The future is in front, to the right, or below: Development of spatial representations of time in three dimensions

Abstract

Across cultures, people frequently communicate about time in terms of space. English speakers in the United States, for example, might “look forward” to the future or gesture toward the left when talking about the past. As shown by these examples, different dimensions of space are used to represent different temporal concepts. Here, we explored how cultural factors and individual differences shape the development of two types of spatiotemporal representations in 6- to 15-year-old children: the horizontal/vertical mental timeline (in which past and future events are placed on a horizontal or vertical line that is external to the body) and the sagittal mental timeline (in which events are placed on a line that runs through the front-back axis of the body). We tested children in India because the prevalence of both horizontal and vertical calendars there provided a unique opportunity to investigate how calendar orientation and writing direction might each influence the development of the horizontal/vertical mental timeline. Our results suggest that the horizontal/vertical mental timeline and the sagittal mental timeline are constructed in parallel throughout childhood and become increasingly aligned with culturally-conventional orientations. Additionally, we show that experience with calendars may influence the orientation of children’s horizontal/vertical mental timelines, and that individual differences in children’s attitudes toward the past and future may influence the orientation of their sagittal mental timelines. Taken together, our results demonstrate that children are sensitive to both cultural and personal factors when building mental models of time.

Although we cannot see or touch time, it is a fundamental component of our lives. But how do we think and reason about it? As evidenced by language, gesture, and cultural artifacts, people from cultures around the world use space to represent time. Spatial language is frequently used metaphorically to talk about time (e.g., many languages describe the future as being in front and the past behind; Clark, 1973; Lakoff & Johnson, 1980), co-speech gestures add a spatial component to temporal language (Núñez & Cooperrider, 2013; Núñez & Sweetser, 2006), and external tools that project time onto space, like clocks and calendars, are common (e.g., Barnett, 1999; Gell, 1992; Whitrow, 1989). Associations between space and time also run deeper than external cultural tools or linguistic metaphors, as decades of work have suggested that representations of time and space are also linked in the mind (e.g., Boroditsky, 2000; Casasanto et al., 2010; Casasanto & Boroditsky, 2008; Gibbs, 1994; Lakoff & Johnson, 1980). Using space to represent time is thought to be cognitively beneficial by facilitating our ability to reason about temporal order and intervals (Boroditsky, 2000; Casasanto, 2008; Casasanto et al., 2010; Gentner, 2001). Notably, however, spatial representations of time can take varied forms across cultures, and within a culture it is common to use different dimensions of space (e.g., the left-right and front-back axes) to represent different types of temporal concepts (Núñez & Cooperrider, 2013).

Two of the most prevalent mental mappings of time onto space across cultures take the form of a line. One form, commonly referred to as the mental timeline, is a mental model of time based on a linear reference frame oriented on the horizontal or vertical axis (for reviews, see Bender & Beller, 2014; Bonato et al., 2012). With the mental timeline, an observer takes an external perspective, viewing events from outside of the timeline. In other words, the observer is mapping time onto a line that does not intersect with the body. A second type of linear spatiotemporal mapping reflects how individuals represent the future and past relative to themselves. This spatiotemporal mapping employs the sagittal axis, the axis running between the front and back of the body. Unlike the horizontal/vertical mental timeline, the sagittal mental timeline involves an internal perspective with the observer in the center of the axis (Núñez & Cooperrider, 2013). Here, we use the term ‘horizontal/vertical mental timeline’ to refer to what most prior work has referred to as the ‘mental timeline’ and we use the term ‘sagittal mental timeline’ to refer to mapping time onto space in the sagittal axis. As we will review below, the orientation of these mappings—such as whether the future is mapped to the right or left, or to the front or the back—can vary from culture to culture.

The present study explored the development of these two forms of spatiotemporal mappings across early and middle childhood, in a sample of 6- to 15-year-old children in Gujarat, India. We address two overarching questions. First, when—and through what trajectory—might children develop conventional horizontal/vertical and sagittal mappings? Do children initially map time onto space with idiosyncratic orientations that only later become aligned with cultural conventions, or are these mappings constructed in alignment with cultural conventions from the outset? Moreover, do horizontal/vertical and sagittal mappings follow independent developmental trajectories or are these mappings constructed in parallel? Second, what factors influence the orientation of children’s mappings? Are these mappings guided solely by characteristics of the spoken and written languages that children learn, or might they also be influenced by exposure to cultural artifacts like calendars, and reflect individual differences in how much children value the past or future? Below, we review prior research on the horizontal/vertical and sagittal mental timelines before describing our study.

The horizontal/vertical mental timeline

In many Western cultures, the horizontal/vertical mental timeline maps time onto a horizontal spatial axis, frequently with a left-to-right orientation such that earlier events occur on the left and later events occur on the right. However, this orientation is not universal: in some cultures, it is oriented right-to-left, while in others it is oriented top-to-bottom (e.g., Bergen & Lau, 2012; Fuhrman & Boroditsky, 2010; Ouellet et al., 2010; Tversky et al., 1991). One hypothesis for why these differences exist is that the orientation of the horizontal/vertical mental timeline is influenced by the orientation of written language within a culture (Bergen & Lau, 2012). Consistent with this hypothesis, the left-to-right orientation found in many Western cultures parallels the way texts in English and other European languages are read. Conversely, the horizontal/vertical mental timeline in Hebrew speakers is oriented with earlier events occurring on the right and later events occurring on the left, consistent with the right-to-left direction in which Hebrew is read (Fuhrman & Boroditsky, 2010). Similarly, Taiwanese speakers of Mandarin, who read top-to-bottom, orient their horizontal/vertical mental timelines consistent with this vertical orientation (Bergen & Lau, 2012).

An alternate hypothesis, however, is that it is not writing direction, per se, that influences the orientation of the horizontal/vertical mental timeline, but specifically the way that external representations of time, like calendars, are oriented (Pitt & Casasanto, 2020). In the United States, for example, the canonical wall calendar is oriented with the days of the week going from left-to-right (Figure 1A). In other countries, vertical calendars, in which the days of the week are oriented from top-to-bottom, are also common (Figure 1B). Although both of these calendar types each involve the horizontal and vertical axes (whether through the progression of days in the week or weeks in the month), here we classify calendar orientation based on the axis that represents the days of the week. Because many cultures orient this axis of their calendars in the same direction as their written language (e.g., in Arabic-speaking countries, written language is right-to-left and calendars are also oriented with the days of the week going from right-to-left), it is difficult to determine whether it is visual scanning patterns associated with written language versus the way that time is spatialized specifically that influences the orientation of the horizontal/vertical mental timeline. To tease apart these explanations, it is necessary to assess the orientation of the horizontal/vertical mental timeline in a culture in which writing direction is not confounded with calendar orientation.

Figure 1.

Example calendars drawn by our participants illustrating (A) a horizontal orientation and (B) a vertical orientation.

The sagittal mental timeline

In thinking about how oneself moves through time, many languages and cultures describe the future as being in front of the observer and the past as behind (Boroditsky, 2000; Clark, 1973; Lakoff & Johnson, 1980; Torralbo et al., 2006; Ulrich et al., 2011). This front-future orientation may arise from experience with moving through space. As one moves forward physically, the locations to be reached are in front, whereas the locations already passed are behind. However, despite this universal pattern of locomotor experience, there is variation in the way languages map time onto the sagittal axis. Some languages, including Vietnamese and the Andean language Aymara, use spatiotemporal metaphors that refer to the past, which is known, as being in front of the body where it can be seen, whereas the future, which is unknowable, is represented behind the body where it cannot be seen. These linguistic metaphors appear to have cognitive underpinnings – speakers of these languages also conceptualize time with a front-past orientation in the sagittal axis, as evidenced by co-speech gestures (Núñez & Sweetser, 2006; Sullivan & Bui, 2016).

Interestingly, recent work suggests that both cultural attitudes and individual differences can influence the orientation of the sagittal mental timeline, independent of the spatiotemporal metaphors present in a language (Callizo-Romero et al., 2020; de la Fuente et al., 2014; Gu et al., 2019; Li & Cao, 2017; Sullivan & Bui, 2016). These studies measured participants’ sagittal mental timelines using a temporal diagram task (Casasanto, 2009a), which presents participants with a top-down view of a cartoon character with one box in front of the character and one box behind (Figure 2). Participants are told a story that involves one activity the character did in the past and another activity the character is planning to do in the future, and participants are asked to indicate using the boxes when each activity occurred. In the first study involving this task, de la Fuente and colleagues (2014) compared how Spanish and Moroccan Arabic-speaking participants completed the diagrams. Consistent with the front-future encoding of spatiotemporal metaphors in both Spanish and Arabic, Spanish speakers were more likely to place the future activity in the box in front of the character and the past activity in the box behind the character. However, Moroccan participants were more likely to put the past activity in the front box and the future activity in the box behind the character.

Figure 2.

Example of a temporal diagram task stimulus with the face oriented upwards. Participants were given stickers corresponding to past and future events to put in the empty boxes.

de La Fuente and colleagues additionally found that these opposing temporal diagram response patterns were related to cultural differences in the way Spaniards and Moroccans value and focus on the future versus the past. They developed a temporal focus questionnaire to assess individual differences in the value one places on past events and traditional values versus future events and technological progress. The results of this questionnaire indicated that Spanish participants were more future focused, whereas Moroccan participants were more past focused, consistent with their responses on the temporal diagram task. In other words, participants typically mapped the time period (i.e., future or past) that they valued most to the front, and the other time period to the back. A similar relation between participants’ temporal focus and the type of mapping they produce on the temporal diagram task has also been found when comparing younger versus older adults, history and archeology students versus computer science and engineering students, and even museum goers visiting exhibits related to ancient versus contemporary art (de la Fuente et al., 2014; Gu et al., 2019; Li & Cao, 2017; Sullivan & Bui, 2016). A recent study that analyzed all existing data from these measures, as well as new data from additional cultural groups, indicates that this relation is robust: across more than 2,000 participants from more than 10 cultural and subcultural groups, an individual’s temporal focus score predicted significant variance in how likely that individual was to map the future to the front versus the back (Callizo-Romero et al., 2020). Taken together, these findings suggest that both individual and group-level differences in temporal focus influence how people map time onto the sagittal axis.

The present study

Given these well-documented effects of culture and individual differences on the orientation of the horizontal/vertical and sagittal mental timelines, a key open question concerns the development of these mappings. Relatively few studies have investigated the developmental origins of spatiotemporal mappings, and these past studies have focused primarily on the origins of the horizontal/vertical mental timeline (Busby Grant & Suddendorf, 2009; Tillman et al.,2017; Tillman et al., 2018; Tversky et al., 1991; but see Marghetis et al., 2014; Burns et al., 2019 for investigations of children’s temporal gestures). One open-ended method that has been used to assess the development of the horizontal/vertical mental timeline is to ask children to place stickers on a sheet of paper to represent the relative order of multiple events (Tillman et al., 2018; Tversky et al., 1991). For example, Tillman and colleagues (2018) used a variant of this open-ended timeline task to assess the development of linear horizontal or vertical representations of time in English-speaking preschoolers, kindergarteners, and adults. They found that the likelihood of spontaneously producing linear arrangements, and particularly left-to-right-arrangements, increased dramatically between preschool and kindergarten and continued to increase into adulthood. This pattern suggests that children may initially be flexible in how they orient their horizontal/vertical mental timelines; but with age and increasing exposure to cultural artifacts, this orientation may become more conventional.

Here, we aimed to build on prior work by investigating factors that contribute to the orientation and development of the horizontal/vertical mental timeline and the sagittal mental timeline in order to gain a holistic picture of how children construct their mental models of time. We collected data from 6- to 15-year-old children in Gujarat Province, India. All participants were enrolled in English-language schools and most spoke Hindi and Gujarati, in addition to English. Both Hindi and Gujarati are written left-to-right and use spatiotemporal metaphors that place the future in the front and the past in the back, similar to English language metaphors. Because previous studies on the development of spatiotemporal associations have focused on English-, Hebrew-, and Arabic-speaking children (Burns et al., 2019; Tillman et al., 2018; Tversky et al., 1991), the present study provides an additional datapoint for identifying how the organization and development of mental models of time may vary across cultures. As we describe below, studying children in India also provided a unique opportunity to investigate the influence of both writing direction and calendar orientation on the orientation of the horizontal/vertical mental timeline. We had two primary goals centered around understanding the development of spatial representations of time in three dimensions.

Our first goal was to investigate the trajectory with which children develop conventional horizontal/vertical and sagittal mappings. To assess the orientation of children’s horizontal/vertical mental timelines, we used the open-ended timeline task adapted from Tillman et al. (2018). We assessed 1) the age at which children first began to spontaneously produce linear timelines on the open-ended timeline task, as well as 2) when a particular orientation (e.g., left-to-right or top-to-bottom) might become conventional, such that children within an age group respond similarly to each other (e.g., organizing events left-to-right or top-to-bottom), and 3) whether the conventional response changes with age. Similarly, to assess the orientation of children’s sagittal mental timelines, we used a temporal diagram task (Casasanto, 2009a; de la Fuente et al., 2014). We focused on three aspects of children’s responses: 1) children’s consistency of responding across trials (an index of whether children exhibit an internally consistent sagittal mapping), 2) the conventionality of children’s responses, (i.e., whether children within an age group respond similarly), and 3) whether the conventional response changes with age.

Our second goal was to explore the factors that influence the orientation of the horizontal/vertical and sagittal mental timelines. Because Hindi and Gujarati are written left-to-right, but both horizontal calendars (i.e., days of the week ordered left-to-right) and vertical calendars (i.e., days of the week ordered top-to-bottom) are common in India, children in India are an ideal population for exploring how writing direction and calendar orientation each influence the orientation of the horizontal/vertical mental timeline. In particular, if some children orient their horizontal/vertical mental timelines from top-to-bottom, consistent with the orientation of vertical calendars, this would suggest that cultural artifacts that explicitly spatialize time, such as calendars, play a unique role in the organization of the horizontal/vertical timeline, independent of writing direction.

With respect to the orientation of the sagittal mental timeline, we investigated whether individual differences in children’s temporal focus would have an effect on the orientation of their sagittal mental timelines, in line with what has been observed in adults (Callizo-Romero et al., 2020; de la Fuente et al., 2014; Gu et al., 2019; Li & Cao, 2017; Sullivan & Bui, 2016). Because younger adults tend to be more focused on the future than older adults and are correspondingly more likely to endorse a front-future mapping (de la Fuente et al., 2014), one might predict that children would be even more future-focused and likely to map the future to the front. Alternatively, children may exhibit individual differences in their focus on the future versus the past – regardless of their age – and these differences may influence the direction of their sagittal mental timelines. To test this, participants were given an adapted, child-friendly version of the temporal focus questionnaire (de la Fuente et al., 2014) to determine the degree to which individual children valued the future versus the past. We then assessed whether children who placed a greater value on the past were more likely to endorse a front-past mapping compared to children who placed a greater value on the future.

We were also interested in how the developmental trajectories of the horizontal/vertical mental timeline and sagittal mental timelines relate to one another. On the one hand, because these two types of mappings might rely on different types of experience (i.e., exposure to written language and calendars for the horizontal/vertical timeline and locomotor experience and exposure to spatial metaphors for the sagittal timeline), these two types of mappings might exhibit independent developmental trajectories. In addition, because the sagittal mental timeline may arise from locomotor experience and is encoded in language whereas horizontal/vertical mappings are not (e.g., English describes future events as being in front but not to the right), it is possible that children develop a sagittal mental timeline earlier than they construct a horizontal/vertical mental timeline, which may be more reliant on cultural artifacts that young children have limited experience with. These different trajectories could play out in terms of the emergence of consistent spatiotemporal mappings or with regards to the emergence of culturally conventional mapping orientations. On the other hand, if children are constructing a holistic three-dimensional model of how to map time onto space, perhaps as they gain increasing exposure to linguistic metaphors and cultural artifacts, they may show evidence of constructing or changing the orientation of both types of spatiotemporal mappings in parallel.

Method

Participants

Data from 149 participants were included in the final analyses, including data from 36 first grade students (mean age = 6.7 years, range: 5.8-8.0, 11 female), 37 third grade students (mean age = 8.6 years, range: 7.9-10.0, 10 female), 40 sixth grade students (mean age = 11.4 years, range: 10.7-12.4, 23 female), and 36 ninth grade students (mean age = 14.3 years, range: 13.5-15.3, 23 female). All participants were recruited from and tested at two English-language K-12 schools in and around the city of Vadodara, in Gujarat Province, India. We aimed for a sample size of 36 children per grade, but data from slightly more students were collected to enable all eligible students within a classroom to participate. This sample size was chosen based on data from a previous study that used the timeline task to explore the development of the horizontal/vertical mental timeline (Tillman et al., 2018). Data from an additional two sixth grade students were excluded due to using the incorrect colored stickers in the timeline task (N=1) or putting both stickers in the same box for all trials of the temporal diagram task (N=1). Parents of participants gave informed consent in writing prior to their child’s participation, and each child gave either verbal assent (for children aged eight and under) or written assent (for children over age eight) before beginning the study. Children received a small gift for their participation. The study was approved by the Institutional Review Board at the University of California, Berkeley. Data were collected in January 2019.

Procedure

All participants first performed the timeline task, followed by the temporal diagram task. These tasks were administered in this fixed order so that the orientation of the temporal diagram stimuli would not influence responses on the timeline task. Participants in the sixth and ninth grades then completed the temporal focus questionnaire. Finally, a subset of participants in the third and ninth grades were asked to draw a calendar for the month of January. Participants were tested individually (first and third grade participants) or in small groups (sixth and ninth grade participants). Participants that were tested in groups were seated far apart from one another so that they could not see each other’s responses. Depending on the participant’s preferred language, participants tested individually were tested in either English, Hindi, Gujarati, or a combination of the three. Hindi and Gujarati scripts were translated from the English scripts and then back-translated to ensure accuracy. Group testing for sixth grade students was conducted by two experimenters together, one who spoke English and a multilingual experimenter who spoke English, Hindi, and Gujarati and could translate as needed. Group testing for ninth grade students was conducted in English.

Timeline task

The timeline task was adapted from Tillman et al. (2018). To begin, the participant was given a sheet of paper with a 4” x 4” square marked on it and a star sticker in the center of the square. The experimenter introduced the game by saying, “This is a sticker game! First, you will get a card with one sticker already on it. Then, you will put two new stickers on the card. Listen carefully, because I’m going to tell you where to put the stickers. Let’s start!”

For the times of day trial, the experimenter said, “I want you to think about different times of the day. One time of day is morning, another time of day is noon, and another time of day is night. Okay, so the sticker that is already on this card is for noon time. So what is this sticker for?” [Participant(s) should respond “noon,” if not, experimenter corrected the participant(s).] Then the experimenter continued, “your job is to put down a new sticker for night time and another new sticker for morning time. Okay, so the sticker on this page is for noon [point to sticker]. Now, can you put this red sticker down for night time? And can you put this blue sticker down for morning time?”

For the days trial, the script was the same except that participants were prompted to think about yesterday, today, and tomorrow. The sticker that was already on the card was for today, and participants were prompted to put down new stickers for yesterday and tomorrow. Green and yellow stickers were used instead of blue and red. Two different item orders were administered, which varied the order in which the times of day and days trials were administered and the order in which the two stickers were given to the participant to place on the card (i.e., on one trial participants were asked to place the earlier occurring event first, and on the other trial participants were asked to place the later occurring event first).

The sticker arrangements in the timeline task were coded based on the criteria given in Tillman et al. (2018). A sticker arrangement was coded as a line if the smallest angle between the three stickers was been 140 and 180 degrees, and the stickers placed by the participant were on opposite sides of the central sticker (i.e., the stickers were placed in temporal order). Arrangements coded as lines were also classified by direction. All lines within 30 degrees of vertical or horizontal were coded as left-to-right, right-to-left, top-to-bottom, or bottom-to-top. Lines that did not fall within 30 degrees of these axes were coded as diagonal. Two independent coders coded each sticker arrangement, and a third coder resolved any discrepancies. Interrater reliability was very high (Cohen’s kappa = 0.99). Examples of observed linear and nonlinear sticker arrangements can be found in the Supplementary Figure 1.

Temporal diagram task

The temporal diagram task was adapted from de la Fuente et al. (2014) and Li and Cao (2017). On each trial, participants saw a top-down view of a cartoon head with one box in front and one box behind (Figure 2). In contrast to previous applications of this task that used only a single test trial with a diagram of the head facing up, toward the top of the page (Callizo-Romero et al., 2020; de la Fuente et al., 2014; Gu et al., 2019; Li & Cao, 2017; Sullivan & Bui, 2016), we presented participants with four test trials: two that depicted diagrams with the head facing up and two depicting diagrams with the head facing down toward the bottom of the page. Using four trials enabled us to determine how consistent participants were in their response patterns. Alternating the orientation of the head enabled us to differentiate response patterns based on the perspective of the head (i.e., in front versus behind) from those based on the sheet of paper itself (i.e., the top versus bottom of the page).

Before the test trials, the experimenter introduced the task by saying, “You are going to hear some short stories. With each story we will show you a picture that you will use to help you remember what happened in the story.” Participants were then shown a diagram with the cartoon head facing toward the top of the page and the experimenter said, “Do you see this picture? There is a person’s head in the middle, see? [point] And look, the person’s nose here is pointing up, so we can see that this person’s head is facing up. There’s one box in front of the person, here, [point] and one box behind the person, here. [point] Can you put a red sticker in front of the person? And can you put a blue sticker behind the person?” The experimenter then walked the participant through a second practice trial, this time with the cartoon head facing downward.

After the practice trials, the experimenter said, “Great! Okay, so now we are going to listen to a story about this person [point to person’s head]. I’m going to tell you about one thing this person already did in the past, and one thing they are going to do in the future. Then you will put a sticker in each box to help you remember what the person already did and what they are going to do in the future.” Each of the four test trials had a similar format and was accompanied by stickers that matched the events in the story. For example, “Here is a picture of Aditi - you can tell she is facing downwards because of where her nose is. There is one box in front of her [point] and one box behind her [point]. Yesterday, Aditi played football. Tomorrow, Aditi is going to go to the beach. To help you remember, please put the beach sticker in the box for tomorrow. Thanks! Now please put the football sticker in the box for yesterday.” On half of the trials, the cartoon head was facing upwards, such that the box in front was on the top of the page, and on the other half of the trials the cartoon head was facing downwards, such that the box in front was on the bottom of the page. Two of the trials contrasted events that occurred yesterday and tomorrow, and two of the trials contrasted events that occurred last week and next week. In addition, two of the trials featured a character with a stereotypical Indian female name, and two of the trials featured a character with a stereotypical Indian male name. The trials were administered in two different orders that counterbalanced the order in which the trials appeared and whether participants were first asked to place the sticker for the earlier or later occurring event. Complete scripts for all trials can be found in the Supplemental Information.

We classified each participant’s responses as belonging to one of four categories. Participants were classified as front-future mappers if they placed the sticker corresponding to the future event in front of the character on at least three out of four trials. Participants were classified as front-past mappers if they placed the sticker corresponding to the past event in front of the character on at least three of out four trials. Participants were classified as vertical mappers if they placed the stickers in a consistent vertical pattern (e.g., placing the past event sticker in the top box), regardless of the direction the character was facing, on at least three out of four trials. Participants who did not respond consistently on three out of four trials were classified as inconsistent mappers.

We also included a novel temporal diagram interpretation task in which participants were asked to interpret already-completed diagrams. For example, on one trial participants were shown a completed diagram and told “Here is a picture like the ones you just made. This is a picture of Prithi [point]. You can tell Prithi is facing upwards because of where her nose is pointing, and there is one box in front of her and one behind her. Look at this sticker! [point] It means “reading a book.” I want you to tell me if Prithi already read a book last week, or if she is going to read a book later next week. Do you see which box the sticker is in? Did Prithi read a book last week, or is she going to read a book next week?” Participants completed two interpretation trials, one with the character facing upwards and one with the character facing downwards, and the order was counterbalanced across subjects. Complete scripts for all trials can be found in the Supplemental Information.

We classified each participants’ interpretation responses as belonging to one of four categories. Participants were classified as front-future mappers if they said a sticker in front of the character corresponded to a future event on both trials. Participants were classified as front-past mappers if they said a sticker in front of the character corresponded to a past event on both trials. Participants were classified as top-to-bottom mappers if they said a sticker on top of the character corresponded to a past event on both trials. Participants were classified as bottom-to-top mappers if they said a sticker on top of the character corresponded to a future event on both trials.

Temporal focus questionnaire

The temporal focus questionnaire was adapted from de la Fuente et al. (2014) and modified to be shorter and more appropriate for use with children aged 10 years and older. The questionnaire consisted of 10 statements, and participants indicated their level of agreement with each statement using a 5-point Likert scale (1 = dislike very much, 5 = like very much). Five of the statements expressed opinions that reflect greater focus and value on the past (e.g., “I think that people in the past were happier than people are now”) and five statements expressed opinions that reflect greater focus and value on the future (e.g., “I think that new technologies like computers and cell phones make people happier”). The statements were presented in a fixed order, and all participants completed the questionnaire in English. Only students in the sixth and ninth grades completed the questionnaire. Cronbach’s alpha measure of reliability was 0.4 for the future-focused scale and 0.45 for the past-focused scale. The full questionnaire can be found in Supplemental Information.

Calendar drawing task

To assess children’s stereotypes of calendar orientation, we asked a subset of third (n = 27) and ninth grade (n = 17) students to draw a calendar for the month of January on a blank sheet of paper. Calendar drawings were scored as either horizontal (days of the week going left-to-right) or vertical (days of the week going top-to-bottom) (Figure 1).

Data analysis and sharing plan

All data were analyzed in R. The complete dataset, all experimental scripts and materials, and the RMarkdown script used to produce the results section can be found at https://osf.io/kadpt/.

Results

Timeline task

In our first series of analyses, we explored the emergence of a linear horizontal/vertical mental timeline, as well as which orientations were most common. First, we assessed the proportion of trials in which participants placed the stickers in a linear configuration. The proportion of linear arrangements increased across grades (χ²(1) = 26.78, p < 0.001). Although most first grade students did not spontaneously create linear arrangements, by the third grade, the majority of students produced linear arrangements. Specifically, first grade students created linear arrangements on 29% [20-41%] of trials, third grade students created linear arrangements on 64% [52-74%] of trials, sixth grade students created linear arrangements on 79% [69-87%] of trials, and ninth grade students created linear arrangements on 68% [57-78%] of trials.

The above analyses collapsed across the different question types (days or day times) and trial orders. To investigate these additional factors, we performed a mixed-effects logistic regression with fixed effects for grade, question type, and trial order and a random effect of subject. We compared a model that predicted the likelihood of producing a line based on grade to one that included these additional factors. The full model accounted for significantly more variance than the model containing only grade (χ² = 6.46, p = 0.04). However, only grade was a significant predictor (B = 0.22, p < .001), which suggests that no single one of these additional factors contributed additional unique variance.

Figure 3 depicts the proportion of sticker arrangements into each line type for each age group. We analyzed the proportion of trials in which participants placed the stickers in a configuration that reflects a conventional calendar orientation in India (i.e., left-to-right or top-to-bottom for earlier to later events)¹. We found that the proportion of conventional lines increased with grade (χ²(1) = 28.11, p < .001; first grade: 12% [7-22]; third grade: 27% [18-38]; sixth grade: 55% [44-65]; ninth grade: 47% [36-59]; Figure 4). Finally, we tested whether left-to-right or top-to-bottom lines were more common. In all grades, students were equally likely to create both types of lines (χ²s < 3.12, ps > 0.08; first grade: 7 vs 6%; third grade: 9 vs 18%; sixth grade: 34 vs 21%; ninth grade: 17 vs 31%). In sum, the prevalence of students making linear and conventional sticker arrangements increased with age, and both left-to-right and top-to-bottom arrangements were equally common.

Figure 3.

Proportion of timeline arrangements in each grade by line type.

Figure 4.

Proportion of participants in each grade who created conventional (left-to-right or top-to-bottom) lines in the timeline task.

Temporal diagram task

Figure 5 depicts how children of each age group responded in the temporal diagram task. We explored when children begin to exhibit internally consistent sagittal mental timelines and when these mappings begin to converge across participants into a conventional orientation. The proportion of participants in each grade who responded consistently was stable across grades (Chi-squared test for trend in proportions, χ²(1) = 2.07, p = 0.15; first grade: 67% [50-80%]; third grade: 65% [49-78%]; sixth grade: 72% [56-83%]; ninth grade: 81% [65-90%]). However, the proportion of front-future mappers increased with grade (χ²(1) = 10.27, p = 0.001), whereas the proportion of front-past mappers decreased with grade (χ²(1) = 4.96, p = 0.026). Students in the first and third grades were equally likely to be front-future or front-past mappers (χ²s < 0.55, ps > 0.46; first grade: 33% vs 33%; third grade: 38% vs 27%), whereas the majority of students in the sixth and ninth grades were front-future as opposed to front-past mappers (χ²s > 4.51, ps < 0.03; sixth grade: 49% vs 23%; ninth grade: 69% vs 11%). Therefore, although students in all grades exhibited primarily internally consistent mappings, the front-future mapping orientation became increasingly prevalent with age.

Figure 5.

Proportion of children in each grade classified into each mapping direction on temporal diagram production trials.

The above analyses collapsed across the different diagram orientations (face pointing up or down), timeframes (next week/last week or tomorrow/yesterday) and trial orders. To investigate these additional factors, we performed a mixed-effects logistic regression with fixed effects for grade, orientation, timeframe, and trial order, and a random effect of subject. We compared a model that predicted mapping direction (front-future or front-past) based on grade to one that included these additional factors. The full model accounted for significantly more variance than the model containing only grade (χ² = 14.92, p = 0.002). In addition to grade, orientation was a significant predictor (grade: B = −0.15, p < .001; orientation: B = −0.59, p < .001). Participants were more likely to produce front-past diagrams when the face was pointing upwards compared to when it was pointing downwards (upwards: 47% [41-52%]; downwards: 34% [29-40%]). This result suggests that participants may have been referencing a vertically oriented timeline with earlier events on the top and later events on the bottom, in addition to the orientation of the face, when creating these diagrams.

Participants’ temporal diagram interpretations showed a similar developmental trend to their performance on the diagram production task (Figure 6). In first and third grade students, all diagram interpretation patterns were equally common (e.g., front-future, front-past, top-to-bottom, and bottom-to-top; Chi-squared test for difference in proportions, first grade: χ²(1) = 3.85, p = 0.278; third grade: χ²(1) = 6.45, p = 0.092). However, by sixth grade, the front-future interpretation was more common than the other interpretations (sixth grade: χ²(1) = 19.25, p < .001, proportion of front-future mappers: 51% [36-66%]; ninth grade: χ²(1) = 19.14, p < .001, proportion of front-future mappers: 50% [34-66%]). Therefore, children’s production and interpretation of the temporal diagrams exhibit consistent developmental trajectories, with the front-future mapping becoming the dominant mapping orientation by sixth grade.

Figure 6.

Proportion of children in each grade who interpreted the temporal diagrams by each mapping direction.

Relation between temporal focus and temporal diagram mapping directions

Next, we tested how differences in temporal focus related to the orientation of the sagittal mental timeline as indexed by the temporal diagram production task in sixth and ninth grade students (younger students did not complete the temporal focus measure). We used a median split on the temporal focus questionnaire scores to split participants into a future-focused group (mean temporal focus score = 0.17 [0.11 - 0.23]) and a past-focused group (mean temporal focus score = −0.05 [−0.07 - −0.03]; Figure S2). Participants in the future-focused group were significantly more likely to be front-future mappers than participants in the past-focused group (χ²(1) = 13.42, p < .001; proportion of front-future mappers in the future-focused group: 89% [72-96%]; proportion of front-future mappers in the past-focused group: 37% [22-56%]). In addition, participants’ temporal focus scores were significantly correlated with the number of trials in which they produced a front-future diagram (Kendall’s Tau (N = 53) = 0.27, p = 0.009), which indicates that the higher a participant’s temporal focus score, and thus the greater their agreement with future-positive statements, the more likely they were to consistently map the future to the front and the past to the back.

Relation between performance on the timeline and temporal diagram tasks

We used a linear regression model to predict conventional responding on the temporal diagram task (front-future mapping) based on a student’s grade and the number of conventional lines they produced in the timeline task (left-to-right or top-to-bottom). This analysis revealed that both grade and timeline task performance predicted front-future mapping (F(2,145) = 11.46, p < .001; grade: B = 0.22, p = 0.007; timeline task: B = 0.22, p = 0.008), which indicates that conventional responding on the timeline task explained unique variance in the likelihood of being a front-future mapper, even after accounting for the variance explained by grade. These data suggest that conventional mappings between time and space emerge in parallel for both types of spatiotemporal representation.

Calendar drawing task

Students in the third grade were equally likely to draw horizontal versus vertical calendars (χ²(1) = 0, p = 1; proportion of horizontal calendars: 52% [34-69%]), whereas students in the ninth grade were more likely to draw horizontal calendars than vertical calendars (χ²(1) = 4.24, p = 0.04; proportion of horizontal calendars: 71% [47-87%]). However, the proportion of horizontal calendars drawn was not significantly different between the two groups (χ²(1) = 0.84, p = 0.36). These data confirm that the children in our sample were familiar with both horizonal and vertical calendar orientations (see Figure 1 for examples of calendar drawings).

Discussion

Within a culture, different spatial axes are frequently used to represent different temporal concepts. One common representation is the horizontal/vertical mental timeline, a line external to the body that is often oriented horizontally with either earlier events occurring on the left and future events on the right or the reverse (Bergen & Lau, 2012; Casasanto & Bottini, 2014; Tversky et al., 1991). In addition, when thinking about moving through time, many cultures map time onto the sagittal axis with the future represented to the front of the body and the past to the back (Boroditsky, 2000; Clark, 1973; Lakoff & Johnson, 1980; Torralbo et al., 2006; Ulrich et al., 2011). In the present study, we explored the developmental trajectories of both the horizontal/vertical and sagittal mental timelines in children, as well as factors that influence the orientation of these mappings. Children in the first, third, sixth, and ninth grades completed three different spatiotemporal tasks: a timeline task to assess the horizontal/vertical mental timeline (Tillman et al., 2018; Tversky et al., 1991) and two temporal diagram tasks to assess the sagittal mental timeline (de la Fuente et al., 2014; Li & Cao, 2017). Children in the sixth and ninth grades also completed an adapted temporal focus questionnaire (de la Fuente et al., 2014). Together, these data contribute to our understanding of how children map time onto space in three dimensions.

A primary goal of this study was to explore the developmental trajectory of children’s mental timelines along the horizontal/vertical and sagittal axes—specifically, when children begin to systematically map time onto space along these axes and when these mappings begin to exhibit culturally-conventional orientations. Similar to prior work with American and Israeli children (Tillman et al., 2018; Tversky et al., 1991), we found that the horizontal/vertical mental timeline in Indian children becomes increasingly linear and aligned with cultural conventions as children become older and are exposed to more years of formal schooling. The youngest children rarely spontaneously produced linear representations of time in the timeline task and only began to reliably do so by third grade. However, when children did create linear representations, they were equally likely to produce left-to-right horizontal lines and top-to-bottom vertical lines. By sixth grade, the majority of children were producing left-to-right or top-to-bottom linear arrangements. Together, these data suggest that shortly after the point when children begin to consistently produce linear representations of time, they do so in culturally conventional ways.

Overall, the proportion of children who produced linear arrangements on the timeline task was lower than in previously reported samples (e.g., Tillman et al., 2018). Whereas Tillman and colleagues found that that the majority of American kindergarteners produced linear arrangements, we found that the majority of Indian first graders in our sample did not produce linear arrangements. However, given the open-ended nature of this task, we are hesitant to conclude that our data indicate that the development of the horizontal/vertical mental timeline is delayed in Indian children relative to American children. It is difficult to quantify the environmental and socioeconomic differences between these two samples, including the types and amounts of cultural representations of time to which each of these groups of children have been exposed (e.g., the presence of explicit timelines in children’s classrooms). And of course, there may have been differences in how these two groups of children interpreted the task. The primary takeaway of our data—in comparison to prior data from American children—is not the absolute age in which linear responses become dominant, but instead the similarities in developmental trajectories observed across both groups, in which linear and culturally conventional responses increase with age and schooling experience.

With respect to the developmental trajectory of sagittal mappings, we found that children in the first grade, who were mostly 6 and 7 years old, already exhibited internally consistent sagittal mental timelines. That is, most children were consistent in the way they mapped the future and the past to the front and back across trials. What set younger children apart from older children was the direction of these mappings: younger children were equally likely to map the future to the front or the future to the back, whereas older children were more likely to map the future to the front. This was true both for children’s production of temporal diagrams as well as for their interpretation of already-completed diagrams. Taken together, our findings suggest that young children readily employ mappings between time and space in the sagittal plane, but with increasing age and linguistic experience (e.g., exposure to front-future metaphors in language), the direction of these mappings becomes conventionalized toward the front-future orientation. Generalizing across the two types of tasks and mappings, our data suggest that by sixth grade, the majority of children are orienting both their horizontal/vertical and sagittal mental timelines in a culturally conventional way.

A second overarching goal of this study was to understand factors that may influence the orientation of the horizontal/vertical and sagittal mental timelines. Beginning with the horizontal/vertical timeline, we found that children in all age groups were equally likely to spontaneously produce timelines with horizontal left-to-right and vertical top-to-bottom orientations, and that by sixth grade, the majority of children were producing timelines with these orientations. These findings are important because they suggest that external representations of time, like calendars, may influence the orientation of the horizontal/vertical mental timeline. Although prior studies have suggested that the orientation of the horizontal/vertical mental timeline is strongly related to writing direction (Bergen & Lau, 2012), in most cultures writing direction and calendar orientation are confounded. By contrast, although our participants used languages—Hindi, Gujarati, and English—that are written left-to-right, they were also familiar with vertically oriented calendars, in which the days of the week are represented top-to-bottom rather than left-to-right. Indeed, when we asked our third and ninth grade participants to draw a calendar, horizontal and vertical calendars were equally common. Thus, the fact that children in our study spontaneously produced both left-to-right and top-to-bottom orientations suggest that it is not just writing direction but also the way that cultures explicitly spatialize time that influences the orientation of the horizontal/vertical mental timeline (Pitt & Casasanto, 2020).

Turning to the sagittal mental timeline, prior studies have suggested that people tend to orient this mapping such that the temporal era that they value or attend to most is located in front of the body (Callizo-Romero et al., 2020; de la Fuente et al., 2014; Gu et al., 2019; Li & Cao, 2017; Sullivan & Bui, 2016). One finding from this line of work has been that older adults are more likely to endorse the front-past mapping in comparison to younger adults, and that this tendency may be explained by the greater value that older adults place on the past (de la Fuente et al., 2014). Perhaps surprisingly, we found that younger children (first and third graders) were more likely to endorse the front-past mapping in comparison to older children (sixth and ninth graders). Our finding that the front-future and front-past sagittal mental timelines were equally common in first grade children is also unexpected under the view that the front-future mapping is grounded in locomotor experience with the world (Boroditsky, 2000; Clark, 1973; Lakoff & Johnson, 1980; Rinaldi et al., 2016). However, our results suggest that locomotor experience may be just one of many factors that contribute to the orientation of this mapping. For example, we found that children who placed a greater value on the past and traditional values were more likely to endorse the front-past mapping relative to their peers who placed a greater value on the future and the benefits of technology. These results suggest that temporal focus may have a similar influence of the orientation of the sagittal mental timeline in both children and adults.

By assessing mental timelines in the horizontal/vertical and sagittal axes within the same children, we were also able to assess how the developmental trajectories of these timelines might relate to one another. We found that conventional responding on the timeline task (producing lines oriented left-to-right or top-to-bottom) predicted conventional responding on the temporal diagram task (front-future mappings), above and beyond the effects of age. This suggests that children seem to become sensitive to the cultural conventions that guide the orientation of the horizontal/vertical mental timeline and the sagittal mental timeline around the same time. Thus, even though these two types of mappings could be thought of as being shaped by different environmental inputs (e.g., calendars and writing direction for the horizontal/vertical mental timeline, metaphorical language and co-speech gestures for the sagittal mental timeline), they appear to develop in tandem.

In the present study, we assessed children’s spatiotemporal mappings using external diagrams. Other studies have taken a different approach, asking children to explain the difference between temporal concepts (e.g., yesterday vs. tomorrow) and observing co-speech gestures (Marghetis et al., 2014; Burns et al., 2019). Using this method, Burns et al. (2019) found that 6- to 7-year old children in the UK primarily produced lateral gestures in the horizontal axis and rarely produced sagittal gestures, whereas adults were equally likely to produce both types of gestures. These finding suggests that, when not explicitly cued to use the sagittal axis (as we did with children in the present study in our temporal diagram tasks), children may be more likely to represent time along the lateral horizontal axis than along the sagittal axis. An important direction for future work will be to investigate the situations and contexts in which different spatiotemporal representations are engaged (Winter et al., 2015).

The youngest children tested in our study were in the first grade, which means that they had already begun formal schooling and been exposed to spatiotemporal metaphors in language. To fully understand the origins of spatiotemporal mappings, and particularly the role of cultural and linguistic influences, it is necessary to investigate these mappings in individuals who have not had these experiences. Some research suggests that such cultural experiences may not be necessary for developing linear representations of other magnitude domains. For example, adults from the Yupno Valley of Papua New Guinea, who do not have consistent exposure to cultural artifacts that involve linear order, spontaneously create linear arrangements, though the orientations of these arrangements are much more variable than those seen in adults from the US (Cooperrider et al., 2017). Furthermore, this propensity for linear organization may not even be specific to humans: orangutans and gorillas also display a tendency to arrange magnitudes in a linear fashion, though again the orientation of these arrangements is highly variable (Gazes et al., 2017). Although these studies have not explicitly focused on spatiotemporal associations, they suggest that the tendency to think about nonspatial magnitudes in terms of space is present even in the absence of exposure to cultural artifacts. Moreover, previous work with infants in the first year of life suggests that infants may spontaneously align some aspects of space and time, namely spatial extent and temporal duration (de Hevia et al., 2014; Lourenco & Longo, 2010; Srinivasan & Carey, 2010). Whether these associations extend to temporal order and representations of the future and past in order to support the types of spatiotemporal associations investigated here will be a key direction for future work.

One limitation of the present data, which applies both to the temporal diagram tasks and to the timeline task, stems from the tasks being displayed on a sheet of paper on a table in front of the child. As a result, the vertical axis of the paper (e.g., top-to-bottom) was overlayed on the child’s sagittal axis (e.g., front-to-back). In the temporal diagram tasks, there was an illustration of a person in the middle of the page who could be facing either toward the top or the bottom of the page, and the boxes on either side were described as being in front of and behind that person. Because the majority of children altered their responses depending on which way the person was facing, we believe that children were interpreting the diagram in terms of the sagittal axis running between the front and back of the person. However, we also found some evidence that children may have been influenced by vertical orientation of the boxes on the page in addition to the orientation of the depicted person’s face. Although the front-future mapping was the most common response mapping by sixth grade, some children consistently referenced the vertical axis in addition to the sagittal axis when deciding which event happened in the future versus past in both the diagram production and interpretation tasks. Our data demonstrate the importance of including multiple temporal diagram trials (involving the face pointing both upwards and downwards) in order to fully capture the range of spatial axes that participants may reference when completing this task.

An additional benefit of including multiple trials of the temporal diagram task is to help differentiate between intentional and random response patterns. Because our version of temporal diagram task involved four trials, we only classified children as employing a specific orientation if they used that orientation on at least three out of four trials. This requirement helped us differentiate consistent response patterns from children who were responding randomly. Because there were only two trials in our temporal diagram interpretation task, it is possible that children could have responded randomly and still been assigned to a specific mapping orientation. However, the consistent developmental patterns in the data found between these two versions of the task suggest that these tasks are sensitive to developmental change.

Another limitation of our data is that the internal reliability of our temporal focus questionnaire was relatively low. Although earlier studies that used the adult version of this scale (de la Fuente et al., 2014, Li & Cao, 2017) did not report reliability measures, a recent study found that reliability of the adult scale is fairly high, and higher than what we found with our version we adapted for children (Callizo-Romero et al., 2020). Our results are consistent with prior research in suggesting that individual differences in temporal focus throughout the lifespan are related to the orientation of sagittal spatiotemporal mappings, but the relatively low reliability of our measure indicates that additional work is needed to determine how temporal focus may influence children’s sagittal spatiotemporal mappings.

In sum, the present study suggests that two different types of spatiotemporal mappings—the horizontal/vertical mental timeline and the sagittal mental timeline—are constructed in parallel throughout childhood and become increasingly aligned with cultural norms. We provide evidence that multiple factors may influence the orientation of these mappings, including experience with calendars and individual differences in children’s attitudes towards the future. Taken together, our results demonstrate that children are sensitive to both personal and cultural factors when building three-dimensional mental models of time.

Research Highlights.

• Investigated the development of two types of spatiotemporal mappings in children

• Both calendar orientation and writing direction influenced the mental timeline

• Proportion of children who mapped the future to the front increased with age

• Temporal focus predicted whether the future was mapped to the front or back

• Conventional orientations of both mapping types emerged in parallel