Tracking contextual histories: how long-term and recent lexical experience shapes our processing of words
This thesis investigated the role of contextual information on lexical processing. Using secondary data analysis and online experiments, we tested how written language processing is shaped by recent experience and by the lexical histories words have accrued across the lifetime. Chapter 1 is a litera...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2023
|
| Subjects: |
| _version_ | 1797108972322291712 |
|---|---|
| author | Taylor, E |
| author2 | Kate, N |
| author_facet | Kate, N Taylor, E |
| author_sort | Taylor, E |
| collection | OXFORD |
| description | This thesis investigated the role of contextual information on lexical processing. Using secondary data analysis and online experiments, we tested how written language processing is shaped by recent experience and by the lexical histories words have accrued across the lifetime. Chapter 1 is a literature review. It reviewed the different ways of conceptualising semantic information, and discussed the importance of ‘semantic diversity’ (Jones et al., 2012; Hoffman et al., 2013), a metric which captures the variation in words’ contextual histories. Semantic diversity has been conceptualised and operationalised in different ways. Chapter 2 compared four different measures of semantic diversity by exploring their relationships with other item-level variables and their effects on lexical decision data. Small differences between the different measures demonstrated the importance of the corpus size and sample from which diversity is calculated. In Chapter 3, 3,000 English words were normed for context availability and sentence availability, two variables that reflect the ease with which words can be associated with a context or a sentence. These norms allowed for the exploration of lexical-semantic space on a larger scale. Context availability and sentence availability tapped different facets of semantic knowledge, and each explained unique variance in lexical decision performance. Context availability appeared to capture information relating to our interactive and imageable experience with words, while sentence availability captured linguistic variation in a similar way to semantic diversity. As sentence availability is a new metric and warrants further investigation, we employed this in Chapter 4 as a marker of long-term linguistic experience. Chapter 4 investigated the role of recent experience, by testing the effects of repeated exposure to words across sentence reading and sentence availability rating tasks. By presenting the same words in each task, we tested how longer-term lexical knowledge (indexed by sentence availability) interacts with the short-term influence of recent encounters (induced by priming). Words high in sentence availability were associated with more efficient performance in both tasks, therefore validating the measure. There were no main effects of priming but the interaction between priming and sentence availability was significant during sentence reading: words with high sentence availability were read faster during task 1 (effectively unprimed) compared to task 2 (primed). We ventured that this might be due to a difference between the context participants themselves produced in the rating task and the context they read in the subsequent task, but this result was difficult to interpret. Chapter 5 also tested long-term and recent experience using a more robust experimental design. Long-term experience of words was indexed by their semantic diversity and imageability values, which were manipulated orthogonally to compare high vs low imageability and high vs low semantic diversity. Participants first encountered words via either lexical decision or semantic decision (task 1; serving as unprimed performance), then all participants responded to the same words in a subsequent lexical decision task (task 2; serving as primed performance). All participants therefore encountered the same words, but half took part in the same task pre-test and post-test, while for the other half of participants, the task differed. Analysing the results from task 1, we found that semantic diversity was associated with more efficient lexical decision and less efficient semantic decision performance, while imageability was associated with more efficient performance in both tasks. These observations replicate those in the literature. Lexical decision performance was better in task 2 than 1, suggesting that repetition facilitated performance. Semantic diversity was no longer associated with lexical decision performance when it was administered as task 2, suggesting that recent experience (via task 1, regardless of whether that was lexical decision or semantic decision) might have temporarily masked the effects of a word’s long-term contextual history, as estimated by its semantic diversity.
Together, the results of this thesis demonstrate the importance of prior lexical experience on current linguistic behaviour. Through studying item-level variables, it shows the importance of both theoretical conceptualisations and computation methods in calculating semantic measures. It supports the idea of semantic richness as a complex and multifaceted construct (Pexman et al., 2002) and identifies potential distinct facets of meaning-based knowledge. It also gives a basis for understanding the relationship between long-term experience and recent word encounters through priming studies. The thesis closes with a general discussion that offers reflections and suggestions for future research. |
| first_indexed | 2024-03-07T07:35:27Z |
| format | Thesis |
| id | oxford-uuid:c77149dd-7daa-4011-8217-42c71ddb53c0 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T07:35:27Z |
| publishDate | 2023 |
| record_format | dspace |
| spelling | oxford-uuid:c77149dd-7daa-4011-8217-42c71ddb53c02023-03-02T11:43:35ZTracking contextual histories: how long-term and recent lexical experience shapes our processing of wordsThesishttp://purl.org/coar/resource_type/c_db06uuid:c77149dd-7daa-4011-8217-42c71ddb53c0PsychologyEnglishHyrax Deposit2023Taylor, EKate, NThis thesis investigated the role of contextual information on lexical processing. Using secondary data analysis and online experiments, we tested how written language processing is shaped by recent experience and by the lexical histories words have accrued across the lifetime. Chapter 1 is a literature review. It reviewed the different ways of conceptualising semantic information, and discussed the importance of ‘semantic diversity’ (Jones et al., 2012; Hoffman et al., 2013), a metric which captures the variation in words’ contextual histories. Semantic diversity has been conceptualised and operationalised in different ways. Chapter 2 compared four different measures of semantic diversity by exploring their relationships with other item-level variables and their effects on lexical decision data. Small differences between the different measures demonstrated the importance of the corpus size and sample from which diversity is calculated. In Chapter 3, 3,000 English words were normed for context availability and sentence availability, two variables that reflect the ease with which words can be associated with a context or a sentence. These norms allowed for the exploration of lexical-semantic space on a larger scale. Context availability and sentence availability tapped different facets of semantic knowledge, and each explained unique variance in lexical decision performance. Context availability appeared to capture information relating to our interactive and imageable experience with words, while sentence availability captured linguistic variation in a similar way to semantic diversity. As sentence availability is a new metric and warrants further investigation, we employed this in Chapter 4 as a marker of long-term linguistic experience. Chapter 4 investigated the role of recent experience, by testing the effects of repeated exposure to words across sentence reading and sentence availability rating tasks. By presenting the same words in each task, we tested how longer-term lexical knowledge (indexed by sentence availability) interacts with the short-term influence of recent encounters (induced by priming). Words high in sentence availability were associated with more efficient performance in both tasks, therefore validating the measure. There were no main effects of priming but the interaction between priming and sentence availability was significant during sentence reading: words with high sentence availability were read faster during task 1 (effectively unprimed) compared to task 2 (primed). We ventured that this might be due to a difference between the context participants themselves produced in the rating task and the context they read in the subsequent task, but this result was difficult to interpret. Chapter 5 also tested long-term and recent experience using a more robust experimental design. Long-term experience of words was indexed by their semantic diversity and imageability values, which were manipulated orthogonally to compare high vs low imageability and high vs low semantic diversity. Participants first encountered words via either lexical decision or semantic decision (task 1; serving as unprimed performance), then all participants responded to the same words in a subsequent lexical decision task (task 2; serving as primed performance). All participants therefore encountered the same words, but half took part in the same task pre-test and post-test, while for the other half of participants, the task differed. Analysing the results from task 1, we found that semantic diversity was associated with more efficient lexical decision and less efficient semantic decision performance, while imageability was associated with more efficient performance in both tasks. These observations replicate those in the literature. Lexical decision performance was better in task 2 than 1, suggesting that repetition facilitated performance. Semantic diversity was no longer associated with lexical decision performance when it was administered as task 2, suggesting that recent experience (via task 1, regardless of whether that was lexical decision or semantic decision) might have temporarily masked the effects of a word’s long-term contextual history, as estimated by its semantic diversity. Together, the results of this thesis demonstrate the importance of prior lexical experience on current linguistic behaviour. Through studying item-level variables, it shows the importance of both theoretical conceptualisations and computation methods in calculating semantic measures. It supports the idea of semantic richness as a complex and multifaceted construct (Pexman et al., 2002) and identifies potential distinct facets of meaning-based knowledge. It also gives a basis for understanding the relationship between long-term experience and recent word encounters through priming studies. The thesis closes with a general discussion that offers reflections and suggestions for future research. |
| spellingShingle | Psychology Taylor, E Tracking contextual histories: how long-term and recent lexical experience shapes our processing of words |
| title | Tracking contextual histories: how long-term and recent lexical experience shapes our processing of words |
| title_full | Tracking contextual histories: how long-term and recent lexical experience shapes our processing of words |
| title_fullStr | Tracking contextual histories: how long-term and recent lexical experience shapes our processing of words |
| title_full_unstemmed | Tracking contextual histories: how long-term and recent lexical experience shapes our processing of words |
| title_short | Tracking contextual histories: how long-term and recent lexical experience shapes our processing of words |
| title_sort | tracking contextual histories how long term and recent lexical experience shapes our processing of words |
| topic | Psychology |
| work_keys_str_mv | AT taylore trackingcontextualhistorieshowlongtermandrecentlexicalexperienceshapesourprocessingofwords |