Different fiction genres take children’s memories to different places

ABSTRACT Fiction, more than expository text, nurtures intimate connections between text and the reader’s life experiences. This dimension of reader response is underexplored in relation to children. Adapting methods from Empirical Literary Studies to educational research objectives, the authors employed the concept of ‘remindings’, i.e. reminiscing prompted by text, in studying children’s life-resonant responses to self-selected leisure books. Six workshops were run in primary classrooms during which participants (N = 148; age 8–11) engaged in remindings and mental imagery. Written remindings were then analysed for systematic variation across fiction book genres (Real-world vs. Fantasy; Relationships vs. Adventure). They found that Real-World Adventure books prompted remindings of discrete life events, while Real-World Relationships books prompted remindings of more diffuse experiences. Fantasy Adventure books were the least likely to prompt remindings. Further genre-based differences emerged in the distribution of themes within remindings. The authors consider the consequences of these insights for supporting young readers.


Introduction
There is an uncontested role for direct world experience in children's expository reading (Afflerbach, Pearson, & Paris, 2008;Kispal, 2008). For example, comprehending a written description of the water cycle requires at least some previous experience of condensation and evaporation, e.g. first-hand knowledge of morning dew, steaming kettles or clothes drying on the line. That is why, in classrooms, practical experiences commonly accompany the use of informational texts conveying abstract concepts.
But in what ways do personal experiences matter when a text is fiction and deals with life more generally, rather than with a clearly defined learning topic? Beyond literal comprehension, what is the potential of invoking such experiences when fostering fiction reading for pleasure? And will the possible manifestations and uses of invoking such experiences in the classroom vary depending on the type of fiction being read -and taught? These are the research questions informing the current empirical study.
In expository reading, lifelong learning progress is marked by advancing deductive abilities which entail abstracting away from discrete life experiences. Evidence regarding fiction, on the contrary, suggests that reading becomes more 'self-implicating' (Kuiken, Miall, & Sikora, 2004) with practice, age and the increasing variety of fiction that one reads (Charlton et al., 2004). Therefore, learning to read literature for pleasure not only entails learning to draw on one's world knowledge in an automated fashion, but also learning to read in a life-resonant manner. This involves engaging in the metacognitive exercise of noticing links between text and life.
It is true that for any given story, the chances of experiencing what researchers have termed 'personal resonance' (Seilman & Larsen, 1989) or 'personal relevance' (Kuzmičová & Bálint, 2019) will vary according to the unique background of each reader. Primary school teachers, however, can play a crucial role in shaping children's general propensity to identify and productively employ personal connections in written narratives, and by extension their sustained reading for pleasure and positive attitudes to literature that may last for life (Cremin, Mottram, Powell, Collins, & Safford, 2014;Kucirkova, Littleton, & Cremin, 2017).
In many educational systems, the reading as proficiency agenda and a focus on quantifiable outcomes predominate (e.g. Hempel-Jorgensen, Cremin, Harris, & Chamberlain, 2018). This reduces the space available for exploratory, life-resonant reading practices (Cremin et al., 2014). Policy documents sometimes acknowledge that learners should be 'encouraged to link what they read or hear read to their own experiences' in order to deepen their 'pleasure in reading, motivation, vocabulary and understanding' (Thornton, 2018, p. 212), yet these ideas tend to have little influence on practice. Teachers' grassroots initiatives (e.g. Open University, 2020) therefore seek interdisciplinary support to help them foster what Rosenblatt (1968Rosenblatt ( /1938 called 'the aesthetic stance', i.e. learners' deeply personal engagement with text (revisited in Kucirkova & Cremin, 2020). Designed in response to these needs, the current study draws on conceptual frameworks originally formulated by researchers outside education, in the field of Empirical Literary Studies, yet which are transferrable and relevant to pedagogical research and practice.
Specifically, we ran six exploratory workshops in primary classrooms during which children were invited to: a) engage in verbally guided 'remindings' (personal reminiscing) and mental imaging; b) take note of the respective magnitude of these two processes when prompted by an audio excerpt from a children's narrative; and c) describe, in writing, their remindings and mental imagery associated with a familiar leisure book of their own choosing. The study design and findings regarding children's written remindings are examined in this article.
Remindings were first studied by the psychologists Seilman and Larsen (1989) in the early years of the now burgeoning ELS initiative. Like many subsequent and more widely known ELS inquiries (e.g. Kidd & Castano, 2019), Seilman and Larsen addressed the idea that literary text is read differently to expository text. Specifically, they were interested in how these broad text types relate to readers' empirical self-knowledge. To this end they developed the method of 'self-probed retrospection' and asked university students to read either a short story by an award-winning author, or a chapter from a non-fiction book. While reading, students made pencil markings every time they thought of something that they had previously experienced or learned/heard about. Afterwards, they categorised the nature of these remindings. Compared to the expository text, the literary text elicited significantly more remindings where the participant was positioned as an Actor in an autobiographical event, rather than a passive Observer of an event or Receiver of information.
Later research moved beyond this design, with participants being asked to describe their remindings freely in writing (e.g. Halász, 1991;Therman, 2008). In a series of related studies, Halász (1991) asked secondary students to either read a short story or an expository text while completing various tasks, including written descriptions of remindings. Halász found that the largest part of remindings (approximately 50%) fell thematically into Seilman and Larsen's (1989) Receiver category, in that they referred to knowledge and experience gained through text and other media. The literary text predominantly reminded participants of other fictions (literature, film, theatre), whereas the expository texts reminded them of other non-fiction materials. However, he recognises this may have been skewed by his choice of literary text (The Vulture, a short story by Franz Kafka). As Halász contends, the pronounced fantastic features of this story are likely to override any sense of verisimilitude. This acknowledgement of differences within the literary/fictional text type, and of the potentially special status of fantastic literature, is relevant to the current study.
The evidence given here is not to suggest that remindings should be the primary focus of all literature-based learning. In secondary and tertiary education, where students are also expected to develop more abstract understanding of literary style and symbolic signification, ELS researchers have found self-referencing to be at the expense of such understanding at times (Sikora, Kuiken, & Miall, 2011;Therman, 2008). Even in these higher educational tiers, however, fiction reading ideally serves personal growth and pleasure alongside subject-specific knowledge of literature. Schrijvers, Janssen, Fialho, and Rijlaarsdam's (2019) review of pedagogical interventions in secondary settings shows compelling evidence that, if students' 'insight into human nature' is to be fostered in the classroom, they should 'participate in writing activities that focus on activating, annotating, and reflecting on personal experiences in relation to fictional texts and themes' (Schrijvers et al., 2019, p. 36). This entails verbal accounts of remindings triggered by the themes rendered in fiction.

Research design
To explore the basic usability of the concept of remindings in primary literacy learning, we designed a themed 30-minute intervention. Revolving around the two life-resonant experience categories of remindings and mental imagery, this 'Stories in Our Bodies' workshop was framed as dealing with some of the hidden magic of books.
We began with a Warmup Exercise in which the children were asked to close their eyes and conjure mental images in multiple sensory modalities, all related to viewing, manipulating and tasting an imaginary lemon. The procedure was adapted from Johnson, Cushman, Borden, and McCune's (2013) ELS study, which found that such multisensory imaging, if undertaken prior to reading, deepens one's engagement with text, yielding positive effects for transportation into the story world and empathy with characters. In our design this exercise aimed at introducing an introspective mood while also demonstrating that, despite the expression's visual connotations, 'mental images' can occur in any sensory modality including movement, touch or taste (Kuzmičová, 2014).
The Warmup Exercise was followed by an interactive Presentation on mental imagery and remindings. It was explained that in addition to generic mental images (the taste or texture of a lemon), words and stories sometimes make us remember more specific things or occasions (our regular rounds in the supermarket, or the time we made lemon cake mixture in Design & Technology), and that these so-called remindings can but do not have to have mental images attached to them. Throughout the Presentation, children were invited to provide examples from their own experience. It was emphasised that some readers notice more mental imagery and/or remindings than others and that everyone's images and/or remindings are unique. Examples of stories mentioned in the Presentation were varied between classes to ensure familiarity to all children.
After this followed a Listening and Tally Marks activity. During this activity, the children listened to two consecutive excerpts (1.5 minutes each) from the audio track of Michael Rosen's (2011) narrative poem 'Chocolate Cake'. While listening to the first excerpt, the children made tally marks on a Post-it note pasted onto an activity sheet every time they noticed a reminding. While listening to the second excerpt, they made tally marks on a different colour Post-it note every time they noticed a mental image. Although we anticipated that our participants typically preferred prose when reading for pleasure, we opted for a narrative poem because its density of sensory content and fast pace allowed for the use of a short excerpt without significant loss in meaning. The poem tells the tale of a boy who, in the dead of night, inadvertently eats up the family's cake. Albeit featuring poetic devices, it is written in free verse and, when heard without seeing how it is typeset, could easily be perceived as a short prose story.
Finally, the children were given two Writing Tasks on the reverse page of their activity sheet ( Figure 1). They were asked to respond to the following two prompts about their self-chosen text: 'This book reminded me of . . . ' and 'This book made me imagine (see, feel, taste, smell . . .).' In view of the innate difficulties of verbalising mental imagery, the imagery prompt was placed second, to be tackled once the children had already completed some writing. The prompts were accompanied by pictograms to remind the children of the distinction between remindings and images. For this activity, the children had been invited to bring their favourite book, which lay on the table in front of them. They could flip through the book to refresh their memories of it.
It was made clear that reporting no remindings or mental images, either in the Listening and Tally Marks activity or in the writing tasks, was perfectly legitimate. Thus, we took care to ease any pressure to perform according to a norm, inviting the children instead to provide as authentic a response as possible. That said, we were aware that the Listening and Tally Marks instruction especially would have caused the children's metacognitive focus to depart -to some extent -from what it is normally like for them to experience a story, similarly to most instructions used in ELS and other experimental research.
The workshops took place as part of Book Week and a themed day, respectively. Following the ethics procedures of the first author's institution, parents and children were informed of the activities beforehand in writing and were offered the possibility to opt out. The children did not write their names to ensure they remained fully anonymous, and they were expressly allowed to keep their activity sheet should they not wish to share their writing in the end.

Sample
The study took place in the West of England. A total of 148 children participated, across six classes divided between two schools: Ashtree (87 participants) and Lanesfield (61 participants; both pseudonyms). The classes ranged in size between 16 and 31 children. Ashtree School was in an academic upper middle-class suburb of a cosmopolitan city, with a wide array of options for organised out-of-school activity (Scouting, clubs and so on) within the catchment area. The school was consistently rated 'Outstanding' by the UK Office for Standards in Education (Ofsted). Only 5% of the children in this school received Pupil Premium funding for socially disadvantaged pupils in the year in which the research was undertaken. Lanesfield School was in an economically struggling working-class town 30 miles from the city that offered little in the way of organised out-ofschool activities. It served a community known for severe unemployment rates and was consistently rated by Ofsted as 'Requires Improvement'. In Lanesfield, the proportion of children receiving Pupil Premium funding was 40%. However, the schools were similar in that they served an overwhelmingly white British population which, with few exceptions, spoke English as their first and only language. The participants from Ashtree were aged 8-9 years, while those from Lanesfield were aged 10-11 years. The two-year age gap was a deliberate design choice based on preliminary conversations with teaching staff in the schools. They perceived that if we worked with younger children in Lanesfield, the data would likely be skewed by their significantly lower baseline reading skills, and simpler choices of reading material, compared to Ashtree. Meanwhile, older children in high-performing Ashtree could selfcensor their responses in favour of reading strategies modelled on the final primary assessment tests for which they had been intensively prepared from the age of nine. Our sample was thus selected to yield rich, varied and comparable data through a balance of homogeneity (linguistic background, relative reading skills) and diversity (socioeconomic and academic environment, out-of-school activities).

General conduct of the workshops; remindings vs. mental imagery
The children seemed to enjoy the workshops, responding spontaneously to interactive prompts, and some commented positively on the tasks. The Listening and Tally Marks exercise appeared to play an important part in this. It was clear that in all six participating classes, the workshop was perceived as distinct from the norm of a literacy lesson.
A general trend in the Writing Tasks was that accounts of remindings were slightly longer on average than accounts of mental images. The average word count was 11.78 (remindings) vs. 10.67 (mental images) words per participant. Importantly, accounts of remindings were also more linguistically complex; they more often included at least one finite verb clause with an explicit grammatical subject (e.g. 'It makes me remember when I trapped my finger in my car' [94L]), rather than nominal phrases only (e.g. 'going to this school/picking blackberries' [8A]). Eighty per cent of all remindings responses included at least one such clause, but this was only true of about 63% of all mental imagery responses. This finding contrasted with the Listening and Tally Marks activity where mental images received, on average, 1.6 times as many tally marks as remindings (see also Figure 1). This inverse relationship between remindings and mental images in interaction with the two activities can be interpreted as follows. On average, sensory images are more easily produced at will and can be a simpler mental form. They required less time to emerge during the Listening and Tally Marks activity but escaped verbalisation to a greater extent than remindings in the Writing Tasks. In addition, some children responded to the mental imagery question by continuing the narrative of their response to the immediately preceding remindings question, making the two conceptually indiscernible. The mental imagery responses are therefore difficult to interpret as an aggregate data set and can only be assessed in the context of corresponding remindings items. For these reasons, our remaining analyses in this article focus on the remindings items alone (i.e. any writing provided under the prompt 'This book reminded me of . . .').

Open coding of written responses: basic categories
Initial open coding of the children's remindings data (silently corrected for spelling) yielded the following categories of response: Event Reminding (E); Free Reminding (F); Alternative (A); Negative (N); Blank (B); and Unassigned (U) ( Table 1).
For subsequent analyses, we collapsed categories A, N and B under a common category of Non-remindings (X), ending up with four basic response categories. Seventeen per cent of all participants then provided responses in the X category, while a combined 64% of all participants clearly reported Proper Remindings, i.e. remindings in either the E or F category. Nineteen per cent of all responses remained unassigned to either category (U) due to brevity.  (Table 2). Across the 13 main themes, the full set of 94 Proper Remindings items was tagged 211 times, i.e. 2.24 times per item on average. Family (28%), Unique place (28%) and Adversity (23%) were the three most prominent themes and the only ones exceeding 20% of the set.

Controlling for between-school differences
Given the overall differences in socioeconomic environment, including vastly different possibilities for out-of-school activity, we used statistical tests to control for betweenschool differences in response patterns. Our objective was not to draw general conclusions concerning the relationship between discrete socioeconomic conditions and susceptibility to remindings. This would have required more nuanced data on individual participants and school pedagogy. We were however obliged to rule out the possibility that the findings reported here mask two latently divergent groups for whom the remindings task played out differently. The tests used for between-school comparisons as well as for all subsequent analyses were Fisher's exact test and, where the size of data sets permitted, Pearson's Chi-squared test. Both tests calculate the likelihood with which varied results emerge by chance rather than reflecting pre-existing significant differences between two or more sets. This likelihood is expressed in p-value, where p < .05 conventionally stands for statistical significance. Pearson's Chi-squared test of independence was run to compare the schools across the four basic response categories. The relationship between school and basic response category was not significant: X 2 (3, N= 148) = .43, p= .93. This means that children in Lanesfield were not more or less likely than children in Ashtree to return responses in either of the four basic categories: Event Remindings, Free Remindings, Non-remindings and Unassigned. Fisher's exact test was then run for each of the 13 main themes represented in the Proper Remindings data set (N= 94, including five Non-fiction items). The only theme that proved to differ significantly between the schools was Travel (p= .01). This theme was more strongly represented in Ashtree.
Connected to this, upon closer scrutiny, the theme of Unique place was predominantly represented by memories of everyday environments such as town parks and hospitals in Lanesfield, whereas Ashtree children frequently referred to overseas and exotic locations. There is however little in the data to counter the conclusion that both kinds of Unique place remindings enabled the children to personally relate to their favourite books. The overall absence of between-school differences suggests promise for educators who may wish to employ remindings in nurturing equitable reading for pleasure pedagogies.

Different genres of fiction
Our remaining analyses focus on narrative fiction, which 138 of 148 children selected as a favourite book. Within the broad category of narrative fiction, we proceeded to explore possible differences in remindings across fiction genres as warranted by previous research. It is frequently argued that, in contrast to expository text, fiction in general presents readers with the unique possibility to simulate full-fledged life situations without having to face real-world consequences (Djikic, Oatley, & Moldoveanu, 2013). However, a body of ELS research also suggests that certain genres of fiction may be more productive in this respect than others, namely fiction emphasising interpersonal relationships over action and adventure (Cupchik & László, 1994;Kidd & Castano, 2019). Based on this research, we hypothesised that categorising our participants' selected books as primarily focusing on Relationships vs. Adventure might yield distinct patterns in the quantity and/ or quality of reported remindings.
While both Relationships and Adventure books necessarily address some relationships and some emotions, experimental studies by Kidd and Castano (2019) have shown that the former category of fiction is associated with greater short-and long-term empathy development, presumably because it portrays more subtle and social emotion facets and makes readers monitor these emotions more minutely, in comparison to the more plotdriven Adventure genre. As the distinction between Relationships and Adventure reportedly concerns emotion and cognition on such deep structural levels, interesting differences in remindings could be expected to emerge beyond obvious themes such as Family or Conflict with others ( Table 2).
Conscious of the possible effect of perceived verisimilitude as referenced earlier (Halász, 1991), we further predicted systematic differences in remindings depending on whether the story of a selected book was largely Real-world vs. Fantasy based. In qualitative studies with young adult readers, Mackey (2019b) notes that fantasy fiction settings can be perceived by readers as triggering mental operations distinct from those of mentally furnishing real-world fictions. They make it less easy to 'import known setting in aid of comprehension', e.g. through instant images/memories of landmarks in one's own physical environment, requiring one instead to laboriously 'imagine everything' (Mackey, 2019b, p. 421). Whilst issues of unrealistic settings were perhaps especially likely to spill over into remindings themes such as Unique place and Travel (Table 2), fantasy worlds also feature fantastic creatures and natural and social laws diverging from reality. Again, the effects of this genre distinction on remindings could potentially be wide ranging.
Combined, the two distinctions yield four broad genre categories which we used in subsequent analyses: Real-world Adventure; Real-world Relationships; Fantasy Adventure; Fantasy Relationships. The same genre divisions are reportedly (Mackey, personal email to the first author, 8 September 2018) used by librarians in reader recommendation training and practice.

Books and genre categories represented
In our overall fiction sample made up of 138 responses, 62 children's fiction authors were represented, 22 were mentioned by at least two children and eight were mentioned in both schools. Between these 62 authors, 104 different books were mentioned, 15 authors being represented by more than one book and six authors being represented by more than three books. These six 'super-authors' were, in order of prominence: David Walliams (10 books, 18 mentions); Jeff Kinney (8 books, 13 mentions); Liz Pichon (7 books, 8 mentions); Jacqueline Wilson (6 books, 6 mentions); Roald Dahl (5 books, 6 mentions); and J. K. Rowling (4 books, 6 mentions). All of these were represented in both schools. Author variation was similar across the schools, with 1.9 children per author in Ashtree and two children per author in Lanesfield.
The most prominent genre category was Real-world Relationships (37 books/49 mentions), closely followed by Fantasy Adventure (34 books/42 mentions). Real-world Adventure (23 books/34 mentions) was about 1/3 less common than the two most prominent genres. Fantasy Relationships turned out to be the least common category, corresponding to less than 1/10 of the sample (10 books/13 mentions). The first author of this article and an independent children's literature expert each coded the full list of the selected books along both Relationships vs. Adventure and Real-world vs. Fantasy dimensions. Coding along the former dimension was more complicated, as some Adventure books had a strong relationships component, while the Relationships category was made up of diverse books whose main common feature was the absence of adventure. Despite these issues, inter-coder agreement reached 92% initially and all remaining issues were resolved upon further revisiting the contents.
It is also important to note that five of the children's authors were represented in more than one of the four genre categories. Three of these authors were among the superauthors mentioned earlier (Roald Dahl, David Walliams, Jacqueline Wilson); the remaining two were less prominent (Enid Blyton, Eva Ibbotson). Most of these authors crossed categories along one dimension only, e.g. Dahl, who was represented in Real-world Adventure and Fantasy Adventure, or Ibbotson, who was represented in Fantasy Relationships and Fantasy Adventure. The only exception was Walliams, the most prominent author overall, who crossed categories along both dimensions and was represented in Real-world Relationships, Real-world Adventure and Fantasy Adventure. Given this crossing of genres by individual authors, the trends reported for each genre in the following sections cannot be attributed to idiosyncratic writing styles, and suggest differences on the level of genre proper.

Genre categories vs. basic response categories
We used statistical tests to explore associations between response type and genre. However, the disproportionately small size of our Fantasy Relationships sample made it impossible to run comparisons of all four genre categories. Instead, we only compared the three larger categories of Real-world Relationships, Real-world Adventure and Fantasy Adventure (Step I, Tables 3-4), and additionally ran tentative tests for All Realworld vs. All Fantasy (Step II, Tables 3-4; includes Fantasy Relationships items) as well as All Relationships vs. All Adventure.
Pearson's Chi-squared test of independence for the three genre categories vs. four basic response categories proved significant, X 2 (6, N = 125) = 18.4, p = .005, suggesting that the basic response categories were not randomly distributed across genres. To further elucidate the exact nature of this finding, we ran a series of Fisher's exact tests Table 3. Four main response categories' frequency across the three major book categories -Realworld Relationships, Real-world Adventure and Fantasy Adventure (Step I) -and across the Real-world vs. Fantasy divide (Step II, includes Fantasy Relationships).
Step I (N= 125) Step II (N= 138) Notes: Bold digits indicate statistical significance; one asterisk indicates p < .05, two asterisks p < .01. Table 4. Thirteen main themes' frequency across the three major book categories -Real-world Relationships, Real-world Adventure and Fantasy Adventure (Step I) -and across the Real-world vs. Fantasy divide (Step II; includes Fantasy Relationships).
Step I (N = 82) Step II (N = 89) within each response category (individual rows in Table 3, Step I; significance level p < .05), with the following results: Event Remindings were significantly associated with Real-world Adventure books (p = .015). A separate comparison with Fantasy Adventure books also proved significant (p = .027), suggesting that the effect could indeed be uniquely attributed to Real-world Adventure books, rather than All Adventure books in general. This means that readers of Real-world Adventure were more likely than all others to recount their remindings in a fundamentally narrative form, i.e. as discrete life events. Free Remindings, on the other hand, were significantly associated (p = .022) with Real-world Relationships books. Readers of Real-world Relationships books were thus most likely to reminisce about their life experience without recounting specific life events.
Non-remindings, finally, proved significantly associated with Fantasy Adventure books (p = .005). They also proved inversely associated with Real-world Relationships books (p = .003). This means that children who wrote about Fantasy Adventure books were more likely than all others to return an answer that did not qualify as a reminding. This likelihood was in turn significantly smaller among readers of Real-world Relationships books as compared to all others. Pearson's Chi-squared test of independence likewise confirmed significant variation of responses between All Real-world and All Fantasy, X 2 (3, N = 138) = 10.22, p = .017. Subsequent Fisher's exact tests confirmed significant association between Non-remindings and All Fantasy (p = .003), specifically.

Genre categories vs. themes in remindings
Next, we explored possible systematic associations between book genres and the 13 main themes found in Proper Remindings. Fisher's exact tests were conducted within each theme (individual rows in Table 4, Step I; significance level p < .05), with the following results: Unique place was significantly associated with Real-world Adventure books (p = .002). However, a separate comparison with Fantasy Adventure books proved nonsignificant (p = .13), while a subsequent comparison of All Relationships with All Adventure proved significant (p = .009). This suggests that All Adventure books, not just the ones set in realistic worlds, may in fact be particularly prone to evoking specific locations.
Meanwhile, the themes of Adversity (p = .039), School (p = .002), Hobbies (p = .04) and marginally also Conflict with others (p = .052) were all significantly associated with Real-world Relationships books. In Step II, we additionally conducted Fisher's exact tests (significance level p < .05) for these themes across the All Relationships vs. All Adventure divide, all of which proved significant or marginally so. Finally, we found a significant association between the theme of Text/medium and Fantasy Adventure books (p < .001) which was also reflected under Step II for All Fantasy books in aggregate (p = .006).
None of the remaining seven themes varied significantly across the three larger genre categories or across the All Real-world vs. All Fantasy/All Relationships vs. All Adventure divides. This is perhaps particularly noteworthy in the case of Family, which was consistently high across all genre categories and, together with Unique place, also the strongest theme (29% of Proper Remindings prompted by fiction, 28% of all Proper Remindings). Our findings are further interpreted in the next section.

Discussion
We found important nuances in life-resonant reading between different genres of fiction. To summarise the main trends for each genre, we found that Real-world Adventure had the greatest power to invoke memories of circumscribed life events (Event remindings). All Adventure books, as compared to Relationships books, were in turn the ones most reliably taking the children's memories to discrete places, geographically speaking (Unique place). Both these findings make intuitive sense. Adventure books primarily build on a discernible plot, i.e. a clear succession of overt actions. These actions are apparent to a notional observer rather than 'merely' happening inside characters' minds and are perhaps more automatically grounded in concrete physical settings.
However, while all Adventure books were better suited for invoking memories of unique places compared to Relationships books, the significant association between Adventure and Event Remindings was overridden in the case of Fantasy Adventure books. This category was the one most likely to yield a Non-remindings response. Compared to Real-world Adventure, Fantasy Adventure books bear relatively little resemblance to the children's lives and are also generally more often read with the express objective of engaging in something different from one's own first-hand experiences. It would be interesting to test on a larger sample whether remindings in the Fantasy Adventure category, when prompted, trend towards a specific variety of the Unique place theme, e.g. more exotic locations.
Perhaps most remarkable is the strong association between Fantasy Adventure books and the Text/medium theme, i.e. the only theme that falls within Seilman and Larsen's (1989) Receiver (rather than Actor) category. This indicates that when Fantasy Adventure readers became reminded of anything, the memory often had to do with the story's artifice (e.g. the language it is written in, the book as an artefact) and/or its medial nature (i.e. other similar stories previously known from books, TV shows, computer games and so on). In other words, remindings prompted by Fantasy Adventure books were more often intertextual, involving connections between a text and other written materials or media products and directing the child's consciousness to story reception on a more abstract level (see also Torr, 2007).
Multiple interpretations converge to explain this finding. Clearly different from the world we live in, fantasy worlds tend to be more readily exploited by media and merchandise producers across platforms. For instance, Kinney's popular realistic series Diary of a Wimpy Kid has generated a minuscule industry compared to Rowling's Harry Potter. However, many of the intertextual references in the Text/medium set referred to stories similar but other than that presented by the book. For instance, our example of Text/medium in Table 2 comes from a reader of Funke's Inkheart, although the reminding concerns earlier experience of Gaiman's Coraline. This suggests a more general tendency in the Fantasy Adventure (or All Fantasy) sample to invoke mediated rather than first-hand experience, whether for lack of resemblance to one's lived reality, or for the saliency of being able to recognise a generic intertextual pattern, in line with Halász's (1991) study.
The last genre category that allowed relatively conclusive findings, Real-world Relationships, was the one least likely to be associated with a Non-remindings response, and the one most likely to generate remindings in the thematic categories of Adversity, School, Hobbies and marginally also Conflict with others. Again, these findings make sense; the themes of Adversity and Conflict with others reflect a recognition of complex social experiences known from everyday reality, while School and Hobbies, two key habitats outside home, emerge as the typical contexts in which these experiences take place. Many of these experiences were also linked to relationships within family. It is remarkable in this respect that Family was consistently represented in all categories of books and not significantly more so in Real-world Relationships.
Finally, we wish to acknowledge the limitations of our research design. Above all, follow-up research will also need to factor in the history of classroom talk about the books and genres represented, for example by combining workshop activities with observations and interviews (as reported in e.g. Arizpe & Styles, 2002). In addition, by opting exclusively to categorise the books along the 'Relationships vs. Adventure' and 'Real-world vs. Fantasy' dimensions, we deliberately refrained from coding for the extent of pictorial storytelling. There was a distinct share of graphic novels and other visual titles among the children's selected books, and their possible specificity or influence regarding remindings has yet to be explored.

Conclusion
This interdisciplinary study extends a tradition of fundamental Empirical Literary Studies research, but its key aim is to benefit educational practice. Whilst many argue that reading for pleasure should be fostered in ways that connect to children's personal experiences (Thornton, 2018), teachers receive little support in unravelling how this could be achieved (Cremin et al., 2014). An established way of exercising some form of personal connections, albeit typically used in basic reading comprehension rather than reading for pleasure, is mental imagery.
However, research suggests that even skilled readers vary greatly in their ability to conjure mental images while reading (Denis, 1982), and some do not consciously conjure mental images at all (Mackey, 2019a). Moreover, the respective remindings vs. mental imagery ratios found in our data confirm that mental imagery, however prominent, can be difficult to share through verbal means. Remindings present a useful complement to mental imagery work, prompting young readers to inspect their autobiographical memory (without which there would be little mental imagery), to help them connect and make sense of the text. Remindings may also enhance their desire to read on and even persevere with texts which they find challenging.
Drawing on observations of expository reading, researchers sometimes assume a straightforward linear relationship wherein 'the richer a child's world experiences and vicarious experiences' (Pressley, 2000, p. 549, cited in Kispal, 2008 14, our italics), the more elaborate their reading comprehension. The findings from our experientially divergent sample suggest that supported with a remindings instruction, fiction may partly obviate this logic. Unlike expository text, it invites an aesthetic (Rosenblatt, 1968(Rosenblatt, /1938, life-resonant reading stance and renders facets of human experience that are shared widely irrespective of one's background, e.g. everyday embodied sensations or the feelings involved in family and friendship dynamics. This remindings study also highlights the possible specific effects of certain broadly defined genres of fiction, which could be used to guide teachers' selections of texts for classes or recommendations to individuals. For example, the data indicate that Fantasy Adventure books may be the best choice when a teacher wishes to stimulate intertextual links (Torr, 2007) to other story experiences; that All Adventure books will invite children to note their responses to fictional environments; that Real-world Adventure books may be the most appropriate for inviting narrative sharing of circumscribed life events; and finally that Real-world Relationships books may be best suited for prompting thinking about the complex fabric of children's social and emotional experiences.
Whilst we recognise that the effects of texts are mediated by accompanying instructions and invitations to interact (Moses & Kelly, 2018) and that incongruent instructions can override a text's inherent potential to elicit life-resonant experiences (Cupchik & László, 1994), we consider that this study has pedagogical ramifications. Strategies deployed in this research could be used to explore children's 'funds of identity', defined as 'historically accumulated, culturally developed, and socially distributed resources that are essential for people's self-definition, self-expression, and self-understanding' (Esteban-Guitart & Moll, 2014, p. 37). We suggest that if teachers can develop their awareness of remindings and utilise them to enhance their knowledge of individual readers' funds of identity, this will, when combined with in-depth knowledge of children's texts, result in more individually tailored text recommendations and opportunities for children to find pleasure in genres that they may have previously rejected or not yet encountered.
Such professional knowledge is likely also to increase the congruence of a teacher's interactions with a text, which will enhance the potential to generate life-resonant experiences, provided that children's remindings are taken seriously and 'integrated [. . .] into the conversation in ways that [. . .] extend their understanding' (Torr, 2007, p. 81). It is in this capacity that we offer our findings, as the basis for developing suggestions of the kinds of remembering, imagining, thinking and talking to stimulate in young readers when working with fiction books across a range of given genres. Enhancing both teachers' and children's reflective awareness of personal remindings affords a potentially valuable and innovative way forward that can engage and support young readers.