Positive illusions, including those that “propel adaptive actions” (target article, sect. 13, para. 6) are maintained over time even (within limits) in the face of recalcitrant evidence. So they require sophisticated intertemporal accounting: Memory and associated forms of mental time travel must be enlisted if positive illusions are to be stable enough to enhance fitness, to be “pervasive, enduring, and systematic” rather than mere temporary errors (Taylor & Brown Reference Taylor and Brown1988, p. 194). So if McKay & Dennett (M&D) are right that certain kinds of ungrounded belief are adaptive, theories of memory are directly implicated. This link extends M&D's account of adaptive misbeliefs, suggesting new questions for memory research.
The sparse literature on functional analyses of remembering addresses the adaptive nature of forgetting and the puzzling luxury of autobiographical memory (Bjork & Bjork Reference Bjork, Bjork, Gruneberg, Morris and Sykes1988; Boyer Reference Boyer2008a; Reference Boyer, Boyer and Wertsch2009; Glenberg Reference Glenberg1997; Nairne 2005; Nairne et al. Reference Nairne, Thompson and Pandeirada2007; Schacter Reference Schacter2001). But the possibility that false memories (or ungrounded memories, which often contingently turn out false) could themselves be adaptive is surprising. False memories are usually seen as unfortunate outcomes of the constructive nature of remembering (Bernstein & Loftus Reference Bernstein and Loftus2009, p. 373), just as the manipulability of general belief-fixation is seen as epistemological trouble. But this standard line of thought is too quick, on two counts: reconstruction is not itself always distortion (Barnier et al. Reference Barnier, Sutton, Harris and Wilson2008; Sutton Reference Sutton, Robbins and Aydede2009), and, as M&D suggest, falsity need not always be maladaptive.
Biased contextual and autobiographical memories filter incoming information, and, in turn, self-enhancing (or otherwise illusory) information is preferentially attended and used to update memory. This bidirectional link between memory and positive illusions lies at the heart, for example, of temporal self-appraisal theory (Wilson & Ross Reference Wilson and Ross2003). Memory directly supports positive illusions: When unrealistic inflation of current self-image is difficult, people still “selectively recall and reconstruct evidence from the past that makes them feel good about their current selves” (Wilson & Ross Reference Wilson and Ross2001, p. 582). Conversely, the motivation to maintain positive self-regard in the present motivates misplaced or false memories, for example, as we subjectively move favourable past events forwards in time and unfavourable past events backward (Ross & Wilson Reference Ross and Wilson2002, p. 800). These loops between autobiography and the control of action, the self remembered and the working self, exhibit considerable variation (Conway Reference Conway2005). Some theorists stress strong reflexive feedback from self-representation into behavior, with ongoing integration lived out between actions and self-ascribed character, emotions, memories, and plans (Velleman Reference Velleman and Velleman2006). Others note gulfs between the story and the life, seeing narrative self-descriptions more like public relations reports that float free of the causal processes behind the government's or organization's behavior (Clark Reference Clark1994; Dennett Reference Dennett1991): This need not be morally or psychologically suspect, due to deliberate suppression or self-deceit, for the narrative and memory capacities – just like PR spokespersons – often don't have nor need the knowledge, or the contacts, or the access, either to get the back-story right or to directly feed in to future choices and actions.
Could false memories, as M&D might suggest, in some circumstances enhance affective, cognitive, and social well-being? A broad positivity bias in autobiographical memory is linked to enhanced emotion regulation: the tendency to accentuate the positive in recalling experiences from the personal past drives improvements in mood (Mather & Carstensen Reference Mather and Carstensen2005). Such memory biases are most securely demonstrated in older adults, whose positive affect is increased when reminiscing about positive past experiences (Pasupathi & Carstensen Reference Pasupathi and Carstensen2003); but younger adults get the same emotionally enhancing effects of biased autobiographical remembering when they remember while focussing specifically on their emotional state rather than on accuracy (Kennedy et al. Reference Kennedy, Mather and Carstensen2004). That this positivity bias drives false memories, not just general selectivity in recall, is suggested by a recent study in which older adults recall more false positive memories than false negative memories across a range of stimuli (Fernandes et al. Reference Fernandes, Ross, Wiegand and Schryer2008).
Narcissistic biases are thus not found only in subpersonal sensory systems. Just as the “narcissistic encoding” of sensory information is driven by our “sensory-motor projects” rather than by any ontological project of correctly cataloguing the world (Akins Reference Akins1996, p. 370), so the range of past experiences to which we are maximally sensitive in remembering are those relevant to embedded action, or salient for current and future decision-making. The gulf between design for fitness maximization and design for truth preservation (target article, sect. 9) may be widest in relation to truths about the past, especially about the distant past and about particular past events. As Boyer notes, the human “capacity to store unique episodes” is strange when “organisms learn about the past mostly to the extent that they can extract from past situations what is not unique about them, and what will be relevant in the future” (Boyer Reference Boyer, Boyer and Wertsch2009, p. 4). So, given the likely recent emergence of mental time travel, its biological function may involve just the broader self-related and social functions at work in the positive illusions described by M&D. In remembering, correspondence with reality is often trumped by coherence, truth by psychosocial utility (Alea & Bluck Reference Alea and Bluck2003; Conway & Pleydell-Pearce Reference Conway and Pleydell-Pearce2000; Wilson & Ross Reference Wilson and Ross2003). Evolved tendencies for forming memories which lack perfectly secure grounds are useful computational short-cuts, minimizing the heavy cognitive costs of precise source monitoring, of tracking the diverse origins of our representations of what is not present in the immediate environment. The idea parallels the suggestion that forgetting is an adaptive response to the computational challenges of retrieval access in context-sensitive systems (Anderson & Schooler Reference Anderson, Schooler, Tulving and Craik2000; Michaelian, submitted). But M&D's approach suggests the extra thought that sometimes-false memories can themselves, given the structure of the social and natural environment, be intelligent errors. Not only are the costs of trusting and acting on them in general less than those of strenuously applying forensic standards of warrant and justification, but (further) misremembering things in particular positive ways might have direct personal, motivational, and social benefits. Neither social nor motivational influence on memory is intrinsically malign: M&D remind memory researchers that cataloguing our susceptibility to misinformation in non-standard, experimentally skewed environments may blind us to the richer functions of remembering.
