The study, conducted by Heng Li of Sichuan International Studies University in Chongqing, China, tested a deceptively simple theoretical proposal. High-power individuals, the argument goes, experience a heightened sense of autonomy, authority, and confidence. That psychological state reduces the pressure to align with others, because powerful people are less dependent on their interlocutors and less motivated to accommodate them. Low-power individuals, by contrast, live in a social world where dependence, vigilance, and the need for acceptance are more salient, which should make them more inclined to mirror the linguistic behavior of the people around them. If that reasoning is correct, then the strength of syntactic priming should vary systematically with power, and the difference should appear even when power is only temporarily induced in the laboratory.

Syntactic priming itself is one of the most robust findings in psycholinguistics. In the classic paradigm established by Kathryn Bock in the 1980s, participants who read or hear a sentence with a particular structure, such as a double-object dative, become more likely to describe an unrelated picture using that same structure. The effect persists across modalities, ages, and languages, and meta-analyses have confirmed its strength in language production. Researchers have proposed two broad mechanistic accounts. One treats priming as transient activation of syntactic representations in a spreading-activation production system. The other treats it as a form of implicit learning, in which the language system updates its expectations about which structures are likely to recur. Either way, the effect operates largely beneath conscious awareness, which makes it a sensitive probe for detecting how psycho-social states modulate linguistic processing.

What makes the new work notable is that most prior research on the social modulation of priming has focused on factors such as social opinion, ostracism, speaker accent, or communicative context. Social power, a foundational construct in social psychology defined as the capacity to influence others while being relatively resistant to influence oneself, had been largely unexplored in this domain. Yet power is known to alter a wide range of perceptual and cognitive processes. Previous studies have shown that high-power individuals process information more abstractly, take others’ perspectives less often, show reduced mirroring responses in the brain, and recognize emotional prosody in voices with lower accuracy than low-power individuals. Power also elevates self-concept consistency and authenticity, and it shapes expectations of social acceptance. If power dampens the tendency to imitate in these domains, the language system should be no exception.

To test this prediction, Li ran two studies using different priming methodologies and different participant populations, a design choice intended to guard against the possibility that any observed pattern is an artifact of a single task or a single group. Study 1 recruited Chinese student participants and used a dialogue-based writing task. Participants were first primed with either high power or low power, a standard social-priming manipulation in which people recall and describe an experience in which they had power over others or in which others had power over them. They then engaged in dialogue with interlocutors whose sentence structures were controlled by the experimenters. The key measure was whether participants subsequently adopted the syntactic structures their interlocutors had used. The results were clear: participants primed with high power were less likely to adopt the syntactic structures used by their interlocutors than participants primed with low power.

Study 2 corroborated the finding with a different task and a different population. This time the participants were community adults rather than students, and the priming methodology was a sentence scrambling task rather than a dialogue writing task. In sentence scrambling paradigms, participants rearrange words or phrases to form sentences, and the structures they produce reveal which syntactic representations have been activated by recent exposure. Once again, participants with high power exhibited a less pronounced syntactic priming effect than participants with low power. The convergence across two tasks, two participant groups, and two experimental formats strengthens the conclusion that the power effect is not tied to any particular methodological quirk but reflects a genuine modulation of the alignment process itself.

The theoretical significance of these results extends beyond the laboratory. Communication accommodation theory, developed by Howard Giles and colleagues, holds that speakers adjust their linguistic behavior toward or away from their interlocutors depending on their social goals, with convergence signaling affiliation and approval. The new findings suggest that power is a hidden moderator of this accommodation process. Powerful speakers, feeling less dependent on their partners, accommodate less, while powerless speakers, whose social outcomes hinge more on acceptance, accommodate more. This asymmetry could help explain a familiar feature of everyday conversation: in interactions between bosses and subordinates, teachers and students, or any pair with a power differential, the grammatical convergence may flow disproportionately from the less powerful side. Over time, such one-sided alignment could shape the linguistic norms of entire communities, since the structures favored by powerful speakers may spread more widely than those favored by powerless ones.

The findings also connect to a broader research program on imitation and social learning. Imitation, whether of gestures, facial expressions, or sentence structures, is a cornerstone of human sociality and a mechanism of cumulative cultural evolution. Previous work by the same author showed that social ostracism, the experience of being excluded, selectively increases language imitation, presumably because excluded individuals are motivated to regain acceptance through affiliation. The power results now complete the picture from the opposite direction: whereas exclusion amplifies the echo of others’ grammar, power mutes it. Together, these studies suggest that syntactic priming is not a fixed property of the language system but a flexible alignment behavior whose strength tracks the speaker’s social circumstances. The perception and production of language, the study concludes, vary significantly between individuals experiencing high and low power, highlighting the profound influence of social dynamics on linguistic processes.

Several caveats and open questions remain. The studies relied on temporarily induced power rather than stable individual differences in power motivation, so it is not yet known whether chronically powerful people, such as long-term leaders, show persistently weaker priming. The participants were Chinese speakers, and although syntactic priming is well documented across languages, including contrastive work on passive constructions in English and Chinese, cross-cultural replication would strengthen the general claim. The data supporting the findings are available from the corresponding author upon reasonable request, and the study received no external funding. Future research could examine whether the power effect interacts with expectation-based accounts of priming, in which alignment strength depends on the prediction error of the prime, or whether it extends to alignment at other linguistic levels, such as word choice and speech rate. It could also test whether the effect reverses in cooperative contexts, since prior work has shown that power and context jointly moderate verbal mimicry.

For now, the message is striking. The grammar we produce in conversation is quietly tuned by how powerful we feel. A person who feels in charge is less likely to borrow the sentence structures of the person they are talking to, while a person who feels subordinate is more likely to echo them. Language, in other words, is not just a vehicle for transmitting content; it is a social instrument whose alignment settings are calibrated by power. Every dialogue, from a hallway exchange to a boardroom negotiation, carries these linguistic echoes of hierarchy, and the new research shows that the echoes are loudest on the less powerful side.

Subject of Research: The effect of social power on syntactic priming and grammatical alignment in dialogue

Article Title: Linguistic echoes of power: high power leads to weaker syntactic priming effects

Article References: Li, H. (2026). Linguistic echoes of power: high power leads to weaker syntactic priming effects. Current Psychology, 45(20), Article 1626. https://doi.org/10.1007/s12144-026-10183-9

Image Credits: AI Generated

DOI: 10.1007/s12144-026-10183-9

Keywords: syntactic priming, social power, psycholinguistics, language production, communication accommodation theory, imitation, dialogue, sentence structure, social psychology, Current Psychology, Linguistic, echoes

Surprisal quantifies how unexpected a word is given the context that precedes it, as judged by a probabilistic model. In psycholinguistics, the concept has a long pedigree: work by John Hale in 2001 and Roger Levy in 2008 formalized the idea that the surprisal of a word predicts how hard it is to process, and subsequent studies linked word-level surprisal to reading times and to neural signatures of comprehension such as the N400 brain response. Words with high surprisal are less predictable and typically demand more cognitive effort; words with low surprisal glide by. Modern large language models, trained to predict the next word in a sequence, provide a convenient and powerful engine for estimating these probabilities at scale.

What makes the new work distinctive is its shift in perspective. Instead of using surprisal to probe how readers process a fixed text, the Spanish team asked whether surprisal could characterize the producer of a text. When someone generates language, the average surprisal of their words reflects how predictable—or unconventional—their discourse tends to be. That raises an obvious psychometric question: is mean surprisal a stable property of the person, or merely a quirk of one particular answer that will not generalize to anything else that person writes? If the former, surprisal joins the ranks of individual differences like vocabulary size or processing speed; if the latter, it is little more than a text-level statistic.

To find out, the researchers conducted a series of four studies using datasets from previously published work, reanalyzing language-based responses produced by the same individuals across multiple different items and tasks. Crucially, they treated the problem with the standard toolkit of psychological measurement: reliability analysis, inter-item correlations, and corrections for measurement error. If surprisal were merely a property of individual texts, scores from different writing prompts completed by the same person should be essentially unrelated. If it were a trait, those scores should correlate, just as different items on a well-constructed questionnaire do.

The results were striking. When surprisal scores were averaged across the different language-based items within a task, reliability was excellent—the measure behaved like a coherent psychological scale. Within individual items, reliability varied, but across items the scores converged. In other words, a person who produced highly predictable language on one prompt tended to produce highly predictable language on another, even when the content differed completely. The team also found, however, that surprisal was significantly influenced by the length of the text answer, a methodological caveat that future studies will need to control for when comparing texts of different sizes.

The researchers then examined how mean surprisal scores related to other variables, drawing on datasets that included measures of language skills, domain-general cognitive abilities, and personality traits. The pattern of correlations suggested that surprisal is not an isolated curiosity. It related to certain cognitive functions, consistent with the idea that producing predictable versus unpredictable language taps into underlying processing efficiency. The authors also propose that surprisal may capture alternative styles of speech or communication—broad tendencies in how people assemble their discourse that go beyond what any single answer reveals.

Technically, the measure is closely related to perplexity, a familiar quantity in computational linguistics; the authors note that surprisal scores can be seen as a linear transformation of perplexity but argue that surprisal is more interpretable from a psychological standpoint. The models used were left-to-right next-word predictors, meaning that each word’s surprisal was computed from the words preceding it, mirroring the incremental nature of human language production. Averaging word-level surprisal across a response yields a single predictability score per text, and averaging across texts yields the individual-difference measure at the heart of the study.

The implications reach well beyond psycholinguistics. Automated scoring of essays, the assessment of constructed responses, and the growing field of speech-based biomarkers for neurological conditions all depend on extracting meaningful, stable signals from language samples. If average surprisal tracks a durable characteristic of the speaker or writer, it could complement existing lexical and semantic features in applications ranging from educational assessment to the detection of cognitive decline, where changes in the structure and predictability of spontaneous speech have been linked to conditions such as Alzheimer’s disease and frontotemporal dementia. The new findings provide the psychometric foundation such applications require: evidence that the signal is reliable, not noise.

The study also carries a cautionary message for the rapidly expanding literature that uses large language models to score human language. Any single text-based score is contaminated by measurement error and by item-specific effects, and correlations between such scores and other variables will be attenuated accordingly. The authors show, for example, that correcting for attenuation can substantially raise observed inter-item correlations, meaning that raw correlations should be read as lower bounds. Transparency was a priority for the team: all surprisal measures, code, and supplementary materials are publicly available through the Open Science Framework, allowing other researchers to scrutinize and extend the approach. As artificial intelligence increasingly mediates how human language is measured, this work demonstrates that concepts born in information theory can be tamed into rigorous psychological instruments—provided researchers ask the same questions of them that psychometricians have asked of tests for a century: Is it reliable? Does it generalize? And what, exactly, does it measure about the person behind the words?

Subject of Research: Using large language model surprisal as a stable individual-difference measure of language production predictability

Article Title: Surprisal from Large Language Models as an individual difference: Capturing the individual regardless of the text answer

Article References: Martínez-Huertas, J. Á., Martínez-Mingo, A., López Rosell, A., Kreitchmann, R. S., & Jorge-Botana, G. (2026). Surprisal from Large Language Models as an individual difference: Capturing the individual regardless of the text answer. Behavior Research Methods, 58(10), Article 288. https://doi.org/10.3758/s13428-026-03117-7

Image Credits: AI Generated

DOI: 10.3758/s13428-026-03117-7

Keywords: surprisal, large language models, individual differences, psycholinguistics, language production, reliability, predictability, Behavior Research Methods, psychometrics, information theory, cognitive skills, automated text scoring