reference to salient events as well as internal states, to a ‘mindful simultaneity’ driven by a more abstract representation of the knowledge and thoughts of the characters from a global perspective of goal-directed action.
