The papers in this issue build on this work and examine the relation between time and viewpoint in various languages (English, Dutch, Mandarin, Classical Greek, and American Sign Language) and narrative genres (fictional narratives, nonfictional news narratives, historical narratives, and conversational stories), focusing on the use of tense, temporal adverbs, negation, demonstratives, and body partitioning. Many of the papers draw on Mental Space Theory (Fauconnier 1985) to elucidate the cognitive management of multiple viewpoints and time frames in narrative processing. In this approach, a narrative is conceptualized as a network of interconnected mental spaces through which speakers and addressees move, guided by linguistic signals that establish relations between the various spaces (Sweetser and Fauconnier 1996). This approach proves useful to the collection of studies in explaining how linguistic and gestural cues guide participants in a communicative setting in representing the progression of narrative time. This representation is shown to involve various cognitive operations which regulate the alignment between the viewpoints of narrator, addressee, and narrative characters.

Tense is one of the linguistic categories central to such operations, as demonstrated in a number of the papers. Verhagen (this issue) discusses fictional narratives in which shifts from the past tense to the present tense signal that the currently depicted story events are, despite being unreal, of immediate relevance to readers. The use of tense thus shapes the intersubjective coordination between the viewpoints of narrative character on the one hand and reader and narrator on the other. Similarly, Dancygier (this issue) argues that the present tense evokes a proximal viewpoint on the narrated events, whereas the past tense evokes a distal viewpoint. The use of the present tense does not automatically imply that time progresses in a way that the narrating time coincides with the narrated time, however. Nijk (this issue) discusses examples from Ancient Greek stories set in the present tense in which time progresses in a compressed manner, indicating that the events are not narrated from the dislocated viewpoint of a character who observes the scenes as they unfold, but rather from a retrospective viewpoint. In news narratives, too, the present tense does not necessarily result in a narration of events from the viewpoint of news characters at the time the events took place. Instead, Stukker (this issue) shows that the difference between past tense and present tense is in the news genre to be understood as a difference between more and less temporal distance between the viewpoints of the news characters in the past and the viewpoints of journalist and reader in the present. In addition, Sanders and Van Krieken (this issue) demonstrate that tense shifts in news narratives prompt viewpoint shifts while marking non-iconical movements through time such that a shift from present tense to past tense may signal a movement forward in time instead of backward.

Tense is not the only category relevant to the representation of time and viewpoint. In studying conversational stories narrated in American Sign Language, a language without verbal tense marking, Janzen (this issue) shows how narrators move back and forth through time between the past narrative space and the present discourse space by means of eye gaze and body partitioning. Virdee (this issue) illustrates how negation elements may build an alternative viewpoint space in which time progresses in the opposite direction compared to the default space, resulting in a narration of events that simultaneously proceeds through two viewpoints with narrative time moving forward and backward at the same time. Finally, Lu (this issue) shows that the relation between the representation of time and viewpoint may differ across languages. In English literary stories, tense marking serves to switch between viewpoints and to blend viewpoints. In Mandarin, a language without obligatory tense marking but with perfect aspect and temporal expressions, translations of these stories do not make use of temporal expressions to achieve the viewpoint effects achieved in the English original versions. Instead, viewpoint management in the Mandarin translations appears to rely on reduplication and perception-related lexical constructions.

The analyses presented in the various papers offer insight into the role and nature of the ground in narrative discourse, i.e. the communicative setting in which the story is told by the narrator and heard or read by the audience (cf. Langacker 1987; compare the speech-interaction ground in Sanders et al. 2012). Specifically, they show that, dependent on language and genre characteristics, the ground may be (1) a physical space in reality shared by narrator and addressee (in conversational signed narratives), (2) a non-physical, anticipated space in which the time of narrating is assumed to, but does not actually coincide with the time of reading (in historical narratives and news narratives), or (3) a non-physical, imagined space with narrator and addressee as participants sharing some common ground in terms of a mutual understanding of hypothetical time (in fictional narratives). In each situation, the ground is anchored to coordinates that are fixed in either hypothetical or actual time and space, such that it functions as a home base from which the journey through narrative time starts and to which it may return at any point.

The genre-dependency of the narrative ground’s nature has consequences for the linguistic and cognitive representation of time and viewpoint. For example, the use of tense functions differently in news narratives compared to fictional narratives as a consequence of differences between the communicative grounds (Stukker this issue). In fictional narratives, the present tense can be thought of as bridging the conceptual distance between the viewpoint of the fictional character on the one hand and the viewpoint of the reader on the other, creating “fictional proximity” between these viewpoints. In nonfictional news narratives, by contrast, the present tense is typically used to report on story events that (partially) overlap temporally with the communicative ground or events that happened in the past but have consequences that stretch over time to the communicative ground, thus expressing “temporal proximity”. In this genre, the present tense guides readers in the process of positioning events as temporally closer to or further away from their own here-and-now viewpoint.

The actual progression of narrative time in news stories is signaled by relative temporal adverbs such as later and then. These adverbs are anchored to the viewpoints of the news actors. By contrast, absolute temporal adverbs such as yesterday and tomorrow always require an interpretation from the ground, i.e. from the viewpoint of the journalist and reader in the present. The cognitive representation of news narratives thus requires readers to mentally manage multiple viewpoints (i.e. their own here-and-now viewpoint and the viewpoints of news actors in the past) in order to move through narrative time and to interpret the newsworthy events alternatively from the present and the past (Sanders and Van Krieken this issue).

The collection of papers reveals that such management of multiple viewpoints is in fact central to the representation of narrative discourse (see also Sweetser 2013). In American Sign Language, narrators build a story space in which the viewpoints of characters in the narrative past are enacted. As the narrative unfolds in time, the narrator regularly returns to the ground (the articulation space) to check in with the addressee and to evaluate the story events. Body partitioning and eye gaze are used, in combination with specific utterance structures, to maintain the enacted viewpoint of characters in the past as well as the intersubjective viewpoint of narrator and addressee in the present (Janzen this issue).

In written fictional stories, the representation of multiple viewpoints is linguistically expressed in various ways. Whereas tense is anchored to the narrative ground, the broadest level of the discourse, demonstratives structure the network of lower-level spaces in terms of time and viewpoint (Dancygier this issue). Both levels are represented in the cognitive processing of narratives. In narratives in which time runs backward, negation is used to maintain two alternative temporal viewpoints on a single sequence of events (Virdee this issue). In first person stories, the past tense separates the viewpoints of I-narrator and reader in the communicative ground from the viewpoint of the I-character in the story, whereas a shift to the present tense blends these viewpoints (Verhagen this issue). This too suggests that as a default, readers cognitively represent multiple viewpoints that, guided by tense shifts, at times blend into a single unified viewpoint. Finally, shifts from the past to the present tense do not always seem to involve a complete shift in viewpoint; rather, they may signal the presence of multiple viewpoints involved in the narration. Seemingly paradoxical co-occurrences of the present tense and distal adverbs can be accounted for by assuming a retrospective viewpoint, distinct from the here-and-now viewpoint of the reader, from which the narrative events and the viewpoints of the narrative characters are accessed (Nijk this issue).

By synthesizing cognitive linguistic research on time and viewpoint in narrative, this special issue advances our understanding of how time and viewpoint are conceptualized and expressed in the language of stories. Taken together, the analyses of a wide variety of narratives point towards an account in which the representation of time and viewpoint is guided by a rich palette of linguistic expressions which, in interplay with genre characteristics, evoke specific functional effects (e.g. irony, immediacy, truthfulness) by regulating the temporal and conceptual alignment between the viewpoints of character, narrator, and reader. These findings can be of value to future research in several ways.

First, the various papers both demonstrate and elucidate how the cognitive representation of time in narrative discourse relies on the management of multiple viewpoints. Results of the analyses suggest that shifts between spaces – whether between various narrative spaces or between narrative spaces and the ground – do not necessarily require readers to (re)locate their deictic center to the space currently being profiled in order to process the progression of narrative time; instead, they project their deictic center to the various viewpoints represented in the narrative while maintaining their own here-and-now viewpoint. This expectation could be tested in experimental studies. For example, the ease with which readers process references to a viewpoint located on the narrative time line versus references to a viewpoint located beyond the narrative time line, for instance in the reader’s present if that present is not part of the narrative, could serve as an indication of the extent to which readers keep multiple deictic centers cognitively available during narrative processing.

Second, the studies assembled in this special issue bring forward new insights into the way language marks time frames and regulates the temporal distance between the viewpoints of narrator, character, and addressee. Yet, much remains to be discovered about the representation of the actual progression of narrative time. This progression seems by default linguistically underspecified (see also Dry 1983), which raises the question as to how readers represent and experience the flow, duration, and speed of narrative time in absence of temporal markers. One possibility is that they rely on genre-specific and/or narrative-specific characteristics to draw inferences about the progression of narrative time, which are tested and – if necessary – adjusted for correctness each time a linguistic expression signals the exact coordinates of the narrative now-point. An alternative possibility is that the temporal gaps are not provisionally filled but only covered at the moment such an expression occurs, much like the way people are, while experiencing situations and events, not always acutely and accurately aware of the progression of time. In both scenarios, the representation of narrative time progression largely occurs ad hoc. Note that these explanations are particularly suited for fictional narrative, in which the narrator unfolds a narrative unknown to the addressee. The reconstruction of narrative time progression may be slightly different in non-fictional narratives that relate a series of events globally known to the reader, such as historic or news stories. In such cases, temporal indicators express the salience of the narrative time line, explicitly marking the progression of narrative time. This is shown in the analyses of both Nijk (this issue) and Sanders and Van Krieken (this issue).

The need for further research on narrative time progression is underscored by cross-linguistic differences in the linguistic expression of time and viewpoint, as identified in the papers collected in this special issue. These differences point towards the possibility that, dependent on language-specific characteristics, the cognitive representation of a given narrative is different if that narrative is expressed in language A compared to language B. Of specific interest are differences in the linguistic demarcation of viewpoints. Languages that explicitly mark the boundaries of a viewpoint space, for example by means of tense or temporal expressions, guide readers into building a network of multiple clearly distinguished viewpoints anchored to distinctive time frames. In moving through this network, the temporal progression of the narrative is closely linked to the representation of viewpoints. The mental space network can be expected to be more fluid for languages that do not explicitly mark viewpoint boundaries or that use non-temporal cues to mark these boundaries, as is analyzed by Lu (this issue). Viewpoints are in this case more or less detached from the narrative time line, leaving much freedom to readers in their cognitive representation of the narrative, specifically in their attribution of narrative information to the viewpoints of narrator and characters. An intriguing question for future research is whether cross-linguistic differences in the linguistic expression of time and viewpoint do indeed influence the cognitive representation of narrative discourse and, more generally, how the notion of linguistic relativity plays out in narrative representations.

The majority of the papers collected in this issue were first presented in July 2017 at the 14th International Cognitive Linguistics Conference, held at the University of Tartu in Estonia, within a theme session we hosted there on “Time and Viewpoint in Narrative Discourse”. A follow-up discussion session was held in the Netherlands in April 2018 at the Centre for Language Studies of Radboud University, Nijmegen. We thank all participants and audience members for sharing their thoughts with us on these occasions. We thank Cognitive Linguistics for providing us with a platform to share this collection of studies on time and viewpoint in narrative discourse. We hope that the special issue will provide fresh impetus to future research in this domain. We are especially grateful to the former Editor-in-Chief John Newman for his support and cooperation in bringing the issue together. Finally, we would like to thank all reviewers for their constructive and valuable reviews.