A brighter side to memory illusions: False memories prime children’s and adults’ insight-based problem solving

https://doi.org/10.1016/j.jecp.2010.08.012Get rights and content

Abstract

Can false memories have a positive consequence on human cognition? In two experiments, we investigated whether false memories could prime insight problem-solving tasks. Children and adults were asked to solve compound remote associate task (CRAT) problems, half of which had been primed by the presentation of Deese/Roediger–McDermott (DRM) lists whose critical lures were also the solutions to the problems. In Experiment 1, the results showed that regardless of age, when the critical lure was falsely recalled, CRAT problems were solved more often and significantly faster than problems that were not primed by a DRM list. When the critical lure was not falsely recalled, CRAT problem solution rates and times were no different from when there was no DRM priming. In Experiment 2, without an intervening recall test, children and adults still exhibited higher solution rates and faster solution times to CRAT problems that were primed than to those that were not primed. This latter result shows that priming occurred as a result of false memory generation at encoding and not at retrieval during the recall test. Together, these findings demonstrate that when false memories are generated at encoding, they can prime solutions to insight-based problems in both children and adults.

Introduction

It is well known that memory is error prone and that errors frequently lead to false memory illusions (i.e., illusions that take the form of a belief that something had actually been present when in fact it had not (Deese, 1959, Roediger and McDermott, 1995)). Such spontaneous errors of commission can be studied using the Deese/Roediger–McDermott (DRM) paradigm, where participants are given a word list (e.g., nap, doze, dream, pillow) whose members are all associates of an unpresented item or critical lure (e.g., sleep). Despite its never having been presented, participants often falsely remember the critical lure as being presented in the list. When studied developmentally, these spontaneous false memories increase with age (e.g., Howe, Wimmer, Gagnon, & Plumpton, 2009).

By now, we are all too familiar with the darker or negative side of false memory illusions, from experts being more prone to false memories in their domains of expertise (Castel, McCabe, Roediger, & Heitman, 2007), to miscarriages of justice (Loftus, 2003), to the outright memory wars of the 1990s (Crews, 1995). However, we argue that there is also a brighter, more positive side to false memories, one that is similar to that usually attributed to true memories. This positive aspect of false recollection is the role they can play in more complex cognitive processes such as insight-based problem solving.

To see this brighter side, consider the notion that false memories, like false beliefs (e.g., McKay & Dennett, 2009), may be the consequence of some creative process. For example, Castel and colleagues (2007) found that because experts have rich and highly interconnected memory networks in their area of expertise, they are more prone to memory errors related to that expertise. It may be, then, that generation of related information, including information not presented (i.e., false memories), is related to the discovery of creative solutions to problems (Sio & Ormerod, 2009), solutions that may depend on spreading activation through well-integrated associative networks that are said to serve as a foundation for human thought (Anderson, 1983, Reder et al., 2009).

One way to investigate this brighter side of false memories is by asking whether false memories can prime solutions to insight-based problems such as those found in compound remote associate task (CRAT) problems. Originally developed by Mednick (1962), these tasks involve the presentation of three words, such as apple, family, and house, all of which can be linked by a single word, in this case tree. To gain insight and solve this problem, theorists have suggested a process involving spreading activation, one that continues until the correct concept has been activated (Bowden, Jung-Beeman, Fleck, & Kounios, 2005). If we also assume that false memories are caused by a spreading activation mechanism (Howe et al., 2009, Roediger and McDermott, 1995), then priming becomes an ideal area of investigation (Anderson, 1983).

In true memory, priming has been interpreted in terms of an enhanced speed and tendency to complete tasks, such as stem completion tasks, when their completion involves the use of a word previously studied (e.g., Graf, Shimamura, & Squire, 1985). McDermott (1997) found that critical lures could also be used to prime word stem and fragment completion tasks, although priming occurred at a level lower than if the items had actually been studied (see Diliberto-Macaluso, 2005, for similar findings with child participants). Similarly, McKone and Murphy (2000) showed that critical lures could prime solutions to both implicit (stem completion) and explicit (stem-cued recall) memory problems. The question addressed here is whether false memories can also prime more complex cognitive tasks.

We suspect that they can because problems requiring a high level of insight may be aided by the spreading activation of concepts in memory, a process similar to the mechanisms proposed in spreading activation models of false memory effects (e.g., Howe et al., 2009, Roediger and McDermott, 1995) as well as Underwood’s (1965) original implicit associative response model. For example, Kershaw and Ohlsson (2004) discovered that insight problem solving involves searching through related concepts in memory for relevant information. Bowden and colleagues (2005) also suggested that insight-related problem solving involves the activation of concepts in memory, including those that are unrelated to the solution, followed later by the weak activation of concepts that are critical to the solution. Indeed, research has already shown that true memories can be used to prime problem solving and reasoning tasks successfully (e.g., Kokinov, 1990), so it might not be too far-fetched to anticipate that false memories may also prime problem solutions. In fact, some evidence has recently emerged showing that, at least for adults, false memories can and do prime solutions to CRAT problems. However, this priming occurred only when the critical lure was falsely remembered on a recall test and not simply due to the presentation of a DRM list whose critical lure was not falsely recalled (Howe, Garner, Dewhurst, & Ball, 2010).

In the current research, we wanted to replicate this finding with adults and, more important, to extend these priming effects to children. This question is important developmentally for any number of reasons. For example, recall that children are less susceptible than adults to spontaneous false memory illusions, especially those induced using the DRM paradigm (e.g., Howe et al., 2009). One reason for this may be because spreading activation is less automatic in children’s memory networks than in those of adults (see Howe, 2005, Howe et al., 2009, Wimmer and Howe, 2009, Wimmer and Howe, 2010). Interestingly, these differences are often attenuated (although not always eliminated) when age-appropriate materials (those that are congruent with children’s knowledge base) are used (e.g., Anastasi & Rhodes, 2008). To the extent that spreading activation is also important to solving CRAT problems, we were interested in whether limitations in automaticity might also constrain children’s problem-solving abilities even when age-appropriate problems were used.

We examined this question by using CRAT problems whose baseline solution rates were relatively high for both children and adults. That is, we used age-appropriate CRAT problems as determined by a norming study that we present next. Thus, although we predicted that the usual age increases in false recall might exist despite using age-appropriate lists, we were interested in whether we could attenuate (or eliminate) age differences in problem-solving rates by using age-appropriate problems. Indeed, age differences in problem difficulty were not, in and of themselves, of interest in this study; rather, we wondered whether false memories could serve the same priming function for children as they can for adults when problem difficulty was equated across age.

Section snippets

Pilot study: Norming CRAT problems for use with children

Before turning to the main experiments, we report a pilot study in which we collected norms for CRAT problems for use with children. Although the other experiments in this article involved only one child age group (11-year-olds) in contrast to adults, the norming study examined CRAT solution rates and times for 7-, 9-, and 11-year-olds.

Participants

A total of 60 individuals participated in the experiment: 30 11-year-olds (M = 11 years 2 months, SD = 4 months) and 30 adults (M = 18 years 5 months, SD = 8 months). Parental consent was obtained for all child participants as well as their own assent on the day of testing. Adult participants provided written informed consent prior to the experiment. Child participants were drawn from predominantly White middle-class schools. All participants were fluent in English.

Design, materials, and procedure

A 2 (Age: 11- or 18-year-olds) × 2 (Priming:

Experiment 2

To resolve this problem, consider three issues. First, because there is a growing consensus that false memories are generated during the encoding phase of the DRM task and not during retrieval (see Dewhurst, Bould, Knott, & Thorley, 2009), the likelihood that the critical lure was generated only during the recall test itself is relatively low. Although there is little doubt that the act of recalling an item during a memory test enhances its activation at that time, we argue that the main

General discussion

Together, these results are consistent with earlier findings with adults (Howe et al., 2010) and are the first to show that false memories, if sufficiently activated during encoding to be erroneously produced on a retrieval test, can prime both children’s and adults’ insight-based problem solving. Moreover, these developmentally invariant priming effects are not due to administering a recall test prior to the problem-solving task; rather, they are due to participants generating the critical

Acknowledgment

This research was supported by a Grant to M.L.H. from the Economic and Social Research Council of Great Britain (RES-062-23-0452).

References (35)

  • J.S. Anastasi et al.

    Examining differences in the levels of false memories in children and adults using child-normed lists

    Developmental Psychology

    (2008)
  • D.A. Balota et al.

    The English Lexicon Project

    Behavior Research Methods

    (2007)
  • E.M. Bowden et al.

    Normative data for 144 compound remote associate problems

    Behavior Research Methods, Instruments, & Computers

    (2003)
  • A.D. Castel et al.

    The dark side of expertise: Domain-specific memory errors

    Psychological Science

    (2007)
  • M. Coltheart

    The MRC psycholinguistic database

    Quarterly Journal of Experimental Psychology

    (1981)
  • F. Crews

    The memory wars

    (1995)
  • J. Deese

    Influence of interitem associative strength upon immediate free recall

    Psychological Reports

    (1959)
  • Cited by (61)

    • Priming older adults and people with mild to moderate Alzheimer's disease problem-solving with false memories

      2020, Cortex
      Citation Excerpt :

      A 3 (Group: Moderate AD vs Mild AD vs OAC) x 3 (Priming: prime/FM vs prime/noFM vs unprimed) mixed design was used, where the first factor was between-participants and the second was a within-participant factor. We followed the same procedure as Akhtar et al. (in press) and Howe et al. (2011, 2016) such that each participant was primed on half the subsequent CRAT problems with preceding DRM lists whose critical lures were also the solution to those CRAT problems. The order of both the DRM lists and CRAT problems was carefully counterbalanced to eliminate order effects.

    View all citing articles on Scopus
    View full text