2026-10-07T02:51:50Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34029318 2026-09-30T03:57:10Z category_37136 category_36167 portal_771 item_type_8 month_year_09_2026

Reading Minds: Priming Implicit Prosody and Individual Differences in Silent Sentence Processing Joy Mills (9448073) Psycholinguistics (incl. speech production and comprehension) Linguistic structures (incl. phonology, morphology and syntax) implicit prosody structural priming individual differences psycholinguistics prosody School: School of Linguistics and Applied Language Studies 470409 Linguistic structures (incl. phonology, morphology and syntax) 520405 Psycholinguistics (incl. speech production and comprehension) 280121 Expanding knowledge in psychology 280116 Expanding knowledge in language, communication and culture Degree Discipline: Linguistics Degree Name: Doctor of Philosophy Degree Level: Doctoral Our inner experience of silent reading affects reading enjoyment, comprehension, and broader measures of literacy. Therefore, how this inner experience differs between individuals is an important question. Yet this phenomenon is remarkably difficult to investigate. This thesis examines how implicit prosody affects sentence processing given a range of differences between individuals. Implicit prosody - the internally generated pitch, rhythm, pauses, and intonation experienced during silent reading - is challenging to study because both its production and perception occur covertly. However, this constraint also presents an opportunity: investigating implicit prosody can illuminate the mental representation of prosody more generally, as well as the ways individuals differ in their reading experience.To manipulate implicit prosody, the study employs a cross modal structural priming paradigm. Participants listened to auditory primes of ambiguous sentences which were disambiguated prosodically. They then silently read an ambiguous sentence of the same type and decided which continuation/interpretation was more likely. For example, after hearing primes like Olivia only mended the SOCKS (i.e. prosodically marked focus on the noun), when they then read the sentence She only ate the cake they should be more likely to choose the continuation she didn’t eat the pie (noun-association). However, after hearing a prime like Olivia only MENDED the socks (focus on the verb), they should be more likely to choose the continuation of the visual target sentence she didn’t make it (verb-association). That is, the structure in the auditory prime (syntactic and/or prosodic) should influence how participants read and then interpret the following visual sentence.Most structural priming research has focused on syntactic priming without considering implicit prosody. This thesis addresses how much of structural priming is prosodic rather than purely syntactic in two ways. First, it includes a sentence type (ambiguous only-association) in which interpretations differ in information structure, specifically focus. Sentences with ambiguous relative clause attachment, which vary both syntactically and prosodically, are also included. Second, the study uses both delexicalised auditory primes, which preserve only the prosodic contour, and unmanipulated primes containing full lexical and syntactic information. Presenting these in a within-subjects design allows assessment of whether individuals process these prime types differently.Recent work suggests that individual differences such as Autistic-like traits (AQ) and Working Memory (WM) correlate with variation in implicit prosody. This project seeks to replicate those findings and introduces two additional individual difference measures not previously examined in this context. The first, Spontaneous Synchronisation of Speech (SSS), tests whether participants automatically synchronise whispered rhythmic output with heard syllables; prior work shows a bimodal distribution of high vs. low synchronisers. Because SSS relates to both perception and production, it is relevant to implicit prosody. The second measure, the Bucknell Auditory Imagery Scale, assesses Auditory Imagery (AudIm)—the vividness and manipulability of imagined sounds. Although widely used in music cognition, it has not previously been applied to implicit prosody.The thesis comprises three analyses. The first analysis chapter is confirmatory, extending previous research. This chapter looks at whether structural priming will occur for ambiguous relative clauses and similarly for ambiguous only-association; whether prosodic contours alone (via delexicalised primes) will be sufficient to prime implicit prosody; and whether previously studied individual differences (AQ and WM) will have an effect on the priming of implicit prosody. The most surprising result was the lack of a simple effect of priming in the ambiguous relative clause condition, regardless of lexical information- effectively a failure to replicate a substantial body of structural priming literature. In contrast, robust priming effects emerged in the ambiguous only-association condition. However, no priming was observed in the delexicalised condition, suggesting that for information structural ambiguities, access to lexical information is necessary in addition to the prosodic contour. The confirmatory analysis also showed a significant interaction with Working Memory: higher WM increased the likelihood of unexpected verb association in the verb priming condition.The second analysis chapter is exploratory, examining the two new individual difference measures (SSS and AudIm) and their interactions with the priming of implicit prosody across the two sentence conditions (RC-attachment and only-association), the two lexical conditions (delexicalised and unmanipulated auditory primes) and AQ and WM. The exploratory analysis revealed an interaction between SSS and AudIm: as Auditory Imagery increased, the likelihood of unexpected verb association decreased in the noun condition, but only for high synchronisers. In other words, people who automatically synchronise to speech rhythms and who have vivid auditory imagery are strongly affected by the auditory primes. A three-way interaction between AudIm, Lexical Condition, and Prime Condition was included in the model, though post hoc tests did not show significant slope differences. Additional interactions provided insights into general sentence processing but did not involve Prime Condition and therefore cannot be interpreted as priming effects. The richness of these interactions suggests avenues for future research into individual differences in more general implicit prosodic processing.The third analysis uses eyetracking to determine whether priming, attachment/association decisions, and individual difference effects appear online during reading or exclusively offline while making a judgement. This analysis examines ambiguous only-association sentences and the same individual difference measures as the Confirmatory Analysis above. The results show priming effects in later measures (Total Dwell Time) which mirror effects found in the behavioural data, such as reduced reading time on the verb interest area after hearing a noun-disambiguated prime. Prime Condition interacts with Regressions In as well as a relatively early measure, First Run Dwell Time, suggesting that priming is impacting online reading earlier than anticipated. Models of the earliest measure (First Fixation Duration) do not include interactions with Prime Condition, but do include AQ and Lexical Condition, suggesting that during reading people with high and low AQ may be integrating lexical and prosodic information differently.There are a number of implications of this research. Firstly, a delexicalised prosodic contour does not appear to be enough to prime implicit prosody, although it may for participants with higher Auditory Imagery. This suggests that the abstract prosodic structure requires lexical or semantic information. Secondly, individual differences such as Auditory Imagery, Spontaneous Synchronisation to Speech, Autistic-like Qualities and Working Memory interact both in terms of priming and in general sentence processing. Further research on implicit prosody should take these differences into account, as should replications of earlier work in this area. 2026-09-30T00:00:00Z Text Thesis 10.26686/8912-9z8x https://figshare.com/articles/thesis/Reading_Minds_Priming_Implicit_Prosody_and_Individual_Differences_in_Silent_Sentence_Processing/34029318 CC BY-NC-SA 4.0