How language shapes our perception of time and magnitude

Language does more than carry messages between people. Different languages express time and number in different ways, and that variation appears to affect how their speakers organise temporal and numerical concepts. Are we bound by the categories our language provides, or can we think beyond them? The question has a long history under the name of linguistic relativity, and the experiments of the last two decades give it a more precise answer than either the enthusiasts or the sceptics of earlier generations expected.

Evidence and Interpretation

The evidence reviewed here concerns two domains, the spatial layout of time and the processing of number words. They are taken in that order, and each section moves from the linguistic difference to the experimental evidence for a cognitive one.

Spatial Words and the Lexicalisation of Time

One of the best-documented ways in which language influences thought is through spatial metaphors for time. Many languages describe temporal concepts with spatial terms, as in ‘a long meeting’ or ‘a short break’, and speakers of English tend to place the future ahead of them and the past behind. The mapping of time onto space varies considerably between linguistic communities, however, and that variation allows the underlying cognitive processes to be compared (Casasanto, 2008; Huang & Tse, 2017). Even within a single language, the choice of words matters. Akbuğa and Göksun (2024) asked Turkish speakers to place the events of short scenarios on a diagram that allowed both a front-back axis and a left-right axis. Scenarios written with front-back metaphors, such as an event lying ahead, drew front-back placements. Scenarios without spatial metaphors, or with metaphors of another kind, drew left-right placements, along the axis on which readers of a left-to-right script usually order earlier and later events. The axis on which people laid out time thus followed the spatial words in front of them. Exploratory analyses also related it to participants’ time-management habits and to the temporal distance and order of the events described. Findings of this kind fit the broader picture of a species whose languages differ far more than the classic search for universals allowed (Blasi et al., 2022; Evans & Levinson, 2009).

The Aymara language of the Andes provides one of the most cited cases of linguistic relativity, that is, of an influence of a language on other cognitive and perceptual domains. Núñez and Sweetser (2006) examined how Aymara speakers conceptualise time. Where English speakers, like the speakers of most languages studied, treat the future as ahead and the past as behind, Aymara speakers place the past in front of them and the future behind. The reversal appears in the language, where nayra, the word for eye, sight and front, also refers to the past. The same reversal shows in spontaneous gesture, as speakers point forward when talking about the past and over the shoulder when referring to the future. The interpretation Núñez and Sweetser favour is that the past has been seen and can be inspected, whereas the future is unknown and out of sight. The mapping would then encode a distinction between the known and the unknown, in a language whose grammar pays close attention to whether the speaker witnessed what they report. On this reading, linguistic structure and bodily experience are bound together in Aymara culture (Figure 1), each reinforcing the other.

The Aymara case also shows how quickly a single language can become a stand-in for a whole explanation. A later line of work asked whether the past-in-front mapping might reflect a culture's attitude to time as much as its words. De la Fuente et al. (2014) compared Spaniards with Moroccans, whose culture places more value on tradition and the past. Moroccan participants were more likely to place past events in front of themselves, even though Arabic, like Spanish, describes the future as ahead and the past as behind. Older Spaniards, whose focus on the past lay between that of younger Spaniards and that of Moroccans, also placed past events in front more often than younger Spaniards did. When participants’ attention was directed to their past or to their future, the mapping shifted with it. On this account, which the authors named the temporal-focus hypothesis, language is one of several influences on the spatial layout of time. De la Fuente et al. concluded that implicit mappings of time can change more readily than the spatial metaphors fixed in a language. People may therefore think about time in the opposite direction to the way they talk about it.

Mandarin Chinese offers a further example, and a well-known controversy. Mandarin uses vertical as well as horizontal metaphors for time, with earlier events ‘up’ and later events ‘down’. Last month is shàng ge yuè, literally ‘the up month’, and next month is xià ge yuè, ‘the down month’. Boroditsky (2001) reported that Mandarin speakers verified statements such as ‘March comes earlier than April’ faster after a vertical spatial prime, whereas English speakers were faster after a horizontal one. She concluded that the habitual metaphors of a language shape how its speakers think about time. The Mandarin speakers in that study were Mandarin-English bilinguals responding to English sentences, so the experiment measured a habit carried into a second language during a verbal task. Two independent teams then failed to replicate the effect, Chen (2007) in four experiments, and January and Kako (2007) in six. The debate that followed brought better designs, larger samples and non-linguistic tasks. The later studies found that Mandarin speakers do represent time vertically more readily than English speakers. The difference appeared in an implicit task in which participants judged photographs by pressing keys (Boroditsky et al., 2011). It also appeared in an explicit pointing task, in which bilinguals arranged time vertically more often the more proficient they were in Mandarin, and more often when tested in Mandarin than in English (Fuhrman et al., 2011). In the pointing task, however, the horizontal axis remained the most frequent choice for both groups, and experience with reading vertical text did not predict vertical arrangements.

The front-back axis has since received the same scrutiny. Gu et al. (2019) recorded the co-speech gestures of Mandarin speakers and found that some of them spontaneously gestured with the past in front and the future behind, especially while producing the Mandarin expressions that encode that mapping. A second experiment confirmed that Mandarin speakers can conceive of the future as behind them, and that the wording of the temporal expressions affects which mapping they use. When Gu et al. compared their participants with the Spaniards and Moroccans tested by de la Fuente et al. (2014), the Chinese participants fell between the two groups, both in their focus on the past and in how often they placed the past in front.

How Far Language Reaches Into Temporal Cognition

The research summarised so far indicates that linguistic structures influence how people represent time, but it leaves open how far that influence reaches. These effects might shape cognition in a lasting way, or they might provide alternative construals that speakers pick up and put down as the situation demands. Most of the evidence favours the second possibility. The Turkish speakers studied by Akbuğa and Göksun (2024) switched axes with the wording of each scenario, and Mandarin speakers in the study by Gu et al. (2019) moved between past-in-front and future-in-front mappings depending on the expression at hand. The Aymara case marks a limit on that flexibility. Its reversed construal is reinforced by language and gesture alike, and it is largely static, with only a few attested expressions in which time itself moves (Núñez & Sweetser, 2006). Language, on this view, supplies a set of conventional mappings that interact with other cognitive and cultural influences. In most of the cases studied, linguistic habits constrain thought only loosely, and speakers set them aside when the task calls for it.

The Influence of Language on Magnitude Perception

Beyond time, language also shapes how people handle numbers. Numerical cognition underlies everyday activities from counting objects to performing calculations, and languages structure number words in ways that affect how those words are read, written and manipulated. German offers a natural experiment, because it inverts the order of tens and units in two-digit number words, as do Dutch, Slovenian and a number of other languages. In German, 42 is spoken as ‘two and forty’, whereas English names the tens first. Steiner et al. (2021) compared German- and English-speaking children and adults on a task in which a spoken number had to be matched with its written form. German speakers were generally slower than English speakers to reject a written number that inverted the digits of the spoken one, and the effect persisted into adulthood. The cost of rejecting inverted numbers predicted arithmetic skill only among German-speaking children in their second year of school. By contrast, the general efficiency of matching a spoken number to its written form was related to arithmetic in every group except the German-speaking adults. The structure of number words in a language therefore affects how numbers are learnt and processed, and its consequences for arithmetic are clearest while the basics are still being acquired. Related work with Polish- and German-speaking preschoolers shows that inconsistencies in the grammatical marking of number, another feature that varies across languages, also hinder early numerical development (Haman et al., 2023).

Imagine learning arithmetic as a child in a language where numbers appear in one order on the page and in another in speech. The evidence suggests that calculation would be harder at first, and that the difficulty would ease with age. Adults still carry traces of the inversion. Native speakers of Arabic, which also inverts tens and units, recall fewer spoken two-digit numbers than native speakers of Hebrew when they have to type them as digits (Hayek et al., 2025). When the numbers are repeated aloud, the two groups do not differ reliably. Dutch-English bilingual adults make faster but less accurate decisions when they hear the familiar inverted Dutch number names than when they hear artificial, transparent ones (Xenidou-Dervou et al., 2024). These traces are confined to particular tasks and number ranges, however. Bahnmueller et al. (2015) found no inversion effects when German-speaking adults compared three-digit numbers. Number naming affects how arithmetic is acquired more than how it is eventually performed.

Discussion and Future Directions

Three directions would strengthen this field. The first is stricter control of confounding influences. Comparisons between languages are usually also comparisons between countries, schools and socioeconomic conditions. Cultural values are themselves associated with performance on non-verbal tasks, such as judging the accuracy of one's own responses in mental rotation (Ordin et al., 2024). These variables therefore need to be controlled statistically before an effect can be attributed to the language itself.

Interindividual variation also deserves closer attention. Speakers of the same language differ in how consistently they use its constructions, and where that variation is large, it reduces the power to detect an influence of the language on cognition. Montero-Melis (2021) accordingly recommends measuring how consistently each language applies its patterns, and treating the linguistic biases of individual speakers as predictors. Baseline biases in the non-linguistic domain under study should be measured as well.

Finally, differences within regions and language families call for a closer look (Bernabeu & Tillman, 2019). Quechua, for instance, has counterparts of the Aymara words for front and back, which in Aymara also refer to the past and the future. Whether the Quechua words carry the same meanings, or only mean ‘earlier’ and ‘later’, is disputed (Sinha & Bernárdez, 2015). The cognitive consequences of such differences between neighbouring languages remain an open question. Research resources cannot be spent on every minor contrast, but the scope of the variation should be mapped before the coarse dichotomies of earlier work, such as English against Mandarin or English against Aymara, are taken as the units of analysis.

Conclusion

The studies reviewed here are consistent with language contributing to how people represent time and numerical magnitude, alongside cultural, educational and task-specific influences. The evidence is strongest for associations that depend on context and development, and weakest for rigid or uniform effects of one language on thought. Future work should use designs that separate language from correlated social and educational factors, measure individual variation, and make cross-linguistic sampling less dependent on a few familiar contrasts.

References

Akbuğa, E., & Göksun, T. (2024). The role of spatial words in the spatialisation of time. Quarterly Journal of Experimental Psychology, 77(2), 383–392. https://doi.org/10.1177/17470218231169972

Bahnmueller, J., Moeller, K., Mann, A., & Nuerk, H.-C. (2015). On the limits of language influences on numerical cognition – no inversion effects in three-digit number magnitude processing in adults. Frontiers in Psychology, 6, Article 1216. https://doi.org/10.3389/fpsyg.2015.01216

Bernabeu, P., & Tillman, R. (2019). More refined typology and design in linguistic relativity: The case of motion event encoding. Dutch Journal of Applied Linguistics, 8(2), 163–171. https://doi.org/10.1075/dujal.15019.ber

Blasi, D. E., Henrich, J., Adamou, E., Kemmerer, D., & Majid, A. (2022). Over-reliance on English hinders cognitive science. Trends in Cognitive Sciences, 26(12), 1153–1170. https://doi.org/10.1016/j.tics.2022.09.015

Boroditsky, L. (2001). Does language shape thought? Mandarin and English speakers’ conceptions of time. Cognitive Psychology, 43(1), 1–22. https://doi.org/10.1006/cogp.2001.0748

Boroditsky, L., Fuhrman, O., & McCormick, K. (2011). Do English and Mandarin speakers think about time differently? Cognition, 118(1), 123–129. https://doi.org/10.1016/j.cognition.2010.09.010

Casasanto, D. (2008). Who's afraid of the big bad Whorf? Crosslinguistic differences in temporal language and thought. Language Learning, 58(Suppl. 1), 63–79. https://doi.org/10.1111/j.1467-9922.2008.00462.x

Chen, J.-Y. (2007). Do Chinese and English speakers think about time differently? Failure of replicating Boroditsky (2001). Cognition, 104(2), 427–436. https://doi.org/10.1016/j.cognition.2006.09.012

de la Fuente, J., Santiago, J., Román, A., Dumitrache, C., & Casasanto, D. (2014). When you think about it, your past is in front of you: How culture shapes spatial conceptions of time. Psychological Science, 25(9), 1682–1690. https://doi.org/10.1177/0956797614534695

Evans, N., & Levinson, S. C. (2009). The myth of language universals: Language diversity and its importance for cognitive science. Behavioral and Brain Sciences, 32(5), 429–448. https://doi.org/10.1017/S0140525X0999094X

Fuhrman, O., McCormick, K., Chen, E., Jiang, H., Shu, D., Mao, S., & Boroditsky, L. (2011). How linguistic and cultural forces shape conceptions of time: English and Mandarin time in 3D. Cognitive Science, 35(7), 1305–1328. https://doi.org/10.1111/j.1551-6709.2011.01193.x

Gu, Y., Zheng, Y., & Swerts, M. (2019). Which is in front of Chinese people, past or future? The effect of language and culture on temporal gestures and spatial conceptions of time. Cognitive Science, 43(12), Article e12804. https://doi.org/10.1111/cogs.12804

Haman, M., Lipowska, K., Soltanlou, M., Cipora, K., Domahs, F., & Nuerk, H.-C. (2023). The plural counts: Inconsistent grammatical number hinders numerical development in preschoolers — A cross-linguistic study. Cognition, 235, Article 105383. https://doi.org/10.1016/j.cognition.2023.105383

Hayek, M., Karni, A., & Eviatar, Z. (2025). The effect of inverting decades and units on the retention of two-digit numbers in working memory: A matter of the output mode. Psychological Research, 89(1), Article 2. https://doi.org/10.1007/s00426-024-02046-4

Huang, Y., & Tse, C.-S. (2017). Linguistic relativity in conceptual metaphors. In A. Ardila, A. B. Cieślicka, R. R. Heredia, & M. Rosselli (Eds.), Psychology of bilingualism: The cognitive and emotional world of bilinguals (pp. 3–26). Springer. https://doi.org/10.1007/978-3-319-64099-0_1

January, D., & Kako, E. (2007). Re-evaluating evidence for linguistic relativity: Reply to Boroditsky (2001). Cognition, 104(2), 417–426. https://doi.org/10.1016/j.cognition.2006.07.008

Kilobug. (2010). Aymara ceremony Copacabana 1 [Photograph]. Wikimedia Commons. https://commons.wikimedia.org/wiki/File:Aymara_ceremony_copacabana_1.jpg

Montero-Melis, G. (2021). Consistency in motion event encoding across languages. Frontiers in Psychology, 12, Article 625153. https://doi.org/10.3389/fpsyg.2021.625153

Núñez, R. E., & Sweetser, E. (2006). With the future behind them: Convergent evidence from Aymara language and gesture in the crosslinguistic comparison of spatial construals of time. Cognitive Science, 30(3), 401–450. https://doi.org/10.1207/s15516709cog0000_62

Ordin, M., El-Dakhs, D. A. S., Tao, M., Chu, F., & Polyanskaya, L. (2024). Cultural influence on metacognition: Comparison across three societies. Humanities and Social Sciences Communications, 11(1), Article 1492. https://doi.org/10.1057/s41599-024-04013-1

Sinha, C., & Bernárdez, E. (2015). Space, time, and space–time: Metaphors, maps, and fusions. In F. Sharifian (Ed.), The Routledge handbook of language and culture (pp. 309–324). Routledge. https://doi.org/10.4324/9781315793993

Steiner, A. F., Banfi, C., Finke, S., Kemény, F., Clayton, F. J., Göbel, S. M., & Landerl, K. (2021). Twenty-four or four-and-twenty: Language modulates cross-modal matching for multidigit numbers in children and adults. Journal of Experimental Child Psychology, 202, Article 104970. https://doi.org/10.1016/j.jecp.2020.104970

Xenidou-Dervou, I., van Atteveldt, N., Surducan, I. M., Reynvoet, B., Rossi, S., & Gilmore, C. (2024). Multiple number-naming associations: How the inversion property affects adults’ two-digit number processing. Quarterly Journal of Experimental Psychology, 77(4), 856–872. https://doi.org/10.1177/17470218231181367