Is the future near or far depending on the verb tense markers used? An experimental investigation into the effects of the grammaticalization of the future.

Psycholinguistic approaches that study the effects of language on mental representations have ignored a potential role of the grammaticalization of the future (i.e., how the future manifests linguistically). We argue that the grammaticalization of the future may be an important aspect, as thinking about the future is omnipresent in our everyday life. The aim of this study was to experimentally manipulate the degree of future time references (i.e., present and future verb tense and temporal adverbials) to address their impact on the perceived location of future events. Across four experiments, two in French and two in German, no effect was found, irrespective of our verb and adverbial manipulations, and contrary to our hypotheses. Bayes factors confirmed that our null effects were not due to a lack of power. We present one of the first empirical accounts investigating the role of the grammaticalization of the future on effects of mental representations. We discuss possible reasons for these null results and illustrate further avenues for future research.

Introduction

Languages are built with grammatical structures, such as grammatical gender [1] or grammatical aspect [2, 3]. These structures vary across languages [4] and have been shown to affect how we mentally represent these different aspects of our environment [5-8]. Prior research on the effect of future time reference on temporal discounting [9-11] has established that languages that obligatorily mark the future grammatically show stronger future discounting effects, although the design of these studies remains disputed [12, 13]. In the present study, we examine the impact of the grammaticalization of the future, that is, the grammatical manifestations of how to refer to the future, and how these grammaticalizations may impact our representations of future events. This approach differs from other experimental studies within the field of future time reference and temporal discounting [14, 15], in that we try to unravel the possible underlying cognitive processes involved from a psycholinguistic perspective, as delineated in our previous theoretical work [16].

In a nutshell, we argued that thinking about the future is an activity we engage in, on average, every 16 minutes [17] and that it can affect different mental health outcomes [18]. We therefore regard it as important to examine the link between grammatical manifestations of the future and mental representations of future events. Not only is the future a vital part of our everyday life, but time in general has been shown to be an important dimension for generating and processing mental representations of text and discourse [19]. As mental representations have a perceptual basis [20-22] and grammatical structures can make some aspects of our visual perception more salient [23], we hypothesize that the grammaticalization of the future should influence how we construct grounded representations of future events.

How is the future linguistically realized?

Talking about the future is inherently different than talking about the past or present, in that there is uncertainty about whether an event will take place or not [24, 25]. In European languages there are different nuances of certainty about future events, these nuances are defined by how likely it is that an event will happen. For example, a scheduled meeting or an intended visit are more likely to happen than a predicted outcome of a horse race [24]. This difference in likelihood is reflected in which linguistic devices are used to mark the future: most European languages use the present tense when referring to a very likely event, but when met with uncertainty in a prediction-like context, languages vary with regards to the linguistic devices used [24, 25]. The general notion of referring to a future event is called Future Time Reference (FTR) and can use a variety of linguistic devices [24, 25]. In this study we mainly focus on temporal adverbials (e.g., words that describe the time frame of an action in a clause such as tomorrow [26]) combined with a present or a future tense (e.g., a grammatical construction indicating a temporal context in the future such as ‘I will go.’).

In European languages, FTR varies, to some extent, in the use of linguistic devices in prediction-like contexts, that is, whether a language uses lexical or grammatical structures [24]. In German, for example, the future tense is indicated using a modal verb construction, as in (1).

(1) German

Morgen es regnen.

Tomorrow will it rain:INF.

‘Tomorrow, it will rain.’

In French, the future tense is constructed using an inflectional structure, which means that the future tense is marked within the verb form, as in (2).

(2) French

Demain, il pleuvra.

Tomorrow it rain:FUT.

‘Tomorrow, it will rain.’

An important difference between the German and French future tense is its obligatory use to signal FTR. German future tense use is not obligatory; this means that German speakers can also use the present tense to talk about the future, and they even do this quite often [27]. In contrast, for French speakers it is more common to use a more obligatory form–at least when written [28]. For this paper, we use the terms high and low degrees of FTR to refer to two different situations, as in Jäggi et al. [16]. When comparing French and German, it refers to whether a language requires an obligatory marking (i.e., high degree of FTR) or not (i.e., low degree of FTR). More importantly for this paper, within each language, it also refers to whether we use the future tense (i.e., high degree of FTR) or the present tense to talk about the future (i.e., low degree of FTR).

The effects of different FTR

The difference in degrees of FTR between languages was first used by Chen [9] in his paper on the link between FTR strength and intertemporal choices (i.e., saving money, having retirement assets, or adopting healthy behaviors), where he proposed the Linguistic Savings Hypothesis (LSH). This hypothesis states that speakers of languages with a high degree of FTR (such as French) are less likely to engage in future-oriented behaviors compared to speakers of languages with a low degree of FTR (such as German). The mechanisms suggested for this hypothesis were that speakers of languages with a high degree of FTR perceive the distance to future events as greater and generally less concrete compared to speakers of languages of a low degree of FTR. This is because referring to the future in the present tense leads speakers to perceive the future as if it were occurring in the present (or at least closer to the present). Chen [9] chose to establish a dichotomous criteria between high vs. low degree of FTR languages adapting Dahl’s [24] observations on obligatory FTR use in prediction-like contexts, so that a low degree of FTR was attributed to languages that do not require the verb tense to mark the future (e.g., German) and a high degree of FTR attributed to languages that require the verb tense to mark the future (e.g., English). Chen [9] found a significant effect of the degree of FTR on behavioral outcomes (correlations), even when other economic and demographic parameters were accounted for (e.g., socio-economic status or origin of the legal system in the corresponding country).

Other studies followed Chen’s [9] approach and showed evidence for the LSH with diverse behavioral outcomes, such as corporate savings behavior [10], corporate responsibility [11], research and development investment [11], environmental behavior and policies [29], pro-environmental attitudes [30], future-oriented policies in general [31], as well as religiosity [32]. Nonetheless, some researchers have raised concerns about the methodological approach of these correlational studies [12, 13]. In particular, by reanalyzing Chen’s [9] data but taking into account cultural traits (such as geographical and historical relatedness of languages), the correlations no longer yielded statistical significance [12].

Roberts et al. [12] suggested that experimental designs were better suited to investigate the effect of FTR strength. In fact, researchers that followed an experimental approach did find a difference between German-speaking (low degree of FTR) and Italian-speaking (high degree of FTR) school children in their intertemporal choice preference, in support of the LSH [33]. Chen et al. [15] also compared different degrees of FTR within a language using a time preference task with Mandarin speakers. In their study, however, the findings obtained were not in favor of the LSH [15]. Interestingly, the results showed a trend in the opposite direction, suggesting that participants were more patient about receiving a certain reward when it was presented in the future tense.

The impact of the LSH is still unsettled given the different results reported by various experimental studies [14, 15, 33]. Thus, our goal in the present study is to reassess the LSH and its assumed effects on temporal discounting and delayed gratification from an experimental psycholinguistic approach, as extensively discussed in Jäggi et al. [16]. In using such an approach, not only can we address the issue of causality–and overcome issues related to correlational studies [9-11]–, but we can also examine the underlying cognitive mechanism at the heart of the assumed effect. We predict that the mechanism in question is driven by the fact that the grammatical structures that refer to the future draw attention to temporal information, particularly that of the future.

The thinking-for-speaking hypothesis coined by Slobin [23] provides a potential foundation of the mechanism by which FTR can exert an effect on the mental representations of the future. The thinking-for-speaking hypothesis states that when we prepare our thoughts to be expressed in language, we need to tailor these thoughts into the grammatical structures that our language provides. In turn, this motivates speakers (and listeners) to attend to particular information, which makes certain concepts conveyed through grammar more or less salient. For example, if someone is thinking about a specific event that will happen three weeks from today and expresses it, they would most likely use some sort of FTR to express their thoughts. Depending on whether, in their language, the future tense is or is not marked within the verb form, the attention of speakers (or listeners) might be differently drawn towards the future. The thinking-for-speaking hypothesis has been tested on other domains such as spatial relations [34], grammatical gender [6, 35] and grammatical aspect [7, 8]. We argue that the highlighted differences between the present and the future as suggested by the thinking-for-speaking hypothesis can influence mental representations of the future.

Mental representations of the future

We define mental representations of temporal events as grounded in space. This definition is derived from research on grounded cognition [36-38], which states that cognitive processes (such as mental simulations of events) are not computed amodally, but are rather influenced or grounded in the body, in our perceptual system, or in the physical and/or social environment [37]. Mental representations play an important role when we process language, or more specifically discourse [39, 40]. Mental representations of discourse events are constructed during language processing, and are constantly updated when confronted with new information [19]. Apart from incorporating semantic information, mental representations carry perceptual information, which can be accessed when processing language [20-22]. This pertains to concrete concepts such as performed actions [41, 42] or forms of objects [43, 44], but can also be observed when processing abstract concepts, such as emotions [45], speed [46, 47], space [48], and importantly, time [49, 50].

As time is considered an abstract concept that cannot be defined by itself [38], its definition is based on metonymy (e.g., the iterative event of a clock ticking defines time) and more importantly here, on temporal metaphorization [38]. This latter aspect is crucial in the context of our study, in that the use of temporal metaphors essentially transforms time into a concrete concept. In temporal metaphors, time is grounded in space and/or motion [51]. For example, time is “a moving object” or “we move through time” spatially. Although the spatial translation of time into metaphors is quite universal, these metaphors do vary across languages: while both English- and Mandarin-speakers use horizontal front/back spatial metaphors, Mandarin-speakers also commonly use vertical up/down temporal metaphors [52]. These differing space-time orientations are linked to writing conventions [53]. As the writing convention for French and German is from left to right, we can safely assume that time is perceptually mapped horizontally from left to right. Consequently, and as we do in the present study, using a horizontal timeline to track the spatial differences of grounding time related to using differing degrees of FTR seems appropriate.

The present study

The aim of the present study is to explore whether spatio-temporal representations of time vary as a function of different degrees of FTR, mainly within languages. We look at two languages that have different degrees of FTR between and within them: French, with a higher degree of FTR compared to German, yet with a possibility to use a lower degree of FTR within the language and German, with a lower degree of FTR compared to French yet with the possibility to use a higher degree of FTR within the language. To manipulate FTR experimentally, we use a combination of temporal adverbials and tenses (present/future), as we detail in the Method sections. Experiments 1 and 2 explore possible FTR effects in French, whereas Experiments 3 and 4 explore possible FTR effects in German. To look for between language effects, a post-hoc comparison between Experiment 2 (French) and Experiment 3 (German) is also conducted.

We hypothesize that within each language, readers perceptually represent sentences with a lower degree of FTR (i.e., present tense and temporal adverbials to indicate the future) as spatially closer to the left–representing the present (T0)–than sentences with a higher degree of FTR (i.e., future tense and temporal adverbials to indicate the future). This hypothesis is drawn from the assumption that a higher degree of FTR habitually emphasizes the difference between the present and the future [23] and thus creates perceptually increased distance as time is grounded in space [36]. As a lower degree of FTR is much more common in German than in French, we expect the effect between a low and high degree of FTR in German to be stronger than in French. In other terms, a feeling of novelty when using the future verb tense in German (quite uncommon) might create an even bigger perceptual distance than in French.

Experiment 1

Method

Participants

Sixty-nine participants were recruited for this study. We recruited 20 French-speaking students via convenience sampling at the University of Fribourg. An additional 49 French-speaking participants were recruited using Prolific (www.prolific.co) [11.12.2018], a webservice specialized in online research. The inclusion criterion was that participants’ first language needed to be French. Participants recruited at the University of Fribourg received experimental credits; the Prolific participants received £6.39 per hour for their participation.

We assessed participants’ gender, age, student status and language information such as their first language. Further, we asked participants the first three questions from the Language Experience and Proficiency Questionnaire (LEAP-Q, French version; [54]) to check their level of multilingualism (“Please list all the languages you know in order of dominance.”/”Please list all the languages you know in order of acquisition (your native language first)”) and the amount of time they are immersed in a French surrounding (“Please list what percentage of the time you are currently and on average exposed to each language. (Your percentage should add up to 100%)”). We used the individual language differences to exclude participants that did not meet the inclusion criterion.

The final sample consisted of 30 female and 39 male participants with a mean age of 30.93 years (SD = 10.54). All participants spoke at least one other language–which is very common in Switzerland [55]–and one third of participants reported speaking languages other than French more than 50% of their time.

Materials and procedure

Item construction

We designed 48 sentences that followed a similar pattern: a person starts an event at a given time (e.g., In six months, Julie will join an international organization). The names of the persons were taken from the name registry of the Federal Statistical Office and the 24 most common female and male names from the year 1998 (estimated participant age based on student population) were chosen for our sentences. The events described in the sentences were common events that people experience in their lives, such as starting a new job, moving to a new apartment or finishing a degree.

As the sentences described events at a given time, six adverbials were defined for events taking place in the present (e.g., en ce moment [at this moment], aujourd’hui [today]) and 16 adverbials for events in the future (e.g., dans six mois [in six months], en juin [in June]). Adverbials referring to the future covered a time range between six months up to one year in the future. As temporal adverbials can be placed in different positions within a sentence [56], we decided to construct half of the items with adverbials placed at the beginning of the sentence and the other half with adverbials placed at the end/in the middle of the sentence. This was done to control for a possible position effect.

In addition to the 48 critical sentences, another 48 filler sentences were created with the same structure as the critical sentences. The filler sentences served the purpose of varying the time range and including different adverbials that indicated a time range from tomorrow up to six months.

Scale construction

To measure the perceptual effects of our experimental sentences, we chose to create a visual analogue scale based on the findings of the spatial representation of time [57]. The scale represented a timeline from left to right with the poles tout de suite [right now] to beaucoup plus tard [much later] and was translated as a numerical scale from 0 to 100 when analyzed (i.e., the numerical scale was not visible for participants). The scale was presented at the same time as the sentences and for each sentence, participants were instructed to place the event described in the sentence on the continuous timeline.

Design and procedure

Critical sentences were presented in three different FTR strength conditions: a) present tense–present adverbial (PP), b) present tense–future adverbial (PF), c) future tense–future adverbial (FF) (see Table 1).

Table 1

Example of critical sentences in three different conditions.

Condition	Sentence in French [English translation]
PP	Aujourd’hui, Julie rejoint une organisation internationale. [Today, Julie joins an international organisation.]
PF	Dans six mois, Julie rejoint une organisation internationale. [In six months, Julie joins an international organisation.]
FF	Dans six mois, Julie rejoindra une organisation internationale. [In six months, Julie will join an international organisation.]

Note. Adverbials are marked using italics; tense is highlighted in bold.

Participants were presented with all three conditions, as a repeated-measure design (i.e., 16 sentences in each condition). To ensure that all sentences were also presented in the three conditions across the experiment, three balanced lists were created. The full list of items used in all experiments can be downloaded on the Open Science Framework (OSF) [58].

The experiment was programmed with Qualtrics, an online survey program [59], so participants could participate on their own computers at home. The study link was distributed among the French-speaking university students via social media, flyers and word of mouth advertising. As few students enrolled in the experiment, we decided to further recruit Prolific participants [60].

When starting the experiment, participants first read the study information and were then asked for consent to participate in the study. After consenting, we asked participants for their demographic information (age, gender and student status) and language related information (first language and LEAP-Q). Then participants received additional information on how to place events on the timeline by clicking with their cursor on the estimated position. We emphasized that there were no correct or wrong solutions, but rather, we were interested in their spontaneous opinions. Next, participants were randomly presented with one of the three lists containing 96 sentences to be placed on the timeline–each presented after another in a random order. After placing all 96 events on the timeline, participants were informed that the experiment had ended. On average, participants took 16.5 minutes to complete the study.

Pre-registration

Experiment 1 was pre-registered on the OSF by the Centre for Open Science [61]. Experiments 2 to 4 are modified versions of Experiment 1. As planned in the pre-registration, mixed effects models were calculated with Value as the dependent variable and Time condition as well as Adverbial position as fixed effects with a maximal random effects structure justified by the design that will converge. Further, we pre-registered to use Bayes factor to determine whether the data were sensitive enough to either detect a null hypothesis or the alternative one. Not included in the pre-registration were the additional analyses of the temporal adverbials in Experiment 1 and the between language comparison in Experiment 3.

Ethical consent

Our request for ethical consent with the number 2018–429 to conduct these studies was granted by the ethics committee from the University of Fribourg.

Results and discussion

Before the analyses, we removed participants that did not finish the study or that mentioned that their first language was not French. We ended up with a sample of 69 participants.

To check our hypotheses, a series of linear mixed models were computed and compared using the package afex [62] in R [63]. All linear mixed models used the Value (value of the analogue visual scale) as the dependent variable. The final model contained Time condition (PP vs. PF vs. FF) and Adverb position (Beginning of the sentence vs. End of the sentence) as fixed effects (Adverb position was incorporated to control for a potential effect). Time condition was added as a random slope per Participants and a random slope per Items, and both Participants and Items were set as random intercepts. Both predictors employed treatment contrasts. The reference level for the predictor Time condition was set to the PF condition, so for the effect of Time condition, the intercept is that of the PF condition. For the predictor Adverb position, the reference level was set to the Adverb at the beginning of the sentence condition. The Kenward-Roger method was used to calculate the p-values for the terms in the mixed model. Due to lack of convergence, and as justified by the design, the random structure for the model was only composed of Participants and Items as a random intercept effect. The results of the final model are summarized in Table 2.

Table 2

Summary of the final model: Time condition and Adverb position as fixed effects, Participants and Item as random intercept effect.

Model / Fixed effects	Estimate (β)	df	t-value	p(>|t|)
value ~ Time condition + Adverb position + (1|item number) + (1|participants)
Intercept (PF)	66.40	123.00	47.43	< 0.001
Time condition (PP)	-63.90	3194.37	-111.42	< 0.001
Time condition (FF)	-0.27	3194.37	-0.46	0.64
Adverb position (end of sentence)	0.54	899.40	0.75	0.45

The mean and confidence interval of the different conditions are presented in Fig 1. The mean value for the PP condition was M = 2.74, this value was significantly different from the PF (M = 66.67) and FF conditions (M = 66.42). The comparison between the PF and FF condition did not yield significant differences. Random effects showed that the greatest variability derived from residual variability (Var = 181.24, SD = 13.46), which cannot be attributed to either Participants (Var = 89.47, SD = 9.46) nor Items (Var = 17.67, SD = 4.20).

Fig 1

Mean value per Time condition with confidence intervals (95%).

As the expected difference between the PF and FF did not seem to emerge, we calculated Bayes factor on the lack of effect to assess the relative strength of our evidence. In other terms, we attempted to verify that our data were sufficiently sensitive to detect and strongly support H0 (no effect) over H1 (effect of verb tense) [64, 65]. In order to determine the evidence for H0 over H1 and calculate a Bayes factor, a plausible range of effect is needed. We decided to set it to 32, as it means that PF (expected 34.42) would fall somehow in between PP (2.74) and FF (66.42). We used a half-normal distribution to calculate our Bayes factor, to avoid favoring the probability of supporting H0 over H1 (i.e., half-Cauchy distribution; [66]). In other terms, we statistically tested whether our lack of difference between PF and FF constituted evidence for H0. To do this, we used the difference of -.25 as our sample mean (PF: 66.67; FF: 66.42, and SE = -.46, i.e., the raw difference divided by the t-value given by our model encompassing Time condition). Using the conventional cut-off of .30 suggested by Jefferys [67] the resulting Bayesian analysis showed strong evidence for the null hypothesis over the existence of an effect of Time condition B = .012. This Bayes factor can be taken as substantial evidence for the null hypothesis over the alternative hypothesis (i.e., the alternative hypothesis is .012 times more likely than the null hypothesis). In other terms, our data were sensitive enough to evaluate that the null hypothesis is extremely likely.

To ensure that the adverbials did not influence the placement of events on the timeline an additional analysis was conducted. The analysis of the adverbials revealed an interesting pattern (see Fig 2). Concrete adverbs that mentioned a specific number (e.g., in six months, in eleven months), followed a discrete distribution (i.e., the bigger the number, the bigger the distance on the scale) suggesting that participants used the numbers in the adverbials to create a mental numerical scale and probably mainly focused on the adverbials rather than the tense. This finding is reminiscent of the SNARC effect (i.e., Spatial Numerical Association of Response Codes), which emerges when participants have to classify random Arabic numbers presented on a screen by clicking either the left or right button on a keyboard. They respond more quickly to small numbers with the left hand and more quickly to large numbers on the right hand, suggesting a spatial mental representation of numerical series [68].

Fig 2

Mean value per adverbial and Time condition.

The numerical adverbials (indicated by the numbers six to eleven) show a distinct pattern reminiscent of the SNARC effect [68]. The pattern for other adverbials is less distinct.

In all, the results from Experiment 1 did not support our within language hypothesis. We found that although participants constructed different temporal representations when reading the present tense to refer to the present (PP) and the future tense (FF), there were no differences in their representations when reading the future tense (FF) and present tense to refer to the future (PF). Interestingly, however, the results of our additional analysis on the adverbials indicated a preference for using concrete, numerical adverbials for participants to decide the position of an event on the timeline. To examine whether this effect was masking a possible effect of tense, in Experiment 2 we modified the temporal adverbials to avoid using concrete ones in order to direct participants’ attention to the verb tense.

Experiment 2

Sixty-five French-speaking participants were recruited using convenience sampling. Participants were recruited from the student population at the University of Fribourg. Participants’ mean age was 24.86 years (SD = 6.42) and the sample consisted of 40 females, 24 males and 1 other gender. All participants reported knowing at least two other languages. Participants from the University of Fribourg received experimental credits for their participation.

The same 48 items as in Experiment 1 were used in Experiment 2, with the difference that we modified the adverbials to avoid using numerical, concrete adverbials. The new set of adverbials consisted of four adverbials to refer to the present (e.g., en ce moment [in this moment], maintenant [now]) and eight adverbials to refer to the future (e.g., d’ici quelque mois [within a few months], le semestre prochain [next semester]).

We adjusted the poles of the analogue visual scale from tout de suite [right now] and beaucoup plus tard [much later] to proche dans le temps [close in time] and éloigné dans le temps [distant in time]. These changes were considered due to the large dispersion of the data points in the PP condition indicating a semantic ambiguity of the original poles. The new poles emphasized the spatial dimension of time.

No other changes were made, neither to the procedure nor to the experimental design.

We explored the within language hypothesis for the French-speaking participants. We used the same parameters for the mixed effects model as in Experiment 1. Again, a model with Participants and Items as random slopes and intercepts, and Time condition and Adverb position as fixed effects did not converge. However, the final model did, with Value as a dependent variable, Time condition and Adverb position as fixed effects and Participant and Item as random intercepts. The results of the final model are summarized in Table 3.

Table 3

Results of the mixed effects model of the French-speaking participants for Experiment 2.

Model / Fixed effects	Estimate (β)	df	t-value	p(>|t|)
value ~ Time condition + Adverb position + (1|item number) + (1|participants)
Intercept (PF)	56.16	86.40	28.14	< 0.001
Time condition (PP)	-54.84	3007.51	-92.72	< 0.001
Time condition (FF)	0.29	3007.51	0.50	0.62
Adverb position (end of sentence)	0.37	46.00	0.16	0.87

Mean and confidence interval (95%) are presented in Fig 3. The mean for the PP condition (M = 1.52) was significantly different from the PF (M = 56.29) and FF (M = 56.69) conditions. The two future conditions PF and FF did not significantly differ from each other. Again, for the random effect the greatest variability was derived from residual variability (Var = 180.88, SD = 13.45), which cannot be attributed to either Participants (Var = 74.43, SD = 8.63) or Items (Var = 62.53, SD = 7.91).

Fig 3

Mean value per Time condition and confidence intervals (95%).

Again, we chose to conduct a Bayesian analysis to check if our data were sensitive enough to support the H0. For this analysis, we calculated the SE = 0.8, the difference between PF and FF diff = 0.4. For this experiment we set the expected value to 28, as the range was smaller compared to Experiment 1. The obtained Bayes factor B = 0.036 indicated that our data were sensitive enough.

The results revealed, again, no significant difference between the PF and the FF condition for our participants. One interpretation was that French may actually not be the ideal language to test our hypothesis, as PF may only be partly present in French and mostly in spoken language. As such, French-speakers may not be as familiar with reading about the future in the present tense, at least not as familiar as German-speakers [16]. A potential novelty effect could have masked a potential FTR effect. In the following experiment, we opted to examine German, as German-speakers are more familiar with speaking and reading about the future in the present tense, as well as reading in the future tense.

Experiment 3

Sixty-four German-speaking participants were recruited using a convenience sampling method at the University of Fribourg campus. Further, a mailing list from a Swiss German-speaking university was used to advertise the study among German-speaking psychology students. Participants from the University of Fribourg received experimental credits for their participation. We recruited 51 female and 13 male participants with a mean age of 25.25 years (SD = 7.18). All participants reported speaking at least one second language.

The French items from Experiment 2 were translated by a professional translator into German. The first author translated the items back to French to check for consistency and semantic accuracy. The German equivalent of adverbials to refer to the present were, for example, im Moment [in this moment] and jetzt [now]) and examples of adverbials to refer to the future were in einigen Monaten [within a few months] and nächstes Semester [next semester]).

The scale used in Experiment 2 was translated to German. The German translation of the poles were zeitlich nahe [close in time] and zeitlich entfernt [distant in time].

No other changes in the design or procedure were made compared to Experiment 2.

The same analyses as in Experiment 2 were conducted for the German-speaking participants. Also, the same parameters were chosen for the linear mixed model analysis. Again, the model with the random effect structure with Participants and Items as random slopes and intercepts did not converge, so the final model used Participants and Items as random intercepts. The results of the mixed effects model are presented in Table 4; the mean and confidence interval are presented in Fig 4.

Table 4

Results of the mixed effects model for the German-speaking participants of Experiment 3.

Model / Fixed effects	Estimate (β)	df	t-value	p(>|t|)
value ~ Time condition + Adverb position + (1|item number) + (1|participants)
Intercept (PF)	59.74	111.93	31.61	< 0.001
Time condition (PP)	-57.41	2961.88	-89.52	< 0.001
Time condition (FF)	-0.86	2961.88	-1.34	0.179
Adverb position (end of sentence)	0.24	46.00	0.15	0.88

Fig 4

Mean value per Time condition and confidence intervals (95%).

The PP condition (M = 2.4) was significantly different from the PF and FF conditions. The PF (M = 59.83) and FF (M = 59.08) conditions did not differ significantly. For random effects, the greatest variability was derived from residual variability (Var = 209.71, SD = 14.48), which cannot be attributed to either Participants (Var = 137.53, SD = 11.73) or Items (Var = 27.58, SD = 5.25).

A Bayesian analysis was conducted with the difference -0.75, SE = 0.56 and the expected effect = 28. The Bayes factor B = 0.006 confirmed our data to be sensitive enough to detect H0.

Additionally, a post-hoc between-language comparison (Experiments 2 and 3) was conducted. Following the grounded theory approach, languages that mark the future obligatorily would make a larger spatial distinction between lower and higher degrees of FTR. For our experimental setup, that would mean that French-speaking participants would generally perceive the future conditions (PF and FF) as spatially more distant than German-speaking ones.

We assessed this hypothesis using a linear mixed model. Again, Value was taken as a dependent variable. With regards to the Time condition, only the two future conditions (PF and FF) were compared as we did not suspect a difference in the present condition between the two languages. We also took Language as a fixed factor, which simply indicated whether the participants were French- or German-speaking. We used the R package afex [62] and set treatment contrasts with French as the reference level. We used a random slopes and intercepts structure of Participants and Items, which converged. Language did not seem to significantly explain variance in the data, the analysis can be found on the OSF under supplementary material [58].

As for Experiments 1 and 2, in French, although there was a large difference between PP and FF, as expected, there was no difference between PF and FF. As such, the effect that we expected was not just blurred by a lack of familiarity of PF in French. Now, it could be the case that due to the large difference between the PP condition and the other two conditions, an existing effect between PF and FF could be hidden by creating a temporal zoom-out effect. To examine this final possibility, in the final experiment, we removed the PP condition to zoom in on the expected effect between PF and FF. For this final experiment, we maintained German as our focus language, as we assumed the effect to be bigger in German (see also our rationale for moving from French to German at the end of the discussion section of Experiment 2).

Experiment 4

Fifty-four German-speaking participants were recruited for Experiment 4. Participants were recruited using convenience sampling on campus and via social media. As in Experiment 3, the inclusion criterion for participants was defined as German as a first language. Participants consisted of 40 females and 14 males who were, on average, 31.30 years old (SD = 13.92). In total, 52 out of 54 participants spoke at least one second language.

We kept the same 48 critical PF and FF items as in Experiment 3, removed the PP items, and also eliminated four future adverbials that were tied to a specific month or season. To get a better zoom in effect, we also removed the adverbials in filler items that were too close to the present condition (e.g., morgen [tomorrow]).

The scale was not changed with regards to Experiment 3.

Again, for this experiment, we dropped the PP condition. So, each participant saw 24 sentences in the PF and FF conditions. Participants were randomly appointed to one of the two lists. No other changes to the design and procedure were made.

To check the within language hypothesis, we conducted a mixed effects model. We used the same parameters as in the experiments before. Initially, the model contained Value as the dependent variable, Time condition and Adverb position as fixed effects and Participants as well as Items as random slopes and intercepts. As this model did not converge, the final model used Participants and Items as random intercepts. The results of the mixed effects model can be found in Table 5.

Table 5

Results of the fixed effect for mixed effects model with German-speaking participants for Experiment 4.

Model / Fixed effects	Estimate (β)	df	t-value	p(>|t|)
value ~ Time condition + Adverb position + (1|item number) + (1|participants)
Intercept (PF)	53.55	96.40	16.81	< 0.001
Time condition (FF)	0.53	2491.42	0.99	0.32
Adverb position (end of sentence)	1.40	46.00	0.42	0.68

The mean and confidence interval (95%) are depicted in Fig 5. The two future conditions PF (M = 54.21) and FF (M = 54.83) did not significantly differ from each other. The greatest variability from the random effects was derived from Participants (Var = 245.6, SD = 15.67), following the residual variability (Var = 185.7, SD = 13.63) and Items (Var = 129.5, SD = 11.38).

Fig 5

Mean value per Time condition and confidence intervals (95%).

We conducted a Bayes analysis to evaluate our null result. We used diff = 0.62, SE = 0.62 and the expected value of 28, which resulted in a Bayes factor B = 0.05. So, our data were sensitive enough to confirm H0.

The results from the mixed effects model show that our conditions were not suitable to explain the variance in the data, which was supported by the lack of significant difference between the two conditions. As such, removing the PP condition to zoom in any effect between the PF and FF conditions did not result in any difference compared to Experiment 1, 2 or 3.

General discussion

The series of experiments presented in this paper examined the effect of differing degrees of FTR within languages on the grounded representations of future events. We presented four experiments, to test the hypothesis of whether readers ground sentences with lower degrees of FTR as spatially closer to the left–representing the present (T0)–compared to sentences with higher degrees of FTR within a language. The results of the experiments refuted the proposed hypothesis. To confirm that our results were not due to data insensitivity (associated to some lack of power), we also calculated Bayes factors, as advocated by Dienes [64]. These analyses suggested that the null results are more likely true negatives than false ones, especially since all experiments show consistent results. This opens the discussion to analyzing the underlying reasons for these null results.

Could methodological issues account for our findings? As already mentioned, we made consecutive adjustments to the experiments to rule out methodological issues within our proposed methodological setup. Nevertheless, we chose the experimental setup based on assumed spatial differences in mental representations based on grounded cognition induced by varying degrees of FTR. More precisely, as the thinking-for-speaking framework focuses on visual attention as a cognitive mechanism [23], we expected to measure visual-perceptual traces (i.e., perceived spatial location on a timeline). It could, however, be that the shift of attention induced by differing degrees of FTR may have no incidence on mental representations of the future (although it may have an effect on other cognitive mechanisms).

Another methodological issue could have been that the influence of temporal adverbials overpowered the assumed differences in mental representations affected by FTR variations. As mentioned earlier (see the discussion section of Experiment 1), the additional finding on concrete/numbered adverbials is reminiscent of the SNARC effect found in tasks where participants have to classify Arabic numbers presented on a screen by clicking either a left or right button [68]. The SNARC effect is usually found in paradigms measuring response time, although the response pattern suggests that mental representations of numerical or other ordinal sequences, such as days of the week or months of the year, are spatially encoded [69]. This spatially encoded sequence is mainly manifested in working memory [70, 71]. Although, in our task we did not ask participants to memorize ordinal sequences, information may have been encoded in the working memory as new adverbials appeared in subsequent sentences and participants may have implicitly wanted to put new adverbials according to their already answered items. This account could explain why despite reducing the concreteness of the adverbials in the subsequent experiments, participants were still building a mental timeline based on the remaining adverbials. Further research may need to critically evaluate the use of temporal adverbials in similar paradigms to avoid such possible interference.

Should we have focused on cross-linguistic research? In our investigation we examined two languages based on different properties of FTR, but also based on availability as French- and German-speakers are both present in Switzerland, and participants were readily available. Although we included two different languages in our investigation and we conducted a post-hoc between-language analysis, we did not follow the usual cross-linguistic paradigm, where speakers of languages differing more drastically are compared directly. For example, comparing two languages at the end points of grammatical future marking (e.g., French with obligatory grammatical future marking vs. Finnish with no grammatical future marking) may better document these effects, as those differences may create a more categorically distinct mental representation of the future. Similar research, pertaining to language-and-thought, has examined grammatical and lexical differences that allow clean categorical distinctions [72]. In our experiments we focused on within language differences, which we hypothesized to have perceptual consequences based on fluid grammatical features rather than categorical ones. Perceptual traces of different mental representations may surface with more clean-cut categories. To the best of our knowledge, there has been very little research on such comparisons [16]. Although this may be a legitimate possibility, we would still advocate investigating possible effects with a different paradigm than that tested in the present paper, especially in regards with the lack of a signal of possible perceptual effects.

A recent account, which contends the notion of FTR strength as the responsible driver for temporal discounting, argues that rather modality and, therefore, the amount of certainty transmitted in a FTR may be responsible for probability discounting (rather than temporal discounting) [73]. In order to dismiss the notion that temporal distance is encoded in future tense, Robertson and Roberts [73] conducted a similar experimental paradigm as part of their Study 2. They also experimentally manipulated the present and future tense in English- (high degree of FTR) and Dutch-speakers (low degree of FTR) and used temporal adverbials to invoke different temporal locations on a timeline. Their result matches with the results found in this paper: they could not find a within language effect, where the future tense invoked more temporal distance [73]. The authors argued that their results indicated that FTR strength should not only be determined by future tense, but also by modal variations, which often encode FTR. This is especially true given that their results indicated that the English- and Dutch-speakers in their study were most likely probability discounting, and not temporal discounting. The distinction between future tense and modal variations as attentional cues is important, however, in terms of the thinking-for-speaking hypothesis it also raises two important issues. The first one is related to defining the actual linguistic elements that are most prevalent in FTR, and the second is related to defining what exactly can be considered as future verbal tense. Future research may need to address these intertwined issues to get a more exhaustive picture of the mechanisms at stake.

It may be the case that our hypothesis proved to be wrong, in that different FTR may not have any effect on the way we represent the future. Although our data would suggest this to be true, more data would definitely need to be collected, using different signal traces, maybe linked to other cognitive mechanisms, such as memory encoding (as discussed by Jäggi et al. [16]).

As a concluding comment, we would therefore suggest that future studies focus on both perceptual and non-perceptual effects of different FTR markings, across language groups that have clearer FTR differences. Only then could we have a better perspective on the present null results.

30 Jul 2021

PONE-D-21-18131

Is the Future near or far depending on Verb Tense Markers used? An experimental Investigation into the perceptual Effects of the Grammaticalization of the Future

PLOS ONE

Dear Dr. Jaeggi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Your manuscript has benefitted from 4 reviewers, all experts in this field of research. As you will see, they have provided critical and constructive comments that I think will greatly improve your manuscript, should you choose to resubmit and follow their advice. I have also read your manuscript and find it technically sound, which is a prerequisite for publication in PLOS ONE. There is much to like about the approach you take in your research and generally the results will be useful to other researchers working in this area. However, there are several serious issues pertaining to theory, key terms and concepts, and strength/coherence of argument, identified by the reviewers that would need to be addressed before the manuscript can be considered suitable for publication. Specifically, in a revision, you would engage seriously with the theoretical treatment of the relevant literature (reviewers 2 and 4), disambiguate the use of critical terminology, which currently reflects vague understanding of key terms (e.g. use of the terms ‘perception/perceptual, ‘grammaticized/grammaticization’ etc etc, see comments by reviewers 1, 2, and 3), and offer clear articulation of key concepts and theories to demonstrate adequate understanding and present a more coherent argument (reviewer 2). There are also a couple of serious omissions of previous literature (reviewer 4). I will return the revised manuscript to the reviewers for another round of reviews, so please be sure to address each comment in detail as per the instructions below.

Please submit your revised manuscript by Sep 10 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Panos Athanasopoulos, Ph.D

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

3. Please include your full ethics statement in the ‘Methods’ section of your manuscript file. In your statement, please include the full name of the IRB or ethics committee who approved or waived your study, as well as whether or not you obtained informed written or verbal consent. If consent was waived for your study, please include this information in your statement as well.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

Reviewer #3: Yes

Reviewer #4: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: I Don't Know

Reviewer #3: Yes

Reviewer #4: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: Yes

Reviewer #4: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Review Jaggi et al

The authors report that in Experiment 1 participants essentially discriminated between sentences about the present and sentences about the future, irrespective of how the future was written. However, they found a relationship between higher numbers (in the adverbials) and positioning towards the more distant future on the timeline, suggestive of the SNARC effect. This (former) result was repeated across three futher experiments, which modified the type of sentence, whether there were present sentences at all, and which also used German.

I found the hypothesis interesting and sound. I very much liked the methodical step-by-step approach to each consecutive experiment, and each step was entirely merited. Probably like the authors, I was slightly surprised to find no effect of the way the future was grammaticalised on participants' perceptions of temporal distance.

I would suggest expanding a bit more on the reasons why no effect was found. One thing that struck me about the grammaticalisation argument generally is that the lines between 'grammatical', 'modal', 'adverbial', and perhaps even 'lexical' futures are quite blurred, more so perhaps than those traditional areas of language-and-thought research such as colour, grammatical gender, etc (particularly gender). English is a case in point. 'Will' is a modal verb, and the 'going to' future is esentially a present (continuous) tense with an added infinitive. There is, in fact, not real future tense in English. Witness:

The train leaves at 2pm tomorrow.

I'm going to see her later.

I'll see you then.

I'll have been living there for 30 years come February.

Etc.

All of these are modal verbs or present tenses, each of which have alternative meanings that are unrelated to the future, just 'tweaked' with an adverbial that can refer to the future. My understanding of French is that yes, it uses suffixes on verbs like most romance languages, but more English-like options exist, and of course German and English are related anyway. So perhaps the issue is that there is not enough of a 'clean' categorical distinction between (these) languages for the hypothesis in question. I'd recommend adding more to the para beginning on line 415. Another thought is that there are some languages which only use adverbials to refer to the future, I believe (Hebrew comes to mind, though I may be wrong there). Some languages also have 'strong' past tenses, such as Italian. I don't think you need more data for this particular paper, just a thought for the future (pun unintended).

Some other points:

The table of results (Exp 1) seems a bit thin and a tad confusing - please report that the intercept refers to PF. I deduced this after looking at the graph. Lmers in R usually put the alphabetically earliest factor level as intercept but this information didn't help me as you've reorganised the levels. Also, what was the result for the random effects (i.e., how much of the data did they account for)? It would also be useful if you could report any deviations from your pre-registration, which is not accessible at the time of this review.

I am unclear as to the procedure used for the specific Bayesian test used in Exp 1, but this might be because of typos in the values: aren't the correct numbers 66.42 and 66.47 (not 64.42 and 64.47)? It's also a bit unusual to interpret a Bayesian test as showing anything to be 'truly' sensitive; better to say it's just extremely likely that the data were sensitive enough, etc etc. The raw results certainly back this null up. I'd also report what the BFs mean in layman's terms (i.e. the data were x times more likely under the null hypothesis).

Other

line 18: This is a contentious statement to make with references, let alone without.

At the time of reviewing the pre-registration details for Exp 1 were not available.

Abstract line 11 : irrespective > irrelevant

line 244 masking > shadowing

line 386-387: clunky

Reviewer #2: REVIEW OF: Is the Future near or far depending on Verb Tense Markers used? An experimental Investigation into the perceptual Effects of the Grammaticalization of the Future

The authors present a series of four experiments which test the hypothesis that Future Time Reference (FTR) grammaticization impacts how temporally distal people construe future outcomes to be. On the whole, it appears the authors have done some interesting research which undermines a widely-cited account of the semantics of future tenses and their pursuant effects on construals of future events. The authors have developed a new task and tested an interesting hypothesis, and this is commendable. I believe that the empirical work this MS presents has merit and will be published eventually.

However, the presentation of the work is critically flawed and this prevents me from recommending it be published in its present form. This is principally because the manuscript does not give adequate theoretical treatment to the relevant literature, and repeatedly misconstrues and misunderstands key concepts which need to be properly understood and explained for the MS to be convincing. For instance, there is persistent conceptual drift between within- and between-language effects; the introduction is framed in terms of cross-linguistic effects, yet the studies involve within-language methods and the theoretical bridge between these domains is underdeveloped.

Another oddity is that grammaticization is implied to manifest differently in different speakers at different times. For instance, by the account in the MS, when a speaker says It rains this is “weakly grammaticized” and when the same speaker says It will rain this is “strongly grammaticized”. This fundamentally misunderstands grammaticization, which are those language-level processes which see lexical linguistic elements evolve into grammatical ones. Diachronic processes of grammaticization can therefore lead to cross-linguistic differences in the extent to which languages oblige the use of grammatical markers in certain speech contexts. However, the MS collapses differences in speaker-level usage with notions of grammaticization in such a way as to demonstrate that core concepts have been misunderstood. Key references are cited (Bybee, Dahl) but the MS does not in its present form demonstrate that these have been understood. Another example is that “thinking for speaking” effects are invoked as a theoretical motivation, yet the methods do not investigate these. Additionally, the locus of the cognitive effects under investigation does not appear well understood. The MS repeatedly refers to perceptual effects, yet the methods do not investigate perception, and rather focus on explicit judgments of linguistic cues.

The MS gives an incoherent account of how the semantics of temporal adverbials work (tomorrow, next week, etc.). Firstly, it is suggested that use of a temporal adverbial indicates reduced grammaticization, even though temporal adverbials and future tense operate somewhat orthogonally. The contrast between English (strongly grammaticized FTR) and German (weakly grammaticized FTR) is relevant. Even though English obliges the future tense for prediction-based FTR and German does not, it is acceptable in either language to use a temporal adverbial in combination with the future tense. Secondly, temporal adverbials in German and French transparently encode notions of temporal distance so it is unsurprising that participants rate distal adverbials (e.g. nine months) as farther than proximal ones (e.g. one month). Yet the MS treats this as though it is a novel/interesting result. Too much is made of this, which only distracts from the interesting null effect of tense framing on ratings of temporal distance, and the MS is muddled as a result.

Large and critical areas of literature are unreferenced. In particular, there is a large literature which has focussed on Chen’s (2013) hypothesis that the obligation to use a future tense for prediction-based FTR should cause speakers to construe future events as more temporally distal, and that this will therefore lead to cross-linguistic effects of FTR grammaticiation on intertemporal decision making. Numerous studies, including ones whose results bear direct relevance to the present MS, have followed up on Chen’s (2013) hypothesis. Yet none of this literature is cited, even though the MS purports to test an idea which can be directly traced to Chen (2013). Reading the MS in its own right, it is difficult to make out why future tense marking should lead to distal temporal construals, which makes it all the more strange that Chen (2013) is left uncited. Chen (2013) provides a closely reasoned and mathematically presented account of this hypothesis...

Additionally, while the results are interesting, not enough is made of them. For instance, linguists continue to debate the semantics of FTR, and the future tense in particular, and many of these debates revolve around the entanglement modal with temporal notions in future tense semantics. Given the results in the MS indicate there is no effect of tense framing on ratings of temporal distance, this literature should be engaged, which it currently is not. In fact, the MS’s treatment of modality is generally underdeveloped, and this should be addressed.

The individual studies are not well motivated. The rationale for each study often does not bear up to close scrutiny, or relies on reference to materials which were not included in the material to be reviewed, i.e. the linguistic task which was developed. This makes judging the substantive contribution and motivation of each study difficult.

This brings up a final point, which is that the methods are not clearly reported. While I know roughly what was done, the precise nature of the empirical work remains a mystery. The methods in their present form would not, for instance, provide enough information for the experiments to be replicated. This obviously needs to be corrected.

I fear this review has been overwhelmingly negative, and I want the authors to know that I do believe they have done some worthwhile work. In this review I attach a copy of their MS which I have annotated with suggestions for how it might be improved, and I hope they will implement these in the future, regardless of the editorial decision from PLOS ONE.

Personally, I am very interested in this work; I am near to completing a PhD which has focussed largely on a very similar research question. Much of my work is presently under review or in prep, but I would be happy to be contacted by the authors should they wish to discuss or share results. I can be

contacted at cole.robertson@ru.nl

Reviewer #3: MS PONE-D-21-18131

Title: Is the Future near or far depending on Verb Tense Markers used? An experimental

Investigation into the perceptual Effects of the Grammaticalization of the Future

Authors: Jaeggi, Gygax, Gillioz, Sato

Summary

This manuscript presents four experiments exploring the role of different types of linguistic futures on perceptual qualities. Experiment 1 tested French speakers on the placement of event descriptions on a provided timeline. The results revealed no differences in the different types of futures explored, contrary to the hypothesis. Experiment 2 was a replication of Experiment 1, except concrete temporal adverbials were not used, and a more categorical response modality was used. The results were like those of Experiment 1. Experiment 3 was another replication, how this one involved the use of German speakers. The results were like those of Experiments 1 and 2. Experiment 4 was a replication of Experiment 3, but without the present tense items. The results replicated the first three experiments in not showing an effect of different types of futures. The results revealed that the perceptual aspect of future thinking is largely unaffected by the type of future language structure used.

Evaluation

This manuscript addresses an interesting topic. The study was well-conduced, and the resulting data are adequately interpreted. Although the study is largely a set of null effects, they do have important theoretical implications, and should be made available. I only have one concern that is detailed below.

Minor point

1. I was not entirely satisfied with the description of “perceptual representations”. I think that a more adequate term for capturing what is being referred to here is “embodied representations.”

Reviewer #4: This is an interesting study that presents some important results. The authors should be commended for preregistering the study, seeking to publish even with null results, and for the application of the Bayesian power estimate. This is good science.

I have some points that may require revision of the manuscript. The two main ones relate to the relationship between time and space, and the relation between this study and some parallel work on future discounting. Neither should prevent publication, but the authors may want to think about revisions to make their claims more clear.

1.

Is it necessary to include the spatial translation of time in the hypothesis? Experiment 1 asks participants to express the distance on a slider ranging from left (low) to right (high), so a link between the grammaticalisation and perception of distance could appear even if the participants thought of time as flowing from right to left. The left/right mapping might also predict that the effect would be weaker if the slider was in the other direction. But this isn't tested. So I wonder if the spatial mapping is a necessary step in the hypothesis?

2.

I found it surprising that the paper did not link to the hypothesis by economist Keith Chen on FTR and perceptions of time:

Chen, M. K. (2013). The effect of language on economic behavior: Evidence from savings rates, health behaviors, and retirement assets. American Economic Review, 103(2), 690-731.

Note that the cross-cultural correlation has been criticised:

Roberts, S. G., Winters, J., & Chen, K. (2015). Future tense and economic decisions: controlling for cultural evolution. PloS one, 10(7), e0132145.

But the original idea provides several specific models for how grammaticalisation in language might affect perception of time. In particular, the idea relies on habitual requirements to make distinctions between the future and the present. That is, the more a specific person makes this distinction in their language, the greater the effect on their thought. (that is, there may be within-language differences).

A summary of this research is presented in this preprint (by Cole Robertson and others):

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3628501

This also includes a summary of recent experimental work on the link between FTR and time perception. Several of these experiments may be relevant to the current study, and the authors may wish to cite them (especially since they claim that little work has been done).

The preprint above also makes a point about the interpretation of future tense. While many see tense as encoding temporal distance (*when* an event occurs), it also communicates modal possibility (*whether* an event will occur). This is particularly the case for modals (e.g. in English "will" and "might" differ in probability more than they differ in temporal distance).

Cole Robertson, as part of his PhD thesis and a few papers under review, did some studies on FTR and time perception in English, Dutch and German.

One experiment was similar to the present study: Presenting participants with a phrase and asking them to indicate with a slider how far away in time they felt this was referring to. A similar experiment was also done with a slider indicating certainty.

In Robertson's study, within a language, the future tense frame did not predict the rating of temporal distance. This is in line with the current results.

However, the rating of certainty did differ significantly by grammatical form: modals were rated as less likely to occur than the future or present tenses.

Participants in Robertson's study also rated objective times (one month, two months) as a check that participants were rating things sensibly. This also agrees with the analysis of numbers in adverbials in the current study.

In addition, Dutch speakers rated events as more distant than English speakers. A similar thing could be tested in the current study. That is, is it possible to compare the scale scores in the current study for German and French directly? (i.e. the between-language prediction). It seems like the French participants are placing the slider at higher positions on average than the German participants. Isn't this what your hypothesis would predict? I appreciate the items are not exactly the same, but is it still worth noting? I'll also note that the authors are in a good position to do an experiment with bilinguals in several languages, making the comparison more effective (perhaps for future research).

Robertson continued with some experiments to show that an individual person's usage of future vs. modal strategies for talking about probability (collected in a survey similar to Dahl's survey) could be used to predict their attitude to future events (e.g. in a future discounting task).

I appreciate that these studies are not yet publicly available (though I'm sure the author would share their thesis manuscript if asked), and the point is not to undermine the current study. In fact, both lines of research seem to agree in their results. My intention is just to flag this converging evidence, and to ask the current authors whether there is a way of harnessing their current data to investigate the question about modal possibility, in addition to temporal distance.

Minor points:

"These structures vary across languages and have shown to

affect how we mentally represent and perceive different aspects of our environment." - This sentence appears to have no evidence attached. It's not clear whether the citations in the previous sentence cover this. To be safe, there are some seminal works that could be easily cited here.

"The grammaticalization of the future, which represents the grammatical manifestations of how to refer to the future, has been scarcely studied with experimental psycholinguistic methods"

"As perceptual representations, we mostly refer to perceptual representations of

distance" - Does this mean distance in time? Or are you talking more generally about any kind of domain?

Lines 37-38. It's my understanding that "future time reference" refers to the act of talking about the future, rather than the linguistic devices used to do that?

"FTR varies in the degree of grammaticalization across languages (and at times within language): a low degree of grammaticalization is characterized by adverbials and modal verbs; and a higher degree of grammaticalization is characterized by grammatical structures embedded in the verb, like a suffix of the future tense (e.g., in French: ‘j’ irai à Paris’ [I will go to Paris])."

There is a lot of academic work on grammaticalisation from linguistics, and this could be cited here.

The French example is not illustrative unless one understands French. Please give the interlinear gloss in addition to the translation.

"as some languages have been shown to have very low grammaticalized future verb tense and to only use modal verbs to indicate the future tense (e.g., German:"

I'm not sure this is technically correct. German can also use modal modifiers (möglich) or mental state predicates (erwarten). Maybe you're grouping these under modal verbs?

"we expect the effect between low and high degree of FTR in German to be somehow stronger than in French". Why "somehow" - you have an explicit hypothesis?

Table 2 - the significance is easier to interpret if it's explained that PF condition was used as the intercept condition.

Mixed effects models: You are modelling a scale with a floor and ceiling. Does the model take this into account? It looks like there might be floor effects for the PP condition? You could do this with e.g. logit function.

Line 365: "it did not significantly account variance for Value" - some missing words in this sentence? Also, some statistical support for this claim should be added (e.g. difference in variance explained in a model comparison test)

The size of the text in the figures is quite small and may not reproduce well at a smaller scale.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

Reviewer #4: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

30 Sep 2021

Manuscript ID PONE-D-21-18131

Dear Professor Panos Athanasopoulos,

We are pleased to resubmit our manuscript entitled “Is the future near or far depending on verb tense markers used? An experimental investigation into the effects of the grammaticalization of the future” to PLOS ONE.

We appreciate the reviewers’ critiques of the initial submission and believe that our manuscript greatly improved by responding to their comments. Mainly, we rewrote parts of the introduction concerning the linguistic background to make our argument clearer and more consistent. Further, we clarified what was referred to as ‘perceptual representations’ by adding some key literature concerning grounded cognition. Finally, we provided the omitted literature concerning the work of Chen (2013) and effects of FTR on temporal discounting. Each comment provided by the reviewers has also been addressed point-by-point in our response letter. Our answers are in italic font to facilitate the distinction between the reviewers’ comments and our response.

We hope that the revisions we have provided have led to a better version of the manuscript and that you find it fitting for publication in PLOS ONE.

Best regards,

Tiziana Jäggi

Response to Reviews: Manuscript ID PONE-D-21-18131

Response to Reviewer 1:

“The authors report that in Experiment 1 participants essentially discriminated between sentences about the present and sentences about the future, irrespective of how the future was written. However, they found a relationship between higher numbers (in the adverbials) and positioning towards the more distant future on the timeline, suggestive of the SNARC effect.”

Thank you for mentioning the SNARC effect. We were not aware of such effect, but learning about it made us realize that it is highly relevant for our paper especially with regard to the results of Experiment 1. This is why in the Discussion of Experiment 1 and the General Discussion we have added some sentences around the SNARC effect to explain our results. Please see lines 418ff and 641ff (the lines refer to the document with tracked changes).

“This (former) result was repeated across three further experiments, which modified the type of sentence, whether there were present sentences at all, and which also used German.

I found the hypothesis interesting and sound. I very much liked the methodical step-by-step approach to each consecutive experiment, and each step was entirely merited. Probably like the authors, I was slightly surprised to find no effect of the way the future was grammaticalised on participants' perceptions of temporal distance.

I would suggest expanding a bit more on the reasons why no effect was found. One thing that struck me about the grammaticalisation argument generally is that the lines between 'grammatical', 'modal', 'adverbial', and perhaps even 'lexical' futures are quite blurred, more so perhaps than those traditional areas of language-and-thought research such as colour, grammatical gender, etc (particularly gender). English is a case in point. 'Will' is a modal verb, and the 'going to' future is esentially a present (continuous) tense with an added infinitive. There is, in fact, not real future tense in English. Witness:

The train leaves at 2pm tomorrow.

I'm going to see her later.

I'll see you then.

I'll have been living there for 30 years come February.

Etc.

All of these are modal verbs or present tenses, each of which have alternative meanings that are unrelated to the future, just 'tweaked' with an adverbial that can refer to the future. My understanding of French is that yes, it uses suffixes on verbs like most romance languages, but more English-like options exist, and of course German and English are related anyway. So perhaps the issue is that there is not enough of a 'clean' categorical distinction between (these) languages for the hypothesis in question. I'd recommend adding more to the para beginning on line 415. Another thought is that there are some languages which only use adverbials to refer to the future, I believe (Hebrew comes to mind, though I may be wrong there). Some languages also have 'strong' past tenses, such as Italian. I don't think you need more data for this particular paper, just a thought for the future (pun unintended).”

Thank you for this remark. Indeed, we mainly looked at grammatical determinants of FTR, rather than lexical or modal ones. We rewrote a part of the Introduction about linguistic features of the future and added on to the General Discussion discussing other possible FTRs to make this distinction clearer. We hope by doing so we have been able to meaningfully incorporate your concerns. Please see lines 68ff and 662ff.

“Some other points:

The table of results (Exp 1) seems a bit thin and a tad confusing - please report that the intercept refers to PF. I deduced this after looking at the graph. Lmers in R usually put the alphabetically earliest factor level as intercept but this information didn't help me as you've reorganised the levels. Also, what was the result for the random effects (i.e., how much of the data did they account for)?”

Thank you for these comments. As you suggested, we added more information on the models in the respective parts. See lines 368ff.

“It would also be useful if you could report any deviations from your pre-registration, which is not accessible at the time of this review.”

Thank you for this comment. We made the pre-registration available now. The pre-registration answered questions concerning whether any data had already been collected, what the main question and the hypothesis of the project were, what the key variables were, how many and which conditions participants would be assigned to, which analyses would be conducted and how many observations would be collected. See lines 347ff.

“I am unclear as to the procedure used for the specific Bayesian test used in Exp 1, but this might be because of typos in the values: aren't the correct numbers 66.42 and 66.47 (not 64.42 and 64.47)? It's also a bit unusual to interpret a Bayesian test as showing anything to be 'truly' sensitive; better to say it's just extremely likely that the data were sensitive enough, etc etc. The raw results certainly back this null up. I'd also report what the BFs mean in layman's terms (i.e. the data were x times more likely under the null hypothesis).”

Thank you very much for this attentive remark. Indeed, these numbers were typos. Further, we removed the word ‘truly’ and addressed your remarks concerning the accessibility of the Bayesian results. See lines 403ff.

“Other

line 18: This is a contentious statement to make with references, let alone without.

At the time of reviewing the pre-registration details for Exp 1 were not available.

Abstract line 11: irrespective > irrelevant

line 244 masking > shadowing

line 386-387: clunky”

We appreciate these suggestions. You can find the corresponding changes on lines 36f, 347ff, 29, 438 and 621f.

Response to Reviewer 2:

“The authors present a series of four experiments which test the hypothesis that Future Time Reference (FTR) grammaticization impacts how temporally distal people construe future outcomes to be. On the whole, it appears the authors have done some interesting research which undermines a widely-cited account of the semantics of future tenses and their pursuant effects on construals of future events. The authors have developed a new task and tested an interesting hypothesis, and this is commendable. I believe that the empirical work this MS presents has merit and will be published eventually.”

“However, the presentation of the work is critically flawed and this prevents me from recommending it be published in its present form. This is principally because the manuscript does not give adequate theoretical treatment to the relevant literature, and repeatedly misconstrues and misunderstands key concepts which need to be properly understood and explained for the MS to be convincing. For instance, there is persistent conceptual drift between within- and between-language effects; the introduction is framed in terms of cross-linguistic effects, yet the studies involve within-language methods and the theoretical bridge between these domains is underdeveloped.”

Thank you for your important remarks. We rewrote a great part of the introduction, especially, on the linguistic realizations of the future and tried to improve the theoretical bridge between within and between language effects. See lines 68ff and 241ff (the lines refer to the document with tracked changes).

“Another oddity is that grammaticization is implied to manifest differently in different speakers at different times. For instance, by the account in the MS, when a speaker says It rains this is “weakly grammaticized” and when the same speaker says It will rain this is “strongly grammaticized”. This fundamentally misunderstands grammaticization, which are those language-level processes which see lexical linguistic elements evolve into grammatical ones. Diachronic processes of grammaticization can therefore lead to cross-linguistic differences in the extent to which languages oblige the use of grammatical markers in certain speech contexts. However, the MS collapses differences in speaker-level usage with notions of grammaticization in such a way as to demonstrate that core concepts have been misunderstood. Key references are cited (Bybee, Dahl) but the MS does not in its present form demonstrate that these have been understood. Another example is that “thinking for speaking” effects are invoked as a theoretical motivation, yet the methods do not investigate these. Additionally, the locus of the cognitive effects under investigation does not appear well understood. The MS repeatedly refers to perceptual effects, yet the methods do not investigate perception, and rather focus on explicit judgments of linguistic cues.”

Thank you very much for these comments. Indeed, in the first draft of the manuscript the concept of grammaticalization was not clearly explained and used interchangeably with FTR, creating the theoretical inaccuracy you mentioned. Therefore, we edited a great portion of the introduction addressing your comments and hope to have made this part clearer. Please see lines 68ff.

We still feel that thinking-for-speaking effects are an important part of our theoretical development to understand our experimental setup. However, we do agree that the term ‘perceptual’ creates confusion and a certain misunderstanding regarding the cognitive effects than help make the manuscript clearer. Therefore, we rephrased the concerned parts in the manuscript. Please see lines 182ff.

“The MS gives an incoherent account of how the semantics of temporal adverbials work (tomorrow, next week, etc.). Firstly, it is suggested that use of a temporal adverbial indicates reduced grammaticization, even though temporal adverbials and future tense operate somewhat orthogonally. The contrast between English (strongly grammaticized FTR) and German (weakly grammaticized FTR) is relevant. Even though English obliges the future tense for prediction-based FTR and German does not, it is acceptable in either language to use a temporal adverbial in combination with the future tense. Secondly, temporal adverbials in German and French transparently encode notions of temporal distance so it is unsurprising that participants rate distal adverbials (e.g. nine months) as farther than proximal ones (e.g. one month). Yet the MS treats this as though it is a novel/interesting result. Too much is made of this, which only distracts from the interesting null effect of tense framing on ratings of temporal distance, and the MS is muddled as a result.”

Thank you very much for this remark. We adjusted this part, where it is suggested that use of a temporal adverbial indicates reduced grammaticalization. Again, this inaccuracy was caused by our unconsidered use of the word ‘grammaticalization’. Please see lines 131ff.

Indeed, it is unsurprising that participants would rate ‘nine months’ as farther than ‘one month’. However, we did not expect to find such a scale-like distribution for concrete numbers. We feel this finding, although less important, needs to be appropriately addressed, especially given that Reviewer 1 had mentioned that it is reminiscent of the SNARC effect. We have therefore developed on this issue of the SNARC effect as suggested by Reviewer 1. Please see lines 418ff and 641ff.

We have added on to the discussion, where we – thanks to you – can compare our results with your similar experiment on temporal distance. Indeed, it is very interesting that we both could not find an effect. Please see lines 676ff for the discussion.

“Large and critical areas of literature are unreferenced. In particular, there is a large literature which has focussed on Chen’s (2013) hypothesis that the obligation to use a future tense for prediction-based FTR should cause speakers to construe future events as more temporally distal, and that this will therefore lead to cross-linguistic effects of FTR grammaticiation on intertemporal decision making. Numerous studies, including ones whose results bear direct relevance to the present MS, have followed up on Chen’s (2013) hypothesis. Yet none of this literature is cited, even though the MS purports to test an idea which can be directly traced to Chen (2013). Reading the MS in its own right, it is difficult to make out why future tense marking should lead to distal temporal construals, which makes it all the more strange that Chen (2013) is left uncited. Chen (2013) provides a closely reasoned and mathematically presented account of this hypothesis...”

Thank you for this very important comment. In hindsight, we should have mentioned Chen’s (2013) research and the LSH hypothesis. The intention was certainly not to leave Chen unreferenced, rather we wanted to experimentally follow-up on our own theoretical recommendations delineated in the article Jäggi et al. (2020) – where Chen’s research was an important part – and to emphasize the direction of our psycholinguistic/cognitive-psychological approach. We understand now, that without having our other article in mind, it seems odd not to reference Chen (2013). You can find the added paragraph on lines 131ff.

Jäggi, T., Sato, S., Gillioz, C., & Gygax, P. M. (2020). An Interdisciplinary Approach to Understanding the Psychological Impact of Different Grammaticalizations of the Future. Journal of Cognition, 3(1), 10. DOI: http://doi.org/10.5334/joc.100

“Additionally, while the results are interesting, not enough is made of them. For instance, linguists continue to debate the semantics of FTR, and the future tense in particular, and many of these debates revolve around the entanglement modal with temporal notions in future tense semantics. Given the results in the MS indicate there is no effect of tense framing on ratings of temporal distance, this literature should be engaged, which it currently is not. In fact, the MS’s treatment of modality is generally underdeveloped, and this should be addressed.”

Thank you for addressing this point. As already mentioned, we have extended the discussion to include the debated points. Please see lines 676ff.

“The individual studies are not well motivated. The rationale for each study often does not bear up to close scrutiny, or relies on reference to materials which were not included in the material to be reviewed, i.e. the linguistic task which was developed. This makes judging the substantive contribution and motivation of each study difficult.

This brings up a final point, which is that the methods are not clearly reported. While I know roughly what was done, the precise nature of the empirical work remains a mystery. The methods in their present form would not, for instance, provide enough information for the experiments to be replicated. This obviously needs to be corrected.”

Thank you for this comment. We have added the complete list of items for each experiment on the OSF.io (see https://osf.io/s2axr/?view_only=68a8142718814603b36e3bcb14ff349b). Please see lines 328f.

“I fear this review has been overwhelmingly negative, and I want the authors to know that I do believe they have done some worthwhile work. In this review I attach a copy of their MS which I have annotated with suggestions for how it might be improved, and I hope they will implement these in the future, regardless of the editorial decision from PLOS ONE.

Personally, I am very interested in this work; I am near to completing a PhD which has focussed largely on a very similar research question. Much of my work is presently under review or in prep, but I would be happy to be contacted by the authors should they wish to discuss or share results. I can be

contacted at cole.robertson@ru.nl”

It was very nice of you to have reached out to us and to have let us know about your thesis and the similarity of the subjects. In doing so, we were able to integrate some important results to compare our study in the discussion. Please see lines 676ff.

Response to Reviewer 3:

“Summary

This manuscript presents four experiments exploring the role of different types of linguistic futures on perceptual qualities. Experiment 1 tested French speakers on the placement of event descriptions on a provided timeline. The results revealed no differences in the different types of futures explored, contrary to the hypothesis. Experiment 2 was a replication of Experiment 1, except concrete temporal adverbials were not used, and a more categorical response modality was used. The results were like those of Experiment 1. Experiment 3 was another replication, how this one involved the use of German speakers. The results were like those of Experiments 1 and 2. Experiment 4 was a replication of Experiment 3, but without the present tense items. The results replicated the first three experiments in not showing an effect of different types of futures. The results revealed that the perceptual aspect of future thinking is largely unaffected by the type of future language structure used.”

“Evaluation

This manuscript addresses an interesting topic. The study was well-conduced, and the resulting data are adequately interpreted. Although the study is largely a set of null effects, they do have important theoretical implications, and should be made available. I only have one concern that is detailed below.”

“Minor point

1. I was not entirely satisfied with the description of “perceptual representations”. I think that a more adequate term for capturing what is being referred to here is “embodied representations.””

Thank you for this concern. We address this issue by adding literature on grounded cognition and not referring to mental representations as ‘perceptual representations’ anymore. Please see lines 205ff (the lines refer to the document with tracked changes).

Response to Reviewer 4:

“I have some points that may require revision of the manuscript. The two main ones relate to the relationship between time and space, and the relation between this study and some parallel work on future discounting. Neither should prevent publication, but the authors may want to think about revisions to make their claims more clear.”

“1.

Is it necessary to include the spatial translation of time in the hypothesis? Experiment 1 asks participants to express the distance on a slider ranging from left (low) to right (high), so a link between the grammaticalisation and perception of distance could appear even if the participants thought of time as flowing from right to left. The left/right mapping might also predict that the effect would be weaker if the slider was in the other direction. But this isn't tested. So I wonder if the spatial mapping is a necessary step in the hypothesis?”

Thank you for this comment. Indeed, we did not test whether the effect would be weaker by changing the slider direction from right to left. However, we still feel that it is important to include the literature on spatial translation of time as it is relevant to understand what we meant when referring to ‘perceptual representations’. We have changed and added a substantial part of the introduction concerning the spatial translation of time regarding grounded cognition. We hope in rewriting this part, it has become clearer why we would like to keep this part in the manuscript. Please see lines 205ff (the lines refer to the document with tracked changes).

“2.

I found it surprising that the paper did not link to the hypothesis by economist Keith Chen on FTR and perceptions of time:

Chen, M. K. (2013). The effect of language on economic behavior: Evidence from savings rates, health behaviors, and retirement assets. American Economic Review, 103(2), 690-731.

Note that the cross-cultural correlation has been criticised:

Roberts, S. G., Winters, J., & Chen, K. (2015). Future tense and economic decisions: controlling for cultural evolution. PloS one, 10(7), e0132145.

But the original idea provides several specific models for how grammaticalisation in language might affect perception of time. In particular, the idea relies on habitual requirements to make distinctions between the future and the present. That is, the more a specific person makes this distinction in their language, the greater the effect on their thought. (that is, there may be within-language differences).

A summary of this research is presented in this preprint (by Cole Robertson and others):

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3628501

This also includes a summary of recent experimental work on the link between FTR and time perception. Several of these experiments may be relevant to the current study, and the authors may wish to cite them (especially since they claim that little work has been done).”

Thank you for this remark. In hindsight, we should have mentioned Chen’s (2013) research and the LSH hypothesis. The intention was certainly not to leave Chen unreferenced, rather we wanted to experimentally follow-up on our own theoretical recommendations delineated in the article Jäggi et al. (2020) – where Chen’s research was an important part – and to emphasize the direction of our psycholinguistic/cognitive-psychological approach. We understand now, that without having our other article in mind, it seems odd not to reference Chen (2013). You can find the added paragraph on lines 131ff.

Jäggi, T., Sato, S., Gillioz, C., & Gygax, P. M. (2020). An Interdisciplinary Approach to Understanding the Psychological Impact of Different Grammaticalizations of the Future. Journal of Cognition, 3(1), 10. DOI: http://doi.org/10.5334/joc.100

“The preprint above also makes a point about the interpretation of future tense. While many see tense as encoding temporal distance (*when* an event occurs), it also communicates modal possibility (*whether* an event will occur). This is particularly the case for modals (e.g. in English "will" and "might" differ in probability more than they differ in temporal distance).

Cole Robertson, as part of his PhD thesis and a few papers under review, did some studies on FTR and time perception in English, Dutch and German.

One experiment was similar to the present study: Presenting participants with a phrase and asking them to indicate with a slider how far away in time they felt this was referring to. A similar experiment was also done with a slider indicating certainty.

In Robertson's study, within a language, the future tense frame did not predict the rating of temporal distance. This is in line with the current results.

However, the rating of certainty did differ significantly by grammatical form: modals were rated as less likely to occur than the future or present tenses.

Participants in Robertson's study also rated objective times (one month, two months) as a check that participants were rating things sensibly. This also agrees with the analysis of numbers in adverbials in the current study.

In addition, Dutch speakers rated events as more distant than English speakers. A similar thing could be tested in the current study. That is, is it possible to compare the scale scores in the current study for German and French directly? (i.e. the between-language prediction). It seems like the French participants are placing the slider at higher positions on average than the German participants. Isn't this what your hypothesis would predict? I appreciate the items are not exactly the same, but is it still worth noting? I'll also note that the authors are in a good position to do an experiment with bilinguals in several languages, making the comparison more effective (perhaps for future research).

Robertson continued with some experiments to show that an individual person's usage of future vs. modal strategies for talking about probability (collected in a survey similar to Dahl's survey) could be used to predict their attitude to future events (e.g. in a future discounting task).

I appreciate that these studies are not yet publicly available (though I'm sure the author would share their thesis manuscript if asked), and the point is not to undermine the current study. In fact, both lines of research seem to agree in their results. My intention is just to flag this converging evidence, and to ask the current authors whether there is a way of harnessing their current data to investigate the question about modal possibility, in addition to temporal distance.”

We appreciate this information on Cole Robertson’s research a lot and were able to get in touch with them about their thesis. We added the mentioned results to the discussion as well as the insights on modal probability, which we find very interesting. Please see lines 676ff.

While analyzing our results, we actually made the between-language comparison between Experiment 2 and 3. Unfortunately, we did not find a difference between French- and German speakers. We have now added this analysis to the manuscript. Please see lines 541ff.

Due to our experimental design, we cannot analyze the modal possibility from our current data. As all our items use the same sentence structure (adverb, subject, verb, event) and the only modal verb used in some sentences is the German ‘werden’ (to become) as part of the German future tense. We understand that adverbs can also mark possibility (e.g. possible, certain), however, we chose our adverbs to mark temporal information related to different distances from the present to the future. Therefore, we do not see how we could make a claim on modal possibility based on our current data.

“Minor points:

"These structures vary across languages and have shown to

affect how we mentally represent and perceive different aspects of our environment." - This sentence appears to have no evidence attached. It's not clear whether the citations in the previous sentence cover this. To be safe, there are some seminal works that could be easily cited here.”

Thank you for noticing this, we have changed the corresponding sentence. See lines 36f.

“"The grammaticalization of the future, which represents the grammatical manifestations of how to refer to the future, has been scarcely studied with experimental psycholinguistic methods"”

Again, we changed this sentence to clarify that other experimental research has been done. See lines 37ff.

“"As perceptual representations, we mostly refer to perceptual representations of

distance" - Does this mean distance in time? Or are you talking more generally about any kind of domain?”

Thank you for this comment. We realized that our definition of ‘perceptual representations’ was not clear enough. We have changed a great portion of this section. Please see lines 205ff.

“Lines 37-38. It's my understanding that "future time reference" refers to the act of talking about the future, rather than the linguistic devices used to do that?”

Thank you for this question. Future time reference is indeed referred to utterances related to the future. However, FTR entails different linguistic devices. We have changed the corresponding parts to make this clearer. Please see lines 68ff.

“"FTR varies in the degree of grammaticalization across languages (and at times within language): a low degree of grammaticalization is characterized by adverbials and modal verbs; and a higher degree of grammaticalization is characterized by grammatical structures embedded in the verb, like a suffix of the future tense (e.g., in French: ‘j’ irai à Paris’ [I will go to Paris])."

There is a lot of academic work on grammaticalisation from linguistics, and this could be cited here.

The French example is not illustrative unless one understands French. Please give the interlinear gloss in addition to the translation.”

Thank you for these remarks. We have changed a substantial part of introduction concerning the linguistic basis to make the manuscript clearer. We have used the interlinear gloss for our examples. Please see lines 68ff.

“"as some languages have been shown to have very low grammaticalized future verb tense and to only use modal verbs to indicate the future tense (e.g., German:"

I'm not sure this is technically correct. German can also use modal modifiers (möglich) or mental state predicates (erwarten). Maybe you're grouping these under modal verbs?”

We appreciate this comment. So far, we were mainly focusing on grammatical aspects of the future tense rather than on modal possibility. Even though, we think this is a very interesting avenue, it goes beyond the scope of the present paper. We have added a bit to the discussion when comparing with Cole Robertson’s results. Please see lines 676ff.

“"we expect the effect between low and high degree of FTR in German to be somehow stronger than in French". Why "somehow" - you have an explicit hypothesis?”

Thank you for pointing this out. We have adjusted this part in the manuscript. See line 263.

“Table 2 - the significance is easier to interpret if it's explained that PF condition was used as the intercept condition.”

We adjusted this part in the manuscript as well. See lines 372ff.

“Mixed effects models: You are modelling a scale with a floor and ceiling. Does the model take this into account? It looks like there might be floor effects for the PP condition? You could do this with e.g. logit function.”

Thank you for this comment. We wanted to add a zero-one-inflated beta model to the analyses (it belongs to the family of logit models), as it is recommended for slider data by Vuorre. Unfortunately, our data did not converge – most likely due to not enough data points – so that we could not add this analysis to the manuscript.

Please see Matti Vuorre’s R Blog: https://mvuorre.github.io/blog/posts/2019-02-18-analyze-analog-scale-ratings-with-zero-one-inflated-beta-models/

“Line 365: "it did not significantly account variance for Value" - some missing words in this sentence? Also, some statistical support for this claim should be added (e.g. difference in variance explained in a model comparison test)”

We have adjusted this part to clarify this sentence and have added statistical support. Please see lines 592ff.

The size of the text in the figures is quite small and may not reproduce well at a smaller scale.

Thank you for this remark, we enhanced the text in the figures to make them more readable. Please see images Fig1, Fig2, Fig3, Fig4 and Fig5.

8 Nov 2021

PONE-D-21-18131R1Is the future near or far depending on verb tense markers used? An experimental investigation into the effects of the grammaticalization of the futurePLOS ONE

Dear Dr. Jaeggi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Thank you for working conscientiously to address all of the reviewers' comments. From a technical/scientific viewpoint, I am happy to inform you that the paper can be accepted for publication. I have, however, processed the paper as 'minor revision' because of the issue raised by reviewer 1 about language errors. Before submitting the final version, can you please make sure you address this point, by for example having your manuscript thoroughly proof-read by the author team and possibly an external native English speaker as well, so that any language issues can be identified and corrected. Otherwise, many congratulations!

Please submit your revised manuscript by Dec 23 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Panos Athanasopoulos, Ph.D

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #3: All comments have been addressed

Reviewer #4: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #3: Yes

Reviewer #4: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #3: Yes

Reviewer #4: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I am grateful to the authors for their thorough revision of the manuscript, which I agree is much improved. I have only a one comment, and that is that a number of language errors have crept in since the revision, which need to be addressed as they are reasonably frequent and occasionally a tad jarring.

Reviewer #3: (No Response)

Reviewer #4: Thank you to the authors for responding to my questions. All my concerns have been addressed. The attempt to use the zero-inflated beta model addresses my question about floor effects. Non-convergence is not necessarily a good criteria for omitting this (and in a MCMC paradigm should be avoidable), but it's unlikely that the issue would affect the results.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #3: Yes: Gabriel Radvansky

Reviewer #4: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

8 Dec 2021

Response to Reviews: Manuscript ID PONE-D-21-18131R1

Response to Reviewer 1:

“I am grateful to the authors for their thorough revision of the manuscript, which I agree is much improved. I have only a one comment, and that is that a number of language errors have crept in since the revision, which need to be addressed as they are reasonably frequent and occasionally a tad jarring.”

Thank you for this comment. We proof-read the manuscript by a professional editor to remove the language errors.

Response to Reviewer 4:

“Thank you to the authors for responding to my questions. All my concerns have been addressed. The attempt to use the zero-inflated beta model addresses my question about floor effects. Non-convergence is not necessarily a good criteria for omitting this (and in a MCMC paradigm should be avoidable), but it's unlikely that the issue would affect the results.”

Thank you for your comment. We reanalyzed the zero-one-inflated beta models for each experiment. Although, we still did not get a viable result for our experiment with three conditions (i.e., Experiment 1, 2 and 3), for Experiment 4 we got the same result as with our mixed effects model (see Figure 1). As we did not get viable results for all experiments, we decided not to include it in the manuscript.

We followed Matti Vuorre’s R Blog for the zero-one-inflated beta model: https://mvuorre.github.io/blog/posts/2019-02-18-analyze-analog-scale-ratings-with-zero-one-inflated-beta-models/

5 Jan 2022

Is the future near or far depending on the verb tense markers used? An experimental investigation into the effects of the grammaticalization of the future

PONE-D-21-18131R2

Dear Dr. Jaeggi,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Panos Athanasopoulos, Ph.D

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

10 Jan 2022

PONE-D-21-18131R2

Is the future near or far depending on the verb tense markers used? An experimental investigation into the effects of the grammaticalization of the future

Dear Dr. Jaeggi:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Panos Athanasopoulos

Academic Editor

PLOS ONE