Translating in the zone: The joint impact of positive emotions and flow on behavioral and cognitive dynamics in translation

2.1Positive emotions and flow in cognitive activities

Research has shown that positive emotions enhance multiple aspects of cognition. They promote a heuristic and global information processing style, where individuals utilize cognitive shortcuts and leverage general knowledge structures (e.g., Beukeboom and Semin 2006Beukeboom, Camiel J., and Gün R. Semin 2006 “How Mood Turns on Language.” Journal of Experimental Social Psychology 42 (5): 553–566. Beukeboom, Camiel J., and Gün R. Semin 2006 “How Mood Turns on Language.” Journal of Experimental Social Psychology 42 (5): 553–566. ). They also facilitate assimilative processing, whereby new information is more readily integrated into existing mental representations (Bohn-Gettler 2019Bohn-Gettler, Catherine M. 2019 “Getting a Grip: The PET Framework for Studying How Reader Emotions Influence Comprehension.” Discourse Processes 56 (5–6): 386–401. Bohn-Gettler, Catherine M. 2019 “Getting a Grip: The PET Framework for Studying How Reader Emotions Influence Comprehension.” Discourse Processes 56 (5–6): 386–401. ).

Fredrickson’s (2001)Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. broaden-and-build theory provides a systematic framework for understanding how positive emotions enhance human cognition, action, and personal resources. This theory posits that positive emotions serve two functions: they broaden individuals’ momentary thought-action repertoires, and they build enduring personal resources across physical, intellectual, social, and psychological domains. While the broadening effect widens the range of thoughts and actions in individuals’ minds, the building function enhances long-term resilience and well-being by offsetting negative emotions and triggering upward spirals. In addition, Fredrickson (2001)Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. identified a reciprocal relationship between positive emotions, broadened thinking, and the ability to find positive meaning in everyday events, indicating that the effects of positive emotions can accumulate and compound over time.

Csikszentmihalyi (1990)Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins. explored the positive aspects of human experience and conceptualized the state of total involvement with life as flow. Flow is an optimal mental state of deep immersion, where individuals lose awareness of their surroundings and their sense of self. This state is often described as highly enjoyable and intrinsically rewarding, motivating individuals to repeatedly seek out such experiences. Csikszentmihalyi (1990)Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins. identified nine characteristics of flow: action-awareness merging; concentration on task; loss of self-consciousness; a sense of control; challenge-skill balance; clear goals; unambiguous feedback; distorted sense of time; and autotelic experience. Among these dimensions, the balance between challenge and skill is particularly critical for inducing flow. According to the eight-channel model of flow experience, flow occurs only when both perceived challenges and skills are high and in balance (Nakamura and Csikszentmihalyi 2021Nakamura, Jeanne, and Mihaly Csikszentmihalyi 2021 “The Experience of Flow: Theory and Research.” In The Oxford Handbook of Positive Psychology, edited by C. R. Snyder, Shane J. Lopez, Lisa M. Edwards, and Susana C. Marques, 279–296. New York: Oxford University Press.Nakamura, Jeanne, and Mihaly Csikszentmihalyi 2021 “The Experience of Flow: Theory and Research.” In The Oxford Handbook of Positive Psychology, edited by C. R. Snyder, Shane J. Lopez, Lisa M. Edwards, and Susana C. Marques, 279–296. New York: Oxford University Press.).

Flow has profound effects on human cognition. Csikszentmihalyi (1999)Csikszentmihalyi, Mihaly 1999 “If We Are So Rich, Why Aren’t We Happy?” American Psychologist 54 (10): 821–827. Csikszentmihalyi, Mihaly 1999 “If We Are So Rich, Why Aren’t We Happy?” American Psychologist 54 (10): 821–827. highlighted its relationship with attention, describing flow as a state characterized by a complete absence of distraction, where all attention is fully devoted to the task at hand, leaving no capacity to process irrelevant stimuli. Neurocognitive research indicates that flow is associated with reduced activation of the dorsal attention network and temporary suppression of analytical and meta-conscious capacities, resulting in more efficient and effortless cognitive functioning (Dietrich 2004Dietrich, Arne 2004 “Neurocognitive Mechanisms Underlying the Experience of Flow.” Consciousness and Cognition 13 (4): 746–761. Dietrich, Arne 2004 “Neurocognitive Mechanisms Underlying the Experience of Flow.” Consciousness and Cognition 13 (4): 746–761. ; Wang et al. 2023Wang, Yue, Buxin Han, Min Li, Juan Li, and Rui Li 2023 “An Efficiently Working Brain Characterizes Higher Mental Flow That Elicits Pleasure in Chinese Calligraphic Handwriting.” Cerebral Cortex 33 (12): 7395–7408. Wang, Yue, Buxin Han, Min Li, Juan Li, and Rui Li 2023 “An Efficiently Working Brain Characterizes Higher Mental Flow That Elicits Pleasure in Chinese Calligraphic Handwriting.” Cerebral Cortex 33 (12): 7395–7408. ).

Flow experiences are typically accompanied by positive feelings arising from a sense of accomplishment and self-esteem (Csikszentmihalyi 1990Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.); however, they are conceptually distinct from positive emotions. While positive emotions can be conceptualized as brief, multicomponent positive responses to specific events (Fredrickson 2001Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. ), flow refers to a more extended and structured optimal experience that additionally depends on task-related conditions such as challenge-skill balance, clear goals, and a sense of control (Csikszentmihalyi 1990Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.). The cognitive and motivational gains associated with positive emotions can thus reasonably be expected to increase the likelihood that individuals enter flow states, a view supported by findings that foreign language enjoyment predicts students’ flow experiences (Dewaele and MacIntyre 2022Dewaele, Jean-Marc, and Peter MacIntyre 2022 “ ‘You Can’t Start a Fire Without a Spark’. Enjoyment, Anxiety, and the Emergence of Flow in Foreign Language Classrooms.” Applied Linguistics Review 15 (2): 403–426. Dewaele, Jean-Marc, and Peter MacIntyre 2022 “ ‘You Can’t Start a Fire Without a Spark’. Enjoyment, Anxiety, and the Emergence of Flow in Foreign Language Classrooms.” Applied Linguistics Review 15 (2): 403–426. ).

2.2Positive emotions and flow in translation

Research into translation activities has explored how positive emotions originating in source texts and in translators shape translation behavior and performance. For text-driven emotions, Lehr and Hvelplund (2020)Lehr, Caroline, and Kristian Tangsgaard Hvelplund 2020 “Emotional Experts: Influences of Emotion on the Allocation of Cognitive Resources during Translation.” In Multilingual Mediated Communication and Cognition, edited by Ricardo Muñoz Martín and Sandra L. Halverson, 44–68. London: Routledge. Lehr, Caroline, and Kristian Tangsgaard Hvelplund 2020 “Emotional Experts: Influences of Emotion on the Allocation of Cognitive Resources during Translation.” In Multilingual Mediated Communication and Cognition, edited by Ricardo Muñoz Martín and Sandra L. Halverson, 44–68. London: Routledge. showed that compared with a source text (ST) containing negative emotion-laden words, a ST with positive emotion-laden words attracted more attention during ST reading (longer fixation durations) yet reduced target text (TT) production load and required less attention to the TT (smaller pupils and shorter fixations). Their results suggest that positive content initially promotes engagement and then facilitates more efficient, less cognitively demanding production. Complementing this, research on emotional framing effects indicates that positive frames of the ST influence evaluative interpretation and strategy selection: translators exposed to optimistic framing use fewer emphasis and attenuation translation strategies than those exposed to pessimistic framing (Rojo and Naranjo 2021Rojo López, Ana María, and Beatriz Naranjo 2021 “Translating in Times of Crisis: A Study about the Emotional Effects of the COVID19 Pandemic on the Translation of Evaluative Language.” Journal of Pragmatics 176: 29–40. Rojo López, Ana María, and Beatriz Naranjo 2021 “Translating in Times of Crisis: A Study about the Emotional Effects of the COVID19 Pandemic on the Translation of Evaluative Language.” Journal of Pragmatics 176: 29–40. ).

A second strand highlights emotional (in)congruence between the ST and translators’ emotions. Congruent primes, such as words aligned with a translator’s ideological stance, shorten the reaction time to find a translation solution relative to incongruent primes (Rojo and Ramos 2014Rojo López, Ana María, and Marina Ramos Caro 2014 “The Impact of Translators’ Ideology on the Translation Process: A Reaction Time Experiment.” MonTI: 247–271. Rojo López, Ana María, and Marina Ramos Caro 2014 “The Impact of Translators’ Ideology on the Translation Process: A Reaction Time Experiment.” MonTI: 247–271. ). Likewise, congruence between background music and the ST’s narrative has been associated with higher translation creativity (Naranjo and Rojo 2021Naranjo, Beatriz, and Ana María Rojo López 2021 “In and out of Tune: The Effects of Musical (In)Congruence on Translation.” Target 33 (1): 132–156. Naranjo, Beatriz, and Ana María Rojo López 2021 “In and out of Tune: The Effects of Musical (In)Congruence on Translation.” Target 33 (1): 132–156. ). Beyond text-driven emotions, studies have experimentally induced positive and negative emotions in translators by providing bogus evaluative feedback on an initial translation and then assessing the translators’ performance on a subsequent task. These studies found that positive emotions led to higher creativity ratings in terms of meaning, pragmatics, and style, suggesting a role for such emotions in promoting cognitive breadth and flexibility (Lehr 2014Lehr, Caroline 2014 The Influence of Emotion on Language Performance: Study of a Neglected Determinant of Decision-Making in Professional Translators. PhD diss. University of Geneva.Lehr, Caroline 2014 The Influence of Emotion on Language Performance: Study of a Neglected Determinant of Decision-Making in Professional Translators. PhD diss. University of Geneva.; Rojo and Ramos 2016Rojo López, Ana María, and Marina Ramos Caro 2016 “Can Emotion Stir Translation Skill? Defining the Impact of Positive and Negative Emotions on Translation Performance.” In Reembedding Translation Process Research, edited by Ricardo Muñoz Martín, 107–130. Amsterdam: John Benjamins. Rojo López, Ana María, and Marina Ramos Caro 2016 “Can Emotion Stir Translation Skill? Defining the Impact of Positive and Negative Emotions on Translation Performance.” In Reembedding Translation Process Research, edited by Ricardo Muñoz Martín, 107–130. Amsterdam: John Benjamins. ).

Flow experiences have also been documented in translation activities, revealing distinctive influencing factors and task-specific characteristics. Empirical evidence from Mirlohi, Egbert, and Ghonsooly (2011)Mirlohi, Mehdi, Joy Egbert, and Behzad Ghonsooly 2011 “Flow in Translation: Exploring Optimal Experience for Translation Trainees.” Target 23 (2): 251–271. Mirlohi, Mehdi, Joy Egbert, and Behzad Ghonsooly 2011 “Flow in Translation: Exploring Optimal Experience for Translation Trainees.” Target 23 (2): 251–271. confirmed the presence of flow in translation classrooms and found that text genre significantly influenced flow, with descriptive texts triggering it more effectively than expository or narrative ones. Weng and Li (2025)Weng, Yu, and Yanze Li 2025 汉英翻译中的译者心流体验：积极心理机制与认知特征探析 [Flow experience in Chinese–English translation: Positive psychological mechanisms and cognitive characteristics]. Foreign Language Teaching and Research 57 (2): 290–303.Weng, Yu, and Yanze Li 2025 汉英翻译中的译者心流体验：积极心理机制与认知特征探析 [Flow experience in Chinese–English translation: Positive psychological mechanisms and cognitive characteristics]. Foreign Language Teaching and Research 57 (2): 290–303. further characterized flow and identified three key features of flow among student translators: challenge-skill balance, a sense of control, and loss of self-consciousness. Interestingly, they also found that these students, regardless of experiencing flow, generally exhibited certain positive psychological traits during translation, including goal-oriented self-efficacy, autotelic immersion that ignores the passage of time, and the merging of actions and awareness.

While research has confirmed that positive emotions benefit attention and creativity in translation and that flow states occur during translation, little is known about how positive emotions trigger and sustain flow, or how they interact with flow to shape translators’ behavioral and cognitive responses. Broaden-and-build theory posits that positive emotions expand cognitive and attentional capacities (Fredrickson 2001Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. ), thereby creating favorable conditions for entering and maintaining flow. During translation, where comprehension and production must be tightly coordinated, positive emotions may help translators meet these demands by providing cognitive and motivational support that facilitates challenge-skill alignment, goal focus, and a sense of effortless control. Flow, in turn, tends to reinforce positive emotions through its intrinsic reward, establishing a virtuous cycle during translation.

From a process perspective, positive emotions and flow are expected to jointly shape how translation proceeds. Positive emotions widen attentional scope, increase cognitive flexibility, and facilitate the retrieval of preexisting general knowledge structures (Fredrickson 2001Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. ; Lyubomirsky, King, and Diener 2005Lyubomirsky, Sonja, Laura King, and Ed Diener 2005 “The Benefits of Frequent Positive Affect: Does Happiness Lead to Success?” Psychological Bulletin 131 (6): 803–855. Lyubomirsky, Sonja, Laura King, and Ed Diener 2005 “The Benefits of Frequent Positive Affect: Does Happiness Lead to Success?” Psychological Bulletin 131 (6): 803–855. ), which supports richer intake of task-relevant information and easier shifts among plausible translation solutions. Flow then channels these broadened resources toward task goals, eliciting spontaneous processing and limiting distractions (Csikszentmihalyi 1990Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.). In synergy, they speed the path from comprehension to production: more information is integrated at each step, potential problems are resolved earlier without explicit re-evaluation, and the overall processing effort is reduced. In practice, this could appear as less frequent re-checking, smoother progression from ST to TT, a more balanced distribution of attention between reading and formulating, and, where possible, more moments of parallel processing of the two streams. Crucially, these gains do not rely on shortcuts that risk translation quality; rather, positive emotions encourage efficient use of heuristic knowledge, while flow sustains automaticity and concentration. Over time, consistent with broaden-and-build theory, these benefits may accumulate across repeated tasks, yielding a translation process that is increasingly efficient, less effortful, and better coordinated.

2.3Operationalizing behavioral and cognitive dynamics in translation via eye-tracking measures

Eye-tracking is an established method that is highly sensitive to online perceptual and cognitive aspects of reading (Rayner 1998Rayner, Keith 1998 “Eye Movements in Reading and Information Processing: 20 Years of Research.” Psychological Bulletin 124 (3): 372–422. Rayner, Keith 1998 “Eye Movements in Reading and Information Processing: 20 Years of Research.” Psychological Bulletin 124 (3): 372–422. ), making it well-suited to tracing how translation unfolds moment by moment. Saccades, rapid eye movements between fixations, index reading speed; larger saccade amplitudes correlate with faster, more integrated information uptake and thus serve as markers of processing efficiency (Krieber et al. 2016Krieber, Magdalena, Katrin D. Bartl-Pokorny, Florian B. Pokorny, Christa Einspieler, Andrea Langmann, Christof Körner, Terje Falck-Ytter, and Peter B. Marschik 2016 “The Relation between Reading Skills and Eye Movement Patterns in Adolescent Readers: Evidence from a Regular Orthography.” PLOS ONE 11 (1): e0145934. Krieber, Magdalena, Katrin D. Bartl-Pokorny, Florian B. Pokorny, Christa Einspieler, Andrea Langmann, Christof Körner, Terje Falck-Ytter, and Peter B. Marschik 2016 “The Relation between Reading Skills and Eye Movement Patterns in Adolescent Readers: Evidence from a Regular Orthography.” PLOS ONE 11 (1): e0145934. ). Regressions, or backward saccades, signal the need for reprocessing; fewer such returns suggest a leaner monitoring load (Rayner 1998Rayner, Keith 1998 “Eye Movements in Reading and Information Processing: 20 Years of Research.” Psychological Bulletin 124 (3): 372–422. Rayner, Keith 1998 “Eye Movements in Reading and Information Processing: 20 Years of Research.” Psychological Bulletin 124 (3): 372–422. ). During translation, gaze also shifts between ST and TT, revealing the coordination of comprehension and production. Professional translators typically exhibit an integrated coordination mode, minimizing ST–TT transitions and achieving instant, sometimes parallel transformation from ST to TT, whereas less-experienced student translators show more frequent transitions and pauses (Dragsted 2010Dragsted, Barbara 2010 “Coordination of Reading and Writing Processes in Translation: An Eye on Uncharted Territory.” In Translation and Cognition, edited by Gregory M. Shreve and Erik Angelone, 41–62. Amsterdam: John Benjamins. Dragsted, Barbara 2010 “Coordination of Reading and Writing Processes in Translation: An Eye on Uncharted Territory.” In Translation and Cognition, edited by Gregory M. Shreve and Erik Angelone, 41–62. Amsterdam: John Benjamins. ). There is evidence that emotions modulate these eye-movement patterns in language tasks: in a long-text reading study using video-based emotion induction, participants showed larger saccade amplitudes under positive priming, while amplitudes remained stable under negative priming (Mills, Southwell, and D’Mello 2024Mills, Caitlin, Rosy Southwell, and Sidney K. D’Mello 2024 “Sadness Facilitates ‘Deeper’ Reading Comprehension: A Behavioural and Eye Tracking Study.” Cognition and Emotion 38 (1): 171–179. Mills, Caitlin, Rosy Southwell, and Sidney K. D’Mello 2024 “Sadness Facilitates ‘Deeper’ Reading Comprehension: A Behavioural and Eye Tracking Study.” Cognition and Emotion 38 (1): 171–179. ). This pattern has been interpreted as positive emotions prompting a shift toward less detail-oriented, faster reading strategies. Moreover, regressive fixations increased over time in the neutral condition but were comparatively stable under positive or negative priming, indicating emotion-dependent differences in sustained reading (Mills, Southwell, and D’Mello 2024Mills, Caitlin, Rosy Southwell, and Sidney K. D’Mello 2024 “Sadness Facilitates ‘Deeper’ Reading Comprehension: A Behavioural and Eye Tracking Study.” Cognition and Emotion 38 (1): 171–179. Mills, Caitlin, Rosy Southwell, and Sidney K. D’Mello 2024 “Sadness Facilitates ‘Deeper’ Reading Comprehension: A Behavioural and Eye Tracking Study.” Cognition and Emotion 38 (1): 171–179. ). Although no study has directly examined emotional influences on ST–TT transitions in translation, the broadening function of positive emotions — greater cognitive flexibility and attentional breadth (Fredrickson 2001Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. ) — suggests that translators may require fewer reviews of the ST, implying a potential reduced frequency of ST–TT transitions.

Complementing indicators of reading pace and reanalysis, fixation-based measures index cognitive effort. According to the eye–mind assumption, fixation location reveals the object being processed, and fixation duration reflects the processing time (Just and Carpenter 1980Just, Marcel Adam, and Patricia A. Carpenter 1980 “A Theory of Reading: From Eye Fixations to Comprehension.” Psychological Review 87 (4): 329–354. Just, Marcel Adam, and Patricia A. Carpenter 1980 “A Theory of Reading: From Eye Fixations to Comprehension.” Psychological Review 87 (4): 329–354. ). Longer fixations generally correspond to greater cognitive effort and resource investment, whereas shorter fixations indicate less allocation of cognitive resources and reduced cognitive effort. In translation tasks with emotionally valenced STs, mean fixation duration during full-text ST reading was longer for positively toned texts than for negatively toned texts (consistent with the approach behavior commonly associated with positive emotions); by contrast, during subsequent TT reading, fixations were shorter for the TTs produced from positive STs than for those produced from negative STs, suggesting a lower allocation of cognitive resources when working with positive materials (Lehr and Hvelplund 2020Lehr, Caroline, and Kristian Tangsgaard Hvelplund 2020 “Emotional Experts: Influences of Emotion on the Allocation of Cognitive Resources during Translation.” In Multilingual Mediated Communication and Cognition, edited by Ricardo Muñoz Martín and Sandra L. Halverson, 44–68. London: Routledge. Lehr, Caroline, and Kristian Tangsgaard Hvelplund 2020 “Emotional Experts: Influences of Emotion on the Allocation of Cognitive Resources during Translation.” In Multilingual Mediated Communication and Cognition, edited by Ricardo Muñoz Martín and Sandra L. Halverson, 44–68. London: Routledge. ).

Beyond cognitive resource investment, attention allocation between ST and TT can be characterized using two complementary measures. First, the ST–TT fixation duration ratio, calculated as the ratio of fixation durations on ST relative to TT, indexes how attention is distributed across comprehension and production. Values closer to 1 indicate a more balanced allocation, whereas pronounced skews toward either side reflect a stronger focus on one phase. Second, parallel attention units (PAUs) capture gazes on the ST that coincide with active typing of the TT, identified via synchronized eye-tracking and key-logging data (Hvelplund 2011aHvelplund, Kristian Tangsgaard 2011a Allocation of Cognitive Resources in Translation: An Eye-Tracking and Key-Logging Study. PhD diss. Copenhagen Business School.Hvelplund, Kristian Tangsgaard 2011a Allocation of Cognitive Resources in Translation: An Eye-Tracking and Key-Logging Study. PhD diss. Copenhagen Business School.). PAU count percent is computed as the proportion of PAUs relative to the total number of ST attention units, TT attention units, and PAUs. Direct evidence of emotional influences on attention allocation between ST and TT in translation remains limited; nevertheless, the documented benefits of positive emotions (e.g., Fredrickson 2001Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. Fredrickson, Barbara L. 2001 “The Role of Positive Emotions in Positive Psychology.” The American Psychologist 56 (3): 218–226. ) make it plausible that translators adopt a more global focus and distribute attention more evenly between ST and TT.

3.1Aims and hypotheses

The present study examines whether experimentally induced positive emotions, relative to a neutral induction, increase the likelihood of entering flow states during translation, and how positive emotions and flow jointly shape process dynamics captured by eye-tracking measures. To this end, two experimental conditions were designed: translating under positive emotional priming and translating under neutral priming (which served as a baseline to isolate the effect of positive priming). The following five hypotheses were formulated based on the theoretical framework and prior findings discussed above:

3.2Participants

A total of thirty-two11.Given that accurate a priori power analysis for linear mixed-effects models requires pre-specifying the full variance-covariance structure, including random intercepts, random slopes, and their covariances (Raudenbush and Liu 2000Raudenbush, Stephen W., and Xiaofeng Liu 2000 “Statistical Power and Optimal Design for Multisite Randomized Trials.” Psychological Methods 5 (2): 199–213. Raudenbush, Stephen W., and Xiaofeng Liu 2000 “Statistical Power and Optimal Design for Multisite Randomized Trials.” Psychological Methods 5 (2): 199–213. ), and that reliable prior estimates for this structure were not available in our context, the sample size was determined with reference to empirical precedents in the affective priming literature (e.g., Kim et al. 2021Kim, Minchul, Jeeyeon Kim, Jaejoong Kim, and Bumseok Jeong 2021 “Diffusion Model-Based Understanding of Subliminal Affective Priming in Continuous Flash Suppression.” Scientific Reports 11 (1): 11534. Kim, Minchul, Jeeyeon Kim, Jaejoong Kim, and Bumseok Jeong 2021 “Diffusion Model-Based Understanding of Subliminal Affective Priming in Continuous Flash Suppression.” Scientific Reports 11 (1): 11534. ). fourth-year undergraduate students majoring in English at South China University of Technology were recruited to participate in the study on a voluntary basis.22.The research was approved by the ethics committee of the School of Foreign Languages, South China University of Technology. To avoid emotional carryover effects, participants were randomly assigned to either the positive emotion (PE) or neutral emotion (NE) stimulus group. Each participant completed two translation tasks, generating sixty-four observations. Following data screening, twenty-nine participants (90.63%) were included in the analysis (see Section 3.6 for details of data screening procedures), with fourteen in the PE group and fifteen in the NE group. All participants were native Chinese speakers with English as their second language. They had completed two semesters of translation courses and scored above 70 on the Test for English Majors-Band 4 (a national English proficiency assessment mandatory for English majors in China). Participants were verified as touch-typists in a pre-experiment session and had normal or corrected-to-normal vision. Each participant provided informed consent with guaranteed anonymity and confidentiality, and received CNY 50 as compensation for their participation.

3.3Emotional stimuli

Videos are widely recognized as effective emotion elicitation tools, with meta-analytic work reporting larger and more consistent effect sizes for film clips as induction stimuli than for alternatives such as music, interaction, or gifts (Westermann et al. 1996Westermann, Rainer, Kordelia Spies, Günter Stahl, and Friedrich W. Hesse 1996 “Relative Effectiveness and Validity of Mood Induction Procedures: A Meta-Analysis.” European Journal of Social Psychology 26: 557–580. Westermann, Rainer, Kordelia Spies, Günter Stahl, and Friedrich W. Hesse 1996 “Relative Effectiveness and Validity of Mood Induction Procedures: A Meta-Analysis.” European Journal of Social Psychology 26: 557–580. ). Compared with static images or text-based stimuli, the multimodal and dynamic nature of videos affords greater ecological validity by approximating context-rich, real-world-like emotional situations and by sustaining affect over minutes (Gross and Levenson 1995Gross, James J., and Robert W. Levenson 1995 “Emotion Elicitation Using Films.” Cognition & Emotion 9 (1): 87–108. Gross, James J., and Robert W. Levenson 1995 “Emotion Elicitation Using Films.” Cognition & Emotion 9 (1): 87–108. ), which matches the temporal demands of translation tasks. Two video clips with bilingual subtitles were used to trigger either positive or neutral emotions at the beginning of the experiment. The positive stimulus was a 2-minute 47-second segment from The Big Bang Theory, featuring a humorous scene in which the main character, Sheldon, pronounces the names of Chinese dishes with a hilarious accent. The neutral stimulus was a 2-minute 44-second segment from The Universe: Secrets of the Sun, presenting a factual narrative about solar formation. Both clips were selected for their accessibility, containing no obscure concepts or terms that might impede viewer immersion.

To verify stimulus effectiveness, twenty-one postgraduate students (not participating in the main experiment) rated both clips using three positive emotion items (interest, joy, and surprise) from the 5-point Differential Emotions Scale (DES) (Izard et al. 1974Izard, Carroll Ellis, F. E. Dougherty, B. M. Bloxom, and N. E. Kotsch 1974 The Differential Emotions Scale: A Method of Measuring the Subjective Experience of Discrete Emotions. Nashville: Vanderbilt University Press.Izard, Carroll Ellis, F. E. Dougherty, B. M. Bloxom, and N. E. Kotsch 1974 The Differential Emotions Scale: A Method of Measuring the Subjective Experience of Discrete Emotions. Nashville: Vanderbilt University Press.), ranging from 1 (Not at all) to 5 (An extreme amount). A paired sample t-test showed significantly higher positive emotion scores (t(20) = 8.59, p < .001) for the comedy clip (M = 12.29, SD = 2.31) compared to the documentary (M = 8.33, SD = 2.60).

3.4Source texts

The two comparable source texts used for the two Chinese-to-English translation tasks were excerpted from the China Accreditation Test for Translators and Interpreters, Level 3 (2015, November). Both texts discussed Sino–U.S. major-country relations and were rotated across task sessions. Text comparability was established through Chinese readability indices calculated using CRIE 3.0 (Sung et al. 2016Sung, Yao-Ting, Tao-Hsing Chang, Wei-Chun Lin, Kuan-Sheng Hsieh, and Kuo-En Chang 2016 “CRIE: An Automated Analyzer for Chinese Texts.” Behavior Research Methods 48 (4): 1238–1251. Sung, Yao-Ting, Tao-Hsing Chang, Wei-Chun Lin, Kuan-Sheng Hsieh, and Kuo-En Chang 2016 “CRIE: An Automated Analyzer for Chinese Texts.” Behavior Research Methods 48 (4): 1238–1251. ), and subjective translation difficulty ratings with a 5-point Likert scale from twelve postgraduate students not involved in the main experiment (see Table 1). The texts were selected to match participants’ language and translation competence levels, aligning with their translation course curriculum.

3.5Procedures and setup

Following the provision of informed consent and a comprehensive briefing about the experimental procedures, participants completed a typing test to verify their touch-typing ability and familiarize them with the experimental interfaces and devices. The experiment commenced with participants viewing their designated stimulus video clip, after which they completed the aforementioned three-item DES positive emotion scale.

The translation tasks were then completed on the Translog II software (Carl 2012Carl, Michael 2012 “Translog-II: A Program for Recording User Activity Data for Empirical Translation Process Research.” In Proceedings of the 8th International Conference on Language Resources and Evaluation, edtied by Nicoletta Calzolari, Khalid Choukri, Thierry Declerck, Mehmet Uğur Doğan, Bente Maegaard, Joseph Mariani, Asuncion Moreno, Jan Odijk, and Stelios Piperidis, 153–162. Paris: European Language Resources Association.Carl, Michael 2012 “Translog-II: A Program for Recording User Activity Data for Empirical Translation Process Research.” In Proceedings of the 8th International Conference on Language Resources and Evaluation, edtied by Nicoletta Calzolari, Khalid Choukri, Thierry Declerck, Mehmet Uğur Doğan, Bente Maegaard, Joseph Mariani, Asuncion Moreno, Jan Odijk, and Stelios Piperidis, 153–162. Paris: European Language Resources Association.), which displayed the source text at the top of the screen and provided a target text input box at the bottom. To control for potential confounding variables, external resource consultation was not permitted. Instead, participants were provided with glossaries containing four pre-identified challenging terms for each text. These glossaries were made available before each translation task to allow familiarization with the terms but were not accessible during the translation process.

Throughout the translation process, an EyeLink Portable Duo eye tracker (1000 Hz sampling rate) recorded participants’ eye movements in a dedicated eye-tracking laboratory. The laboratory was equipped with controlled lighting conditions and sound insulation to minimize external interference. Participants were seated approximately 65 cm from a 24-inch LCD monitor (1920 × 1080 resolution) in an adjustable chair during the experiment.

Immediately following each translation task, participants completed both the three-item DES positive emotion scale and the Short Flow State Scale (Martin and Jackson 2008Martin, Andrew J., and Susan A. Jackson 2008 “Brief Approaches to Assessing Task Absorption and Enhanced Subjective Experience: Examining ‘Short’ and ‘Core’ Flow in Diverse Performance Domains.” Motivation and Emotion 32 (3): 141–157. Martin, Andrew J., and Susan A. Jackson 2008 “Brief Approaches to Assessing Task Absorption and Enhanced Subjective Experience: Examining ‘Short’ and ‘Core’ Flow in Diverse Performance Domains.” Motivation and Emotion 32 (3): 141–157. ) to assess their emotional and flow states. The flow state scale comprises nine items, each corresponding to one dimension of flow experiences. Participants responded to these items using a 5-point Likert scale, ranging from 1 (I strongly disagree) to 5 (I strongly agree). These procedures resulted in three self-reported emotional state measurements (emotion0 following the video stimulus, emotion1 and emotion2 following each translation task) and two flow measurements (flow1 and flow2 following each translation task). The whole experiment lasted about an hour.

3.6Eye-tracking data screening

Eye-tracking data quality was evaluated using three metrics: mean fixation duration (MFD), gaze time on screen (GTS), and gaze sample to fixation percentage (GFP). GTS represents the percentage of time spent gazing at the screen relative to total task time, while GFP is calculated as the ratio of gaze samples forming fixations to total gaze samples collected. Following established thresholds (e.g., Hvelplund 2011aHvelplund, Kristian Tangsgaard 2011a Allocation of Cognitive Resources in Translation: An Eye-Tracking and Key-Logging Study. PhD diss. Copenhagen Business School.Hvelplund, Kristian Tangsgaard 2011a Allocation of Cognitive Resources in Translation: An Eye-Tracking and Key-Logging Study. PhD diss. Copenhagen Business School.), individual eye-tracking files were required to meet at least two of the three criteria: MFD ≥ 200 ms, GTS ≥ 49.4%, and GFP ≥ 75.7%, with the latter two thresholds corresponding to M −2 SD. If one file failed to meet these criteria, both files from the same participant were excluded. Based on these standards, six files from three participants (9.4% of the total 64 files from 32 participants) were discarded.

3.7Statistical models and analysis procedures

Statistical analyses included linear regression, linear mixed-effects regression (LMER), and correlation analysis as appropriate for each hypothesis test. LMER models were constructed using the lme4 package (Bates et al. 2015Bates, Douglas, Martin Mächler, Benjamin M. Bolker, and Steve C. Walker 2015 “Fitting Linear Mixed-Effects Models Using Lme4.” Journal of Statistical Software 67 (1): 1–48. Bates, Douglas, Martin Mächler, Benjamin M. Bolker, and Steve C. Walker 2015 “Fitting Linear Mixed-Effects Models Using Lme4.” Journal of Statistical Software 67 (1): 1–48. ) in R software, version 4.3.3. For each target measure, separate regression models incorporated fixed effects of stimulus type (PE and NE), flow (total flow scale scores), and task sequence (Task 1 and Task 2), along with relevant interactions. The models accounted for individual variability by including a random intercept for participants. P-values of fixed and interaction effects were calculated using Satterthwaite’s approximation through the lmerTest package (Kuznetsova, Brockhoff, and Christensen 2017Kuznetsova, Alexandra, Per B. Brockhoff, and Rune H. B. Christensen 2017 “lmerTest Package: Tests in Linear Mixed Effects Models.” Journal of Statistical Software 82 (13): 1–26. Kuznetsova, Alexandra, Per B. Brockhoff, and Rune H. B. Christensen 2017 “lmerTest Package: Tests in Linear Mixed Effects Models.” Journal of Statistical Software 82 (13): 1–26. ). For the PAU count percent measure, which is bounded between 0 and 1, beta regression models with relevant fixed effects and participant as a random intercept were implemented using the glmmTMB package (Brooks et al. 2017Brooks, Mollie E., Kasper Kristensen, Koen J. van Benthem, Arni Magnusson, Casper W. Berg, Anders Nielsen, Hans J. Skaug, Martin Mächler, and Benjamin M. Bolker 2017 “glmmTMB Balances Speed and Flexibility among Packages for Zero-Inflated Generalized Linear Mixed Modeling.” The R Journal 9 (2): 378–400. Brooks, Mollie E., Kasper Kristensen, Koen J. van Benthem, Arni Magnusson, Casper W. Berg, Anders Nielsen, Hans J. Skaug, Martin Mächler, and Benjamin M. Bolker 2017 “glmmTMB Balances Speed and Flexibility among Packages for Zero-Inflated Generalized Linear Mixed Modeling.” The R Journal 9 (2): 378–400. ).

4.1Relationship between flow experiences and positive emotions

Analysis of self-reported positive emotion values (emotion0) indicates successful emotion manipulation at the outset of the experiment. Linear regression revealed that immediately after viewing the stimulus videos, the PE group reported significantly higher positive emotion values compared to the NE group (β = 3.00, SE = .77, t = 3.91, p < .001), as shown in Figure 1a. However, while the PE group maintained higher positive emotion values (emotion1 and emotion2) following both translation tasks compared to the NE group (Figures 1b and 1c), these differences were no longer statistically significant, potentially due to the cognitive demands of translation diminishing the initial emotional effect.

Regarding the relationship between positive emotions and flow, a LMER model with stimulus as a predictor revealed no significant impact of stimulus type on flow scores following the translation tasks (β = 1.36, SE = 1.28, t = 1.07, p = .296), though the PE group generally exhibited higher flow scores (Figure 1d).

No significant correlations were found between positive emotion scores and total flow scores after each translation task as well. However, analysis of specific flow dimensions revealed targeted correlations with positive emotions. In Task 1, participants’ positive emotion scores (emotion1) correlated significantly with three flow dimensions: action-awareness merging (r = .38, p = .031), concentration on task (r = −.45, p = .009), and autotelic experience (r = .45, p = .010). In Task 2, positive emotion scores (emotion2) correlated significantly with two dimensions: action-awareness merging (r = .36, p = .044) and autotelic experience (r = .49, p = .004). Additionally, initial positive emotion scores (emotion0) showed significant correlations with distorted sense of time in Task 1 (r = .40, p = .023) and action-awareness merging in Task 2 (r = .44, p = .013).

Overall, regarding H1, which proposes that flow experiences are more prominently triggered by primed positive emotions, our findings only offer partial support. While neither positive emotion scores nor stimulus type significantly affected total flow scores, specific flow dimensions showed consistent correlations with translators’ positive emotional states.

4.2Interplay between flow experiences and stimulus type in the translation process

While total flow scores were not significantly affected by stimulus type, we examined whether stimulus type interacted with flow experiences to influence behavioral patterns and cognitive processes (H2 to H4). LMER models testing H2 revealed that stimulus type significantly interacted with flow scores to affect ST–TT transition (β = −4.93, SE = 2.15, t = −2.30, p = .026), with a marginally significant interaction for mean saccade amplitude (β = .19, SE = .10, t = 1.81, p = .076). In the PE condition, higher flow scores corresponded with fewer ST–TT transitions and larger saccade amplitudes (Figure 2a and 2b), indicating more efficient processing patterns when translators were primed with positive emotions. These relationships were not observed in the NE condition, lending support to H2. However, stimulus type did not significantly interact with flow scores to affect regression count (β = 6.42, SE = 12.72, t = .50, p = .616), but flow alone showed a marginally significant effect in predicting regression count (β = −11.51, SE = 6.17, t = −1.87, p = .067). This result indicates that the higher the flow score, the lower the regression count, which is unaffected by the primed emotional conditions.

Analysis testing H3 revealed a significant interaction between stimulus type and flow scores on MFD (β = −15.20, SE = 7.02, t = −2.17, p = .035). As illustrated in Figure 2c, the PE group showed a negative relationship between flow scores and MFD, whereas the NE group demonstrated a positive relationship. This suggests that in the PE condition, higher flow scores corresponded with reduced task effort. Conversely, in the NE condition, increased flow corresponded with greater task effort. These findings support H3.

Analysis testing H4 revealed a marginally significant interaction between stimulus type and flow scores on ST–TT fixation duration ratio (β = .11, SE = .06, z = 1.79, p = .074), but no significant interaction on PAU count percent (β = −.02, SE = .05, z = −.45, p = .656). As shown in Figure 2d, PE condition participants demonstrated a positive relationship between flow scores and ST–TT fixation duration ratio, suggesting that higher flow scores led to more balanced attention allocation between source and target texts. Conversely, NE condition participants showed increasingly imbalanced attention patterns with higher flow scores, favoring TT over ST processing.

4.3Interplay between flow experiences and task sequence in the translation process

Independent of stimulus type, the LMER model with task sequence as the predictor revealed significantly higher flow scores in Task 2 compared to Task 1 (β = 1.21, SE = .45, t = 2.68, p = .012). Testing the broaden-and-build effect of flow-related positive experiences, we examined interactions between flow scores and task sequence on behavioral and cognitive measures. Among all measures, only those related to attention allocation and parallel processing patterns showed potential interaction trends: ST–TT fixation duration ratio (β = .08, SE = .05, z = 1.67, p = .095) and PAU count percent (β = .04, SE = .03, z = 1.45, p = .148). As illustrated in Figure 3a and 3b, higher flow scores in Task 1 were associated with a more imbalanced ST–TT attention allocation pattern swinging to a TT dominance (the ratio reducing from around .50 to .35) and fewer parallel processing activities. These reversed in Task 2, where higher flow scores corresponded with a more balanced ST–TT attention allocation pattern and more parallel processing. However, these trends did not reach statistical significance and should therefore be interpreted with caution.

5.1Positive emotions are associated with specific flow dimensions

H1 investigates the antecedent effect of primed positive emotions on flow experiences during translation. We examined both total flow scores and specific flow dimensions. Although the PE group generally reported higher flow scores than the NE group, no significant difference emerged in overall flow levels. This may be because different cognitive activities exhibit different flow traits; not all dimensions of flow are necessarily equally sensitive to positive emotions in translation. Existing research shows that the dimensions of challenge-skill balance, sense of control, and loss of self-consciousness are more salient in translation-related flow (Weng and Li 2025Weng, Yu, and Yanze Li 2025 汉英翻译中的译者心流体验：积极心理机制与认知特征探析 [Flow experience in Chinese–English translation: Positive psychological mechanisms and cognitive characteristics]. Foreign Language Teaching and Research 57 (2): 290–303.Weng, Yu, and Yanze Li 2025 汉英翻译中的译者心流体验：积极心理机制与认知特征探析 [Flow experience in Chinese–English translation: Positive psychological mechanisms and cognitive characteristics]. Foreign Language Teaching and Research 57 (2): 290–303.). Accordingly, positive emotions may selectively enhance certain flow dimensions in translation, without significantly altering the overall flow score.

Action-awareness merging and autotelic experience were the dimensions most consistently associated with positive emotions, showing significant positive correlations across both tasks. Translators experiencing more positive emotions were more likely to feel merged with the translation activity and to find the experience inherently rewarding. The correlation between positive emotions and action-awareness merging aligns with Carver’s (2003Carver, Charles S. 2003 “Pleasure as a Sign You Can Attend to Something Else: Placing Positive Feelings within a General Model of Affect.” Cognition and Emotion 17 (2): 241–261. Carver, Charles S. 2003 “Pleasure as a Sign You Can Attend to Something Else: Placing Positive Feelings within a General Model of Affect.” Cognition and Emotion 17 (2): 241–261. , 249) “coasting tendency,” where positive feelings signal that things are going well, prompting individuals to withdraw excessive resource investment and maintain a natural, balanced state. In our study, translators in the PE group may have felt more confident and less tense, which, in turn, reduced their additional processing and monitoring, allowing smoother and more spontaneous engagement with the translation tasks. The link between positive emotions and autotelic experience can largely be attributed to the role of positive emotions in enhancing intrinsic motivation, as suggested by the self-determination theory (Deci and Ryan 2012Deci, Edward L., and Richard M. Ryan 2012 “Self-Determination Theory.” In Handbook of Theories of Social Psychology, edited by Paul van Lange, Arie Kruglanski, and E. Tory Higgins, 416–437. London: SAGE. Deci, Edward L., and Richard M. Ryan 2012 “Self-Determination Theory.” In Handbook of Theories of Social Psychology, edited by Paul van Lange, Arie Kruglanski, and E. Tory Higgins, 416–437. London: SAGE. ). Primed with positive emotions, translators tended to regard the translation activity itself as the reward.

The shift in correlations of initial positive emotions, from a distorted sense of time in Task 1 to action-awareness merging in Task 2, indicates a temporal progression of emotional effects. Initially, primed positive emotions primarily served to enhance interest, enjoyment, and motivation, helping translators adapt to the task and ‘get into a groove’. As task familiarity increased, positive emotions appeared to support a ‘fuel-saving’ and spontaneous working mode (Carver 2003Carver, Charles S. 2003 “Pleasure as a Sign You Can Attend to Something Else: Placing Positive Feelings within a General Model of Affect.” Cognition and Emotion 17 (2): 241–261. Carver, Charles S. 2003 “Pleasure as a Sign You Can Attend to Something Else: Placing Positive Feelings within a General Model of Affect.” Cognition and Emotion 17 (2): 241–261. ), where deliberate cognitive intervention was reduced and the merging of action and awareness became more prominent. The observed negative correlation between positive emotions and concentration in Task 1 further supports this idea, as it reveals that positive feelings may initially compete with focused attention. Given that higher positive emotions resulted in more relaxed and secure feelings, translators would not extensively mobilize attentional resources to the task. Taken together, these H1-related patterns refine our understanding of how positive emotions relate to flow in translation by showing that they selectively foster specific flow dimensions rather than uniformly boosting the overall flow state.

5.2The role of primed positive emotions and flow in enhancing processing efficiency

H2 is concerned with whether the interaction between primed positive emotions and flow makes the translation process more efficient, as indicated by eye-tracking measures of transition frequency between ST and TT, saccade amplitude, and regression count. Our findings provide partial support: under heightened flow, translators in the PE group exhibited significantly fewer ST–TT transitions and showed a marginal trend toward larger saccade amplitudes. However, the interaction was not significant for regression count.

These results suggest that when positive emotions co-occurred with flow, translators were more likely to process larger information chunks within each saccade and were able to proceed with the task with less frequent re-examinations of ST during TT production. This pattern supports the theoretical foundation of H2: broadened attentional scope promoted by positive emotions combines with the focused state of flow to facilitate the comprehension of larger information volume and promote more seamless code-switching. As Dietrich (2004)Dietrich, Arne 2004 “Neurocognitive Mechanisms Underlying the Experience of Flow.” Consciousness and Cognition 13 (4): 746–761. Dietrich, Arne 2004 “Neurocognitive Mechanisms Underlying the Experience of Flow.” Consciousness and Cognition 13 (4): 746–761. proposes, flow may lead to reliance on highly practiced skills encoded in humans’ implicit knowledge base, which has the advantage of being more efficient. Our findings extend this to translation, showing that positive emotions have the potential to amplify this advantage, leading to fewer transitions between ST and TT and larger saccade amplitudes.

Regarding regressions, the data show only a marginal main effect of flow, independent of emotional priming. The trend indicates that as translators’ flow experiences intensified, they tended to have fewer regressions during the translation process. This may be the case because regressions often reflect a strategic or corrective effort to ensure accuracy and coherence, which are integral to maintaining translation quality. Unlike ST–TT transitions and saccade amplitude, regressions are more likely governed by task-specific demands. In translation, especially for complex texts, such corrective efforts may persist regardless of positive emotional states. Hence, while positive emotions enhanced cognitive flexibility and broader processing, deliberate re-examination behaviors remained unaffected, as they may primarily stem from the challenge-skill alignment of task execution.

5.3The role of primed positive emotions and flow in reducing cognitive effort

H3 examines the influence of the interaction between primed positive emotions and flow experiences on cognitive effort indexed by MFD. The result supports our hypothesis: translators in the PE group showed a decreased MFD as flow scores increased, suggesting reduced cognitive effort. In contrast, without the catalyzing of positive emotions, increased flow scores in the NE group led to more deliberate processing and monitoring, which demanded greater cognitive effort.

This finding provides evidence for the heuristic processing style induced by positive emotions in translation tasks, where translators rely on preexisting cognitive shortcuts, thereby resulting in less effortful processing (Lyubomirsky, King, and Diener 2005Lyubomirsky, Sonja, Laura King, and Ed Diener 2005 “The Benefits of Frequent Positive Affect: Does Happiness Lead to Success?” Psychological Bulletin 131 (6): 803–855. Lyubomirsky, Sonja, Laura King, and Ed Diener 2005 “The Benefits of Frequent Positive Affect: Does Happiness Lead to Success?” Psychological Bulletin 131 (6): 803–855. ). The mutually beneficial relationship between positive emotions and the action-awareness merging dimension appears to contribute to this result. Specifically, when translators experience action-awareness merging, their behaviors become more spontaneous and less consciously monitored (Csikszentmihalyi 1990Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.), which does not consume short-term memory capacity and operates independently of the individual’s control (Schneider and Shiffrin 1977Schneider, Walter, and Richard M. Shiffrin 1977 “Controlled and Automatic Human Information Processing: I. Detection, Search, and Attention.” Psychological Review 84 (1): 1–66. Schneider, Walter, and Richard M. Shiffrin 1977 “Controlled and Automatic Human Information Processing: I. Detection, Search, and Attention.” Psychological Review 84 (1): 1–66. ). In the translation process, this may manifest as more automatic comprehension and production phases, allowing translators to maintain task performance while reducing mental workload. Such effects are particularly essential for professional practice, as intense effort investment over long periods may lead to fatigue and decreased efficiency.

5.4The role of primed positive emotions and flow in balancing attention allocation between ST and TT

H4 investigates whether the interaction between primed positive emotions and flow leads to a more balanced attention allocation pattern between ST and TT, alongside increased parallel processing. Our results partially support the hypothesis.

In the PE group, as flow scores grew, the ST–TT fixation duration ratio rose from slightly over .3 to nearly .5, suggesting a more balanced attention allocation pattern between ST and TT. This shift indicates that positive emotions, coupled with flow, tended to weaken the default TT-dominant focus in translation, encouraging translators to invest more attention in ST comprehension. A similar tendency was observed by Rojo and Meseguer (2021)Rojo López, Ana María, and Purificación Meseguer Cutillas 2021 “The Effect of Attitude Towards Catalonia’s Independence on Response Latency When Translating Ideologically Conflicting Press Headlines.” Onomázein (NEVIII): 128–145.Rojo López, Ana María, and Purificación Meseguer Cutillas 2021 “The Effect of Attitude Towards Catalonia’s Independence on Response Latency When Translating Ideologically Conflicting Press Headlines.” Onomázein (NEVIII): 128–145., who found longer reaction times when reading ideologically congruent press headlines during the reading-to-translate phase, and they linked this result to positive affect. In contrast, the NE group exhibited a reinforced TT-dominant pattern, likely reflecting habitual task strategies under the NE condition. This divergence aligns with the enhanced cognitive flexibility and broadened attentional scope associated with positive emotions (e.g., Lyubomirsky, King, and Diener 2005Lyubomirsky, Sonja, Laura King, and Ed Diener 2005 “The Benefits of Frequent Positive Affect: Does Happiness Lead to Success?” Psychological Bulletin 131 (6): 803–855. Lyubomirsky, Sonja, Laura King, and Ed Diener 2005 “The Benefits of Frequent Positive Affect: Does Happiness Lead to Success?” Psychological Bulletin 131 (6): 803–855. ; Beukeboom and Semin 2006Beukeboom, Camiel J., and Gün R. Semin 2006 “How Mood Turns on Language.” Journal of Experimental Social Psychology 42 (5): 553–566. Beukeboom, Camiel J., and Gün R. Semin 2006 “How Mood Turns on Language.” Journal of Experimental Social Psychology 42 (5): 553–566. ). Such cognitive benefits may loosen rigid processing patterns by facilitating a more flexible distribution of attentional resources, while in the NE group, translators’ cognitive resources remained more narrowly directed toward task completion rather than exploration or evaluation of the source text. This attentional flexibility may mitigate the risk of premature TT production, supporting better ST comprehension and reducing errors linked to superficial processing.

However, the interaction between primed positive emotions and flow did not significantly influence the percentage of parallel processing activities. This result may be explained by the working memory constraint for multitasking and participants’ limited translation experience. Parallel processing, which involves simultaneously managing comprehension and production, places greater demands on translators’ working memory (Hvelplund 2020Hvelplund, Kristian Tangsgaard 2020 “Translation, Attention and Cognition.” In The Routledge Handbook of Translation and Cognition, edited by Fabio Alves and Arnt Lykke Jakobsen, 279–293. London: Routledge. Hvelplund, Kristian Tangsgaard 2020 “Translation, Attention and Cognition.” In The Routledge Handbook of Translation and Cognition, edited by Fabio Alves and Arnt Lykke Jakobsen, 279–293. London: Routledge. ). Without adequate training and practice, student translators’ working memory was likely exhausted by a single task, whether comprehension, production, or code-switching, leaving no capacity for parallel processing. This finding echoes Hvelplund’s (2011b)Hvelplund, Kristian Tangsgaard 2011b “Distribution of Attention between Source Text and Target Text during Translation.” In Cognitive Explorations of Translation, edited by Sharon O’Brien, 215–236. London: Continuum.Hvelplund, Kristian Tangsgaard 2011b “Distribution of Attention between Source Text and Target Text during Translation.” In Cognitive Explorations of Translation, edited by Sharon O’Brien, 215–236. London: Continuum. observation that parallel attention occurs less frequently than attention directed at ST and TT, and its duration remains relatively stable. Hvelplund (2020)Hvelplund, Kristian Tangsgaard 2020 “Translation, Attention and Cognition.” In The Routledge Handbook of Translation and Cognition, edited by Fabio Alves and Arnt Lykke Jakobsen, 279–293. London: Routledge. Hvelplund, Kristian Tangsgaard 2020 “Translation, Attention and Cognition.” In The Routledge Handbook of Translation and Cognition, edited by Fabio Alves and Arnt Lykke Jakobsen, 279–293. London: Routledge. also found that parallel processing is a function of experience, as the split attention is likely enabled by typing or reading automaticity, which relies on established habit patterns and schemata and is acquired through repetition and practice. In addition, as a key component of translation automaticity, parallel processing is particularly sensitive to time-related constraints (Weng and Zheng 2024Weng, Yu, and Binghan Zheng 2024 “Automaticity in Translation: Effects of Time Pressure and Work Experience.” Meta 69 (2): 355–379. Weng, Yu, and Binghan Zheng 2024 “Automaticity in Translation: Effects of Time Pressure and Work Experience.” Meta 69 (2): 355–379. ). Since participants in the present study translated under a deadline-free condition, they were less compelled to use parallel processing as a strategic adaptation. Therefore, although the interplay between positive emotions and flow demonstrates diverse cognitive benefits, its influence may not be sufficiently powerful to enhance parallel processing. This finding both qualifies broad claims about the influence of positive emotions on attentional resources in translation and delineates a boundary, indicating that more demanding forms of multitasking remain constrained by working memory, experience, and time conditions.

5.5The broaden-and-build potential on cognitive resource allocation and parallel processing

H5 assumes that the behavioral and cognitive changes observed in H2 to H4 may reflect a broaden-and-build effect over time, whereby flow-induced positive experiences lead to cumulative improvements in processing efficiency, cognitive effort investment, attention allocation, and parallel processing during the second translation task. Our data yields interesting results. The behavioral measures, namely ST–TT transition, mean saccade amplitude, regression count, and MFD, showed no significant interactions between flow and task sequence, suggesting stable behavioral patterns across tasks. On the contrary, measures related to attention allocation and parallel processing (i.e., ST–TT fixation duration ratio and PAU count percent), exhibited a marginal interaction trend consistent with the broaden-and-build hypothesis. In Task 1, higher flow scores corresponded to a more imbalanced, TT-dominant attentional pattern and less frequent parallel processing. In Task 2, this trend reversed: increased flow scores were associated with more balanced attention allocation and more frequent parallel processing activities.

These findings indicate that the dominant manifestations of flow experiences evolve with task execution, affecting how translators allocate their attention between comprehension and production. In Task 1, flow appeared in its early-stage form, characterized by challenge-skill balance, clear goals, heightened concentration, and a strong sense of control (Csikszentmihalyi 1990Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.Csikszentmihalyi, Mihaly 1990 Flow: The Psychology of Optimal Experience. New York: HarperCollins.; Weng and Li 2025Weng, Yu, and Yanze Li 2025 汉英翻译中的译者心流体验：积极心理机制与认知特征探析 [Flow experience in Chinese–English translation: Positive psychological mechanisms and cognitive characteristics]. Foreign Language Teaching and Research 57 (2): 290–303.Weng, Yu, and Yanze Li 2025 汉英翻译中的译者心流体验：积极心理机制与认知特征探析 [Flow experience in Chinese–English translation: Positive psychological mechanisms and cognitive characteristics]. Foreign Language Teaching and Research 57 (2): 290–303.). Under this condition, translators with higher flow scores tended to direct attentional resources primarily toward TT production to ensure task completion. Subsequently, after receiving positive feedback from the successful completion of Task 1, translators’ flow experiences accumulated and intensified in Task 2, transitioning toward a more advanced state marked by action-awareness merging and a loss of self-consciousness. Such superior flow experiences allowed for the broadening of attentional breadth, enabling translators to adopt a more comprehensive cognitive strategy in which attention was more evenly distributed between ST and TT. This shift reflects the broaden-and-build effect within the translation process, where cumulative positive feelings facilitate the transition of attentional patterns from a focused, output-oriented mode toward a more integrated and flexible one.

The potential interaction trend in parallel processing might reflect a short-term training effect rather than the direct influence of cumulative flow-induced positive experiences. As discussed above, parallel processing is largely a function of experience, relying on the development of procedural knowledge and automaticity (Hvelplund 2020Hvelplund, Kristian Tangsgaard 2020 “Translation, Attention and Cognition.” In The Routledge Handbook of Translation and Cognition, edited by Fabio Alves and Arnt Lykke Jakobsen, 279–293. London: Routledge. Hvelplund, Kristian Tangsgaard 2020 “Translation, Attention and Cognition.” In The Routledge Handbook of Translation and Cognition, edited by Fabio Alves and Arnt Lykke Jakobsen, 279–293. London: Routledge. ). In the present study, repeatedly performing tasks likely increased participants’ familiarity with task requirements and the physical setup, reducing cognitive load related to task adaptation and enhancing typing automaticity. Additionally, completing the first task may have enabled translators to develop certain schemata that could be activated when encountering similar translation problems in Task 2. Such cognitive shortcuts allowed them to allocate fewer cognitive efforts to individual sub-tasks, thereby freeing up resources to support multitasking. These practice-based improvements align with cognitive models of skill acquisition, where repeated exposure fosters the development of automatic responses (Tenison, Fincham, and Anderson 2016Tenison, Caitlin, Jon M. Fincham, and John R. Anderson 2016 “Phases of Learning: How Skill Acquisition Impacts Cognitive Processing.” Cognitive Psychology 87: 1–28. Tenison, Caitlin, Jon M. Fincham, and John R. Anderson 2016 “Phases of Learning: How Skill Acquisition Impacts Cognitive Processing.” Cognitive Psychology 87: 1–28. ). Nevertheless, these trends of the broaden-and-build effect were only marginal, likely due to the limited sample size. As such, the interpretations warrant confirmation by future research.

Finally, the lack of significant interaction between flow and task sequence on other behavioral measures may reflect diminishing cognitive gains of flow in translators’ behaviors. Specifically, flow experiences enhanced processing efficiency and reduced cognitive effort at their emergence, leading to relatively optimized behavioral patterns even in the early experimental stage. Once this near-optimal state was reached, further improvements became less likely. These diminished benefits may result from the inherent limit of human working memory, which constrains the number of information chunks that individuals can retrieve, store, and manipulate at a time (Hvelplund 2020Hvelplund, Kristian Tangsgaard 2020 “Translation, Attention and Cognition.” In The Routledge Handbook of Translation and Cognition, edited by Fabio Alves and Arnt Lykke Jakobsen, 279–293. London: Routledge. Hvelplund, Kristian Tangsgaard 2020 “Translation, Attention and Cognition.” In The Routledge Handbook of Translation and Cognition, edited by Fabio Alves and Arnt Lykke Jakobsen, 279–293. London: Routledge. ). As a result, regardless of the cognitive benefits brought by flow, translators must continually acquire new information once the preceding information has been processed in order to advance the task. This necessity forces them to perform a certain number of fixations, saccades, regressions, and ST–TT transitions, making these measures unlikely to be further optimized by cumulative flow-related benefits. Overall, these findings highlight the complex interplay between positive emotions and flow in shaping translators’ cognitive and behavioral dynamics, revealing both tangible process-level gains and potential bottlenecks, and offering a translation-specific interpretation of the broaden-and-build dynamics across extended tasks.

5.6Implications and limitations

The findings presented in this article have practical implications for translation education and professional practice. As far as classroom settings are concerned, they point to the value of designing translation activities that deliberately support both positive emotions and flow by, for example, integrating brief positive emotion-inducing activities before demanding translation tasks, providing supportive feedback, and aligning text difficulty and task goals with students’ perceived skills to foster challenge-skill balance. As far as professional and training contexts are concerned, the results suggest that workflow design might benefit from sequencing tasks in ways that allow translators to build on prior positive experiences rather than sustaining prolonged high-effort states.

Viewed against the background of ongoing developments in machine translation and large language models, the findings further highlight that affective and cognitive factors are essential to how human translators operate within technology-rich, human-in-the-loop workflows. Even in AI-supported settings, translators remain responsible for scrutinizing and revising machine-generated translations, making context-sensitive decisions, and managing their attention across multiple sources of information. This emphasizes the continuing reliance on human expertise in professional translation. Given this context, the present findings suggest that fostering positive emotions and flow may help support this role by enabling more efficient, less effortful processing and more balanced attention allocation.

However, several limitations should be acknowledged. This study is constrained by the modest sample of student translators and the use of laboratory-based priming procedures. In addition, several effects related to attention allocation and task sequence were only marginally significant (possibly due to the modest sample size) and should be interpreted with caution. Future research may examine the generality of this affective-flow interaction in larger samples, across translator cohorts varying in expertise and language pairs, as well as in authentic translation contexts. Establishing dynamic cognitive-behavioral models of translation that integrate eye-tracking and behavioral data may also lay the groundwork for further workflow optimization and the development of process-monitoring tools.