Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Dec 1.
Published in final edited form as: Child Dev Perspect. 2021 Aug 24;15(4):220–227. doi: 10.1111/cdep.12430

We Know More Than We Ever Learned: Processes Involved in the Accumulation of World Knowledge

Patricia J Bauer 1
PMCID: PMC8635395  NIHMSID: NIHMS1731148  PMID: 34868348

Abstract

Accumulating information and knowledge is a major task of development. A common assumption is that we build our storehouse of world knowledge, our semantic memory, through direct experience. Although direct experience is involved, to explain fully how we know all that we know, we also must consider processes that allow for integration of information learned in separate yet related episodes of direct learning, as well as inferential processes that operate over integrated representations and permit productive extension of knowledge. In this article, I describe the self-derivation through integration paradigm my colleagues and I developed to model these processes. Using this paradigm, we charted individual and developmental variability throughout childhood and in young adults. Several findings support the contention that the self-derivation through integration paradigm provides a valid model for how we build semantic knowledge, including the observations that performance on the task correlates with and predicts individuals’ world knowledge and academic success.

Keywords: learning, memory, conceptual development

“Knowledge itself is power” (Francis Bacon, 1597) and “Information is liberating” (Kofi Annan, 1997) are clichés, yet these sayings are also true. This explains why, worldwide, countries spend an average of $10,000 per student at the primary to postsecondary level and $16,300 at the tertiary level (Organisation for Economic Cooperation and Development [OECD], 2021) on formal education. They make these investments because information and knowledge afford not only power and liberation, but also higher educational attainment. Higher educational attainment is linked to higher socioeconomic status in adulthood (OECD, 2021), which in turn is linked to a variety of indicators of physical and mental well-being (e.g., Halleröd & Gustafsson, 2011).

Important as it is, formal educational attainment is only one facet of knowledge. Information and knowledge start to accumulate before formal education begins and continue to accrue well after it has ended. Their origins are not limited to academic subjects such as science and math, or even to informal learning environments such as museums, but extend to all manner of experiences. Moreover, accumulation of information and knowledge is important to societies as well as to individuals. For example, more informed individuals tend to make for a more interested and participatory electorate (e.g., Persson, 2015) and to think further into the future (Jarosz, 2019), which is linked to greater concern for the welfare of others and the environment (e.g., Arnocky et al., 2013).

Relations with individual and societal well-being make clear the importance of accumulating information and knowledge over development. Although research on how we learn information (i.e., facts and figures) has been conducted for a long time, we know less about the processes involved in the accumulation of information and how accumulated information supports knowledge (i.e., understanding of a subject). In this review, I discuss an assumption about the primary way we accumulate information, namely, through direct experience, and why it is inadequate for explaining how we know as much as we know. I then discuss an approach to this question that recognizes the need to integrate separate yet related learning episodes and the opportunity that integrated representations provide for self-derivation of new semantic content. I argue that appreciating the role of the processes of integration and self-derivation in learning will improve our understanding of how we develop world knowledge, bringing with it the potential for improved individual and societal function.

World Knowledge and How We Assume It Is Acquired

Like most processes that are the subject of psychological inquiry, in the formal literature, “world knowledge” has a nom de plume. Our storehouse of factual knowledge is known as semantic memory, one of two forms of consciously accessible long-term memory (Tulving, 1972). Semantic memory is that portion of long-term memory that stores facts, figures, names, and dates. The contents of semantic memory are thought to be timeless and placeless, meaning that the specific episodes in which this knowledge is learned are not retained. In contrast, episodic memory is our storehouse of memories of events located in specific time and place. The distinction is between knowing that the capital of the U.S. state of Georgia is Atlanta (a semantic fact) and remembering learning about state capitals in the fourth grade at Douglas MacArthur Elementary School in Indianapolis, Indiana (an episodic memory).

A common—though usually unstated—assumption is that direct experience is how facts about the world are registered in semantic memory. That is, information is gained through reading, instruction, and museums, to name a few examples. The notion that direct experience is a major (if not the major) source of semantic content is deeply rooted in theories of intelligence, such as the Cattell-Horn-Carroll model (e.g., Flanagan et al., 2000), which features Comprehension-Knowledge (Gc) as a major component of intelligence (historically, crystallized intelligence; Schneider & McGrew, 2012). Gc comprises information stored in long-term memory and reflects “the cumulative effects of exposure to and retention of diverse forms of culturally relevant information” (Schneider & McGrew, 2012, p. 122), including definitions of words and the concepts to which they refer. Productive processes operate over this semantic content, permitting inferences that promote comprehension of a text, for example (e.g., anaphoric reference). Yet the raw materials for such inferences come from direct experience (e.g., text, lecture, museum placard) and the products of the inferences are not themselves considered semantic content. On the face of it, the story of how we build our storehouse of world knowledge is a simple one—we accumulate information through explicit, direct experience.1 However, as I describe next, this account is incomplete.

Dismal Facts about Learning (and Retention)

We learn through direct experience. Yet the assumption that most of the contents of our seemingly boundless semantic memories are acquired in this way is undercut by the stark reality that we rapidly forget much of what we learn. This is not news—the finding dates to Ebbinghaus (2013/1885), who studied himself as he learned and relearned nonsense syllables, charting the shape of his forgetting curve along the way. As Ebbinghaus discovered, an initial steep decline in how much is remembered is followed by a long, slow decay (see Murre & Dros, 2015, for a contemporary replication). The pattern fits well with the power function, which implies that equal ratios of time result in equal ratios of recall (Rubin & Wenzel, 1996). The power function applies not only to memory for meaningless content such as nonsense syllables, but also to what we teach or are expected to learn in school, including word lists, Spanish and Swahili vocabulary terms, and biopsychology facts. Moreover, it applies to personally relevant content, such as the names and faces of high school classmates, and autobiographical (or personal) events and experiences. The same general pattern is observed whether testing is via recall or recognition (e.g., Rubin, 1982; Rubin & Wenzel, 1996; Wixted & Ebbesen, 1991).

One exception to this pattern of forgetting is of special significance to developmental science—in children, forgetting is even more pronounced than in adults. Children’s forgetting of autobiographical memories is best fit by the exponential function (Bauer & Larkina, 2014). Like the power function, the exponential function is characterized by rapid initial forgetting. However, exponential forgetting does not slow over time. Instead, for each unit of time, the amount remembered reduces by half (see Figure 1). Although exponential forgetting in children has been shown in the context of episodic (autobiographical) memory, given the similarity in forgetting curves for episodic and semantic memories (noted earlier), it is reasonable to expect that it would also extend to semantic content. The implications are obvious: Children experience rapid forgetting of much of what they learned, and they continue to forget, leaving little to reside in the permastore of semantic memory. We do not know when in development children’s forgetting slows and approximates that of adults. We do know that exponential forgetting provides a better fit to the data from children as old as 11 years.

Figure 1. Schematic Comparison of Rates of Reduction in Memories Over Time.

Figure 1

Open in a new tab

Rates of reduction in memories over time corresponding to the power function (grey bars) compared to the exponential function (black bars). Both functions are characterized by rapid initial forgetting (change from T1 to T2, though the change is larger in an exponential function). Whereas forgetting characterized by the power function slows over time, forgetting characterized by the exponential function continues to be pronounced.

The fact that we rapidly (and for children, continue to) forget material that we experience directly raises a significant question: How do we ever accumulate information and world knowledge? Learning science has identified activities that promote retention, including but not limited to interleaved practice (e.g., Taylor & Rohrer, 2010), pretesting (e.g., Arnold & McDermott, 2013), and retrieval practice (e.g., Ariel & Karpicke, 2018). The efficacy of these activities is apparent in classrooms as well as laboratories, and for children as well as adults (reviewed in Lucariello et al., 2016). Yet as I describe next, whereas these activities promote retention of explicitly taught information, they are not especially illuminating regarding how we build semantic knowledge.

The Challenge of Building Semantic Knowledge

Building sematic knowledge involves more than learning (and retaining) facts, figures, and other types of information. Indeed, appreciating the challenges inherent in the task requires that information be distinguished from knowledge. Whereas facts and figures constitute information, knowledge implies a theoretical or practical understanding or mastery of a subject. The difference is between the bits and pieces that are information and their organization into meaningful, interconnected wholes. Creating a meaningful whole from separate facts requires the cognitive process of integration; integration makes possible productive processes that operate over integrated memory representations, permitting construction of new knowledge. Combined focus on these processes—which represent an interaction of explicit learning and self-generation—may provide a more optimal model for how we develop world knowledge.

Consider that learning does not occur all in one place and time but is distributed over separate episodes. For example, accumulating information about dolphins might involve a visit to an aquarium, during which dolphins communicate with one another by clicking and squeaking. Later, an online article reveals that dolphins live in groups called pods. Still later, a trivia game or other test poses the question of how members of pods communicate. Although the fact called for was not explicitly learned, integrating the information from the separate episodes about dolphins allows formulation of the correct answer—which is by clicking and squeaking. In this way, integration promotes connections among facts and concepts, rendering them more resistant to forgetting (e.g., Ericsson & Kintsch, 1995). The newly formed representations also make possible the self-generative process of self-derivation, which permits going beyond information that is learned explicitly to construct new knowledge and understanding (e.g., Bauer, 2012).

The fact that information is acquired over distributed episodes means that for it to accumulate, separate yet related episodes of learning must be integrated with one another. Although this requirement is obvious, the study of the cognitive and neural processes involved in integration has a relatively recent history (e.g., Bauer, 2012; Preston & Eichenbaum, 2013). Similarly, although it is obvious that people make inferences based on learned information, the potential of the inferences as sources of new semantic knowledge has been largely unexplored (e.g., Bauer & San Souci, 2010). This is not to say that the power of inferential processes has been overlooked: Topics such as analogy, deduction, generalization, induction, and transitive inference have been the focus of substantial research attention (e.g., Dias & Harris 1988; Gentner, 1989; Goswami, 2011). Moreover, the general capacity for inferential reasoning is a basic component of fluid intelligence (Gf: abstract reasoning that is independent of content and prior knowledge; Schneider & McGrew, 2012). Yet the bulk of this work has focused on the process of forming novel inferences (the fluid ability itself) rather than on the resulting products (the crystalized content). That is, the attention has been on whether an inference is made, not on whether the product or result of the inference becomes incorporated into long-term memory as new semantic content. Indeed, disproportionate attention to fluid abilities has led to characterization of semantic content as “the wallflower of the intellectual trio” (which is made up of Gc, Gf, and Gv [visual-spatial reasoning ability]; Hunt, 2000, p. 124)—largely ignored in favor of flashier fluid and visual-spatial reasoning abilities.

A combined focus on integrating separate yet related episodes of explicit learning and the new semantic content derived through self-generative processes operating over integrated representations provides a way to understand not only how information accrues but also how knowledge accumulates over development. This new lens—elaborated in the next section—provides a seemingly necessary complement to the study of explicit learning, given the vulnerability of explicit learning to forgetting. It also complements the study of inferential processes by focusing on their resulting products as sources of new semantic content.

Integration and Self-Derivation of New Semantic Content

To model accumulation of knowledge, my colleagues and I developed a paradigm for children and adults; participants learn new factual information in distributed episodes, and then are asked questions that can be answered by integrating the related episodes and self-deriving novel facts. We have charted individual and developmental variability in integration and self-derivation that correlates with and predicts individuals’ world knowledge and academic success.

Paradigm and Performance

To study how to build a knowledge base, we used the self-derivation through integration paradigm (introduced in Bauer & San Souci, 2010). As schematically illustrated in Figure 2, the basic paradigm involves explicit learning of true but novel facts. Unbeknownst to participants, some of the facts form related pairs, such as the facts about dolphins mentioned earlier. The facts are presented through stories (for 4- to 8-year-olds) or individual sentences (for 8-year-olds through adults). One member of each pair of facts is presented in an initial learning period, followed by a buffer activity (these activities ensure that information must be retrieved from long-term, as opposed to working, memory). After the buffer, the second member of each pair of facts is presented. After another buffer, participants are asked open-ended questions followed by forced-choice questions that can be answered by integrating related pairs of facts and self-deriving new knowledge.

Figure 2.

Figure 2

Open in a new tab

Schematic Representation of the Self-Derivation Through Integration Paradigm

Children as young as 4 years perform well in forced-choice (50%) but not open-ended (13%) testing (Bauer & Larkina, 2017; see also Bauer & San Souci, 2010). By 6 and 8 years, children successfully self-derive new facts even in open-ended testing (50% and 75%, respectively). Yet even among adults, individual variability is striking, with undergraduate students answering between 3 and 93% of questions correctly (Varga & Bauer, 2017b). Control conditions make clear that integrating the separate episodes is necessary for high levels of performance. When one but not both members of fact pairs is presented, neither children nor adults generate the novel facts that answer the questions posed to them.

A Model of Accumulation of Knowledge

Several findings support the contention that self-derivation through integration provides a valid model for how temporally distributed information is integrated and productively extended. First, across studies, we used a wide range of facts, including for children, material derived from their science curricula (Bauer et al., 2020). The questions posed about the facts required different types of inferences to support self-derivation, including deduction, induction, and part-whole and transitive reasoning. Thus, the paradigm models fact learning and productive extension broadly.

Second, self-derivation through integration was apparent not only in the laboratory (e.g., Esposito & Bauer, 2018), but also in classrooms of children from diverse racial and ethnic, socioeconomic, and language backgrounds (Bauer et al., 2020; Esposito & Bauer, 2017, 2019). Thus, the paradigm models expanding semantic memory in representative populations.

Third, information that was newly self-derived through integration was treated as familiar, even though it was never taught explicitly. Specifically, event-related potentials (ERPs: scalp-recorded electrical activity time locked to stimulus presentation) revealed that after a single presentation, neural responses to familiar facts were similar to responses to facts derived through integration, both of which differed from novel facts (Bauer & Jackson, 2015). This suggests that the new information has been incorporated into the knowledge base (see also Varga & Bauer, 2017a).

Fourth, newly self-derived information was retained for at least one week by children as young as 4 years (Varga et al., 2016) and adults (Varga & Bauer 2017b). Levels of retention were high, reflecting little to no forgetting, thus defying both the power and exponential functions. This benefit likely accrues both because explicitly learned facts are connected to one another through integration, and because self-generated content tends to be retained more effectively than content learned through direct experience (i.e., generation effect; see Bertsch et al., 2007, for a meta-analysis).

Fifth, newly self-derived facts themselves were used productively to further expand knowledge. Both children and adults integrated self-derived facts with explicitly taught facts and used the resulting integrated representations to self-derive yet more new knowledge (Wilson & Bauer, 2021).

Finally, self-derivation through integration related to measures of verbal comprehension, with correlations in the moderate range: rs = .45 and .42 for 6- to 8-year-olds and 8- to 10-year-olds (Esposito & Bauer, 2018) and .53 for college students (Varga et al., 2019). It also related to academic achievement. Among children, relations were observed with reading comprehension and math scores (Esposito & Bauer, 2017); among college students, it related to SAT scores and grade point average (GPA) in the semester of testing, and predicted GPA two years later (Varga et al. 2019).

Together, these findings provide strong support for the contention that the experimental paradigm of self-derivation through integration provides an ecologically valid model for how we build semantic knowledge.

Self-Derivation Through Integration: What Develops?

Self-derivation clearly changes developmentally through integration over childhood (e.g., Bauer & Larkina, 2017). This developmental variability relates to each of several steps hypothesized to be involved in the process (Bauer & Varga, 2017). As illustrated in Figure 3, self-derivation through integration requires Encoding the first true, previously unknown fact—for example, dolphins talk by clicking and squeaking. While encoding the second, related fact, the learner must Reactivate the first fact, based on the common element. Once the separate but related facts are activated, they can be Integrated into a single representation. Because the facts refer to concepts about which learners have some knowledge (though the facts are novel), the integrated representation also may include prior knowledge, such as that dolphins have unique dorsal fins or swim while they sleep. To respond to a specific question, the learner then must Select relevant information. To respond to “How does a pod talk?” she must realize that pod relates to dolphins, and that whereas how dolphins talk is relevant, information about dorsal fins and sleeping is not. The learner then can Self-derive the novel fact that pods talk by clicking and squeaking.

Figure 3. ERISS Model.

Figure 3

Open in a new tab

The ERISS model depicts the processes of Encoding of the first episode of experience; Reactivating the first episode upon encoding of the second, related episode; Integrating the related episodes into a new representation; Selecting the subset of information relevant to a specific question or demand; and Self-deriving new semantic knowledge.

Adapted from Bauer & Varga, 2017

Childhood is characterized by developmental variability in executing the steps in self-derivation through integration. Four-year-olds encoded the novel facts taught them but they typically retained both members of the related pairs of facts only 30% of the time. Even when they learned the facts to criterion, they recalled both only 46% of the time (Bauer & San Souci, 2010). This suggests failed reactivation, integration, or both. Four-year-olds also were challenged to identify the information relevant to a test question (the step of selection). When they were given four potentially relevant facts (talk, group, fins, sleep), open-ended performance was at 0% and forced-choice performance did not differ from chance. In contrast, when they were given only the two necessary facts (talk, group), although open-ended performance remained low (13%), under the more supportive conditions of forced-choice testing, they selected correctly 50% of the time (Bauer & Larkina, 2017).

By age 8, children encoded and retained both novel facts 88% of the time, and even when challenged with multiple potentially relevant facts, they had high levels of open-ended self-derivation (Bauer & Larkina, 2017). Yet they seemed dependent on explicit prompts to reactivate prior information (Miller-Goldwater et al., 2021) and integrate it with new information (Bauer et al., 2021). In contrast, adults reactivated and integrated even before being prompted to do so, and high levels of spontaneous reactivation and integration were related to high levels of self-derivation (Bauer et al., 2020; Miller-Goldwater et al., 2021). Relative to children, adults also were more likely to self-derive new factual knowledge without being prompted (Wilson & Bauer, 2021). These developments likely are supported by several factors, including increases in semantic knowledge and its accessibility (e.g., Bauer et al., 2020), and development of the neural structures and networks that support memory integration and self-generative processes (Bauer et al., 2019).

Implications for Building World Knowledge

The developmental changes in integration and self-derivation observed throughout childhood and into adulthood have important implications for accumulating world knowledge. At a fundamental level, the observation that preschoolers engage in these processes under restricted conditions limits the extent to which they can exploit them to accumulate knowledge. Even in the school years, children’s apparent dependence on prompts to reactivate and integrate related facts with one another and use them to self-derive new factual knowledge represents a rate-limiting variable in accumulating knowledge, since prompts and cues are often absent in the world outside the laboratory. Adults’ more spontaneous engagement of these processes is an important developmental achievement (Bauer et al., 2020).

Although evocation of integration differs developmentally, the processes are available from a young age. This means that even young children have many ways to accumulate world knowledge. They can gain information via explicit learning and, with supports for encoding and at test, they expand their knowledge base by integrating across separate yet related learning episodes. When they do so, they retain their new knowledge over time. Children’s success at self-derivation through integration relates to semantic knowledge broadly, as measured by verbal comprehension (Esposito & Bauer, 2018), and to academic achievement specifically (Esposito & Bauer, 2017). These relations speak to the functional significance of self-derivation through integration as a gateway to a rich knowledge base (Bauer, 2012; Bauer et al., 2020).

Conclusion

Knowledge affords both power and liberation, as well as tangible benefits for individuals and society. With new paradigms for studying memory integration, we have expanded our understanding of how knowledge accumulates over separate yet related learning episodes. With paradigms for studying inferential processes that operate over integrated representations, we have expanded our understanding of how knowledge begets knowledge: Children and adults self-derive new, correct factual knowledge that they retain, incorporate into their knowledge base, and use productively to generate yet more knowledge. These processes of integration and self-derivation make information more resistant to forgetting by creating connected wholes that are more than the sum of their parts, producing more knowledge than we ever learned.

Acknowledgments

The research in this article was supported by grants from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (HD57291, HD067359, HD094716), the National Science Foundation (BCS1528091, 1748293), the Institute of Education Sciences (R305A160240), and Emory College of Arts and Sciences. The author thanks the Memory at Emory laboratory group for their collaborations and for comments on an earlier version of this article, as well as the participants in the research reported herein.

Footnotes

1

This generalization applies to our understanding of how we accumulate the factual content that constitutes knowledge of the world. Many models that focus on word representations—a component of semantic memory—explicitly acknowledge a variety of productive processes and thus consider mechanisms other than direct experience (e.g., Huebner & Willits, 2018). However, these models are focused on language and word meaning; their architects do not extend them to accumulation of knowledge of the world more broadly. Indeed, the state of theorizing about semantic memory more broadly has been characterized as impoverished, especially in contrast to theorizing about episodic memory (Duff et al., 2020).

References

  1. Annan K (1997). United Nations press release SG/SM/6268 23 June 1997. https://www.un.org/press/en/1997/19970623.sgsm6268.html
  2. Ariel R, & Karpicke JD (2018). Improving self-regulated learning with a retrieval practice intervention. Journal of Experimental Psychology: Applied, 24(1), 43–56. 10.1037/xap0000133 [DOI] [PubMed] [Google Scholar]
  3. Arnocky S, Milfont TL, & Nicol JR (2013). Time perspective and sustainable behavior: Evidence for the distinction between consideration of immediate and future consequences. Environment and Behavior, 46(5), 556–582. 10.1177/0013916512474987 [DOI] [Google Scholar]
  4. Arnold KM, & McDermott KB (2013). Test-potentiated learning: Distinguishing between direct and indirect effects of tests. Journal of Experimental Psychology: Learning, Memory, and Cognition, 39(3), 940–945. https://doi: 10.1037/a0029199 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bacon F (1597). Meditationes sacrae [Sacred meditations]. Open Library. [Google Scholar]
  6. Bauer PJ (2012). Facilitating learning and memory in infants and young children: Mechanisms and methods. In Odom SL, Pungello EP, & Gardner-Neblett N (Eds.), Infants, toddlers, and families in poverty: Research implications for early child care (pp. 49–74). Guilford Press. [Google Scholar]
  7. Bauer PJ, Cronin-Golomb LM, Porter BM, Jaganjac A, & Miller HE (2020). Integration of memory content in adults and children: Developmental differences in task conditions and functional consequences. Journal of Experimental Psychological: General. 10.1037/xge0000996 [DOI] [PubMed] [Google Scholar]
  8. Bauer PJ, Dugan JA, Varga NL, & Riggins T (2019). Relations between neural structures and children’s self-derivation of new knowledge through memory integration. Developmental Cognitive Neuroscience, 36(April), 100611. 10.1016/j.dcn.2018.12.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bauer PJ, Esposito AG, & Daly JJ (2020). Self-derivation through memory integration: A model for accumulation of semantic knowledge. Learning and Instruction, 66(April), 101271. 10.1016/j.learninstruc.2019.101271 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bauer PJ, & Jackson FL (2015). Semantic elaboration: ERPs reveal rapid transition from novel to known. Journal of Experimental Psychology: Learning, Memory, and Cognition, 41(1), 271–282. 10.1037/a0037405 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Bauer PJ, & Larkina M (2014). Childhood amnesia in the making: Different distributions of autobiographical memories in children and adults. Journal of Experimental Psychology: General, 143(2), 597–611. 10.1037/a0033307 [DOI] [PubMed] [Google Scholar]
  12. Bauer PJ, & Larkina M (2017). Realizing relevance: The influence of domain-specific information on generation of new knowledge through integration in 4- to 8-year-old children. Child Development, 88(1), 247–262. 10.1111/cdev.12584 [DOI] [PubMed] [Google Scholar]
  13. Bauer PJ, & San Souci P (2010). Going beyond the facts: Young children extend knowledge by integrating episodes. Journal of Experimental Psychology, 107(4), 452–465. 10.1016/j.jecp.2010.05.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Bauer PJ, & Varga NL (2017). Integration of separate episodes of experience: ERISS, a component processes model. Journal of Applied Research in Memory and Cognition, 6(2), 124–128. 10.1016/j.jarmac.2017.01.006 [DOI] [Google Scholar]
  15. Bertsch S, Pesta BJ, Wiscott R, & McDaniel MA (2007). The generation effect: A meta-analytic review. Memory & Cognition, 35, 201–210. 10.3758/BF03193441 [DOI] [PubMed] [Google Scholar]
  16. Dias MG, & Harris PL (1988). The effect of make‐believe play on deductive reasoning. British Journal of Developmental Psychology, 6(3), 207–221. doi: 10.1111/j.2044-835x.1988.tb01095.x [DOI] [Google Scholar]
  17. Duff MC, Covington NV, Hilverman C, & Cohen NJ (2020). Semantic memory and the hippocampus: Revisiting, reaffirming, and extending the reach of their critical relationship. Frontiers in Human Neuroscience, 13, Article 471. 10.3389/fnhum.2019.00471 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Ebbinghaus H (2013/1885). Memory: A contribution to experimental psychology. Annals of Neurosciences, 20(4), 155–156. 10.5214/ans.0972.7531.200408 (Original work published 1885) [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Ericsson KA, & Kintsch W (1995). Long-term working memory. Psychological Review, 102(2), 211–245. 10.1037/0033-295X.102.2.211 [DOI] [PubMed] [Google Scholar]
  20. Esposito AG, & Bauer PJ (2017). Going beyond the lesson: Self-generating new factual knowledge in the classroom. Journal of Experimental Child Psychology, 153(January), 110–125. 10.1016/j.jecp.2016.09.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Esposito AG, & Bauer PJ (2018). Building a knowledge base: Predicting self-derivation through integration in 6- to 10-year-olds. Journal of Experimental Child Psychology, 176(December), 55–72. 10.1016/j.jecp.2018.07.011 [DOI] [PubMed] [Google Scholar]
  22. Esposito AG, & Bauer PJ (2019). Self-derivation through memory integration under low surface-similarity conditions: The case of multiple languages. Journal of Experimental Child Psychology, 187(November), 104661. 10.1016/j.jecp.2019.07.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Flanagan DP, McGrew KS, & Ortiz SO (2000). The Wechsler intelligence scales and CHC theory: A contemporary approach to interpretation. Allyn & Bacon. [Google Scholar]
  24. Gentner D (1989). The mechanisms of analogical learning. In Vosniadou S & Ortony A (Eds.), Similarity and analogical reasoning (pp. 199–241). Cambridge University Press. https://doi: 10.1017/cbo9780511529863.011 [DOI] [Google Scholar]
  25. Goswami U (2011). Inductive and deductive reasoning. In Goswami U (Ed.), Childhood cognitive development (pp. 399–419). Wiley-Blackwell. 10.1002/9781444325485.ch15 [DOI] [Google Scholar]
  26. Halleröd B, & Gustafsson J (2011). A longitudinal analysis of the relationship between changes in socio-economic status and changes in health. Social Science and Medicine, 72(1), 116–123. 10.1016/j.socscimed.2010.09.036 [DOI] [PubMed] [Google Scholar]
  27. Huebner PA, & Willits JA (2018). Structured semantic knowledge can emerge automatically from predicting word sequences in child-directed speech. Frontiers in Psychology, 9. 10.3389/fpsyg.2018.00133 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Hunt E (2000). Let’s hear it for crystallized intelligence. Learning and Individual Differences, 12(1), 123–129. 10.1016/S1041-6080(00)00033-9 [DOI] [Google Scholar]
  29. Jarosz E (2019). Unequal times: Social structure, temporal perspective, and time allocation in Poland. Social Indicators Research, 141(February), 1189–1206. 10.1007/s11205-018-1866-7 [DOI] [Google Scholar]
  30. Lucariello JM, Nastasi BK, Anderman EM, Dwyer C, Ormiston H, & Skiba R (2016). Science supports education: The behavioral research base for psychology’s top 20 principles for enhancing teaching and learning. Mind, Brain, and Education, 10(1), 55–67. 10.1111/mbe.12099 [DOI] [Google Scholar]
  31. Miller-Goldwater HE, Cronin-Golomb LM, Porter BM, & Bauer PJ (2021). Developmental differences in reactivation underlying self-derivation of new knowledge through memory integration. Cognitive Psychology, 129(2021), 101413. [DOI] [PubMed] [Google Scholar]
  32. Murre JMJ, & Dros J (2015). Replication and analysis of Ebbinghaus’ forgetting curve. PLOS ONE. 10.1371/journal.pone.0120644 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Organisation for Economic Cooperation and Development. (2021). Education at a glance 2020: OECD indicators. OECD Publishing. 10.1787/69096873-en [DOI] [Google Scholar]
  34. Persson M (2015). Education and political participation. British Journal of Political Science, 45(3), 689–703. 10.1017/S0007123413000409 [DOI] [Google Scholar]
  35. Preston A, & Eichenbaum H (2013). Interplay of hippocampus and prefrontal cortex in memory. Current Biology, 23(17), R764–R773. 10.1016/j.cub.2013.05.041 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Rubin DC (1982). On the retention function for autobiographical memory. Journal of Verbal Learning and Verbal Behavior, 21(1), 21–38. 10.1016/S0022-5371(82)90423-6 [DOI] [Google Scholar]
  37. Rubin DC, & Wenzel AE (1996). One hundred years of forgetting: A quantitative description of retention. Psychological Review, 103(4), 734–760. 10.1037/0033-295X.103.4.734 [DOI] [Google Scholar]
  38. Schneider WJ, & McGrew KS (2012). The Cattell-Horn-Carroll model of intelligence. In Flanagan DP & Harrison PL (Eds.), Contemporary intellectual assessment: Theories, tests, and issues (p. 99–144). Guilford Press. [Google Scholar]
  39. Taylor K, & Rohrer D (2010). The effects of interleaved practice. Applied Cognitive Psychology, 24(6), 837–848. 10.1002/acp.1598 [DOI] [Google Scholar]
  40. Tulving E (1972). Episodic and semantic memory. In Tulving E & Donaldson W (Eds.), Organization of memory. Academic Press. 10.1007/978-1-4419-0463-8_37 [DOI] [Google Scholar]
  41. Varga NL, & Bauer PJ (2017a). Using ERPs to dissociate the neurocognitive processes underlying knowledge extension through memory integration. Journal of Cognitive Neuroscience, 29(11), 1932–1949. https://doi: 10.1162/jocn_a_01168 [DOI] [PubMed] [Google Scholar]
  42. Varga NL, & Bauer PJ (2017b). Young adults self-derive and retain new factual knowledge through memory integration. Memory & Cognition, 45, 1014–1027. 10.3758/s13421-017-0711-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Varga NL, Esposito AG, & Bauer PJ (2019). Cognitive correlates of memory integration across development: Explaining variability in an educationally relevant phenomenon. Journal of Experimental Psychology: General, 148(4), 739–762. 10.1037/xge0000581 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Varga NL, Stewart R, & Bauer PJ (2016). Integrating across episodes: Investigating the long-term accessibility of self-derived knowledge in 4-year-old children. Journal of Experimental Child Psychology, 145(May), 48–63. 10.1016/j.jecp.2015.11.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Wilson JT, & Bauer PJ (2021). Prompt-facilitated learning: The development of unprompted memory integration and subsequent self-derivation. Memory & Cognition. Advance online publication. https://doi: 10.3758/s13421-021-01155-4 [DOI] [PubMed] [Google Scholar]
  46. Wixted JT, & Ebbesen EB (1991). On the form of forgetting. Psychological Science, 2(6), 409–415. 10.1111/j.1467-9280.1991.tb00175.x [DOI] [Google Scholar]

RESOURCES