ارتباط بین تئوری ذهن و حافظه اپیزودیک: شواهد برای توسعه آگاهی خودادراکی

The study investigated a link between theory of mind and episodic memory involving autonoetic consciousness (Tulving, 1985). Eighty-nine Japanese 4- to 6-year-olds received two versions of a false belief task, a task of aspectuality or knowledge origins, and four memory tests. After controlling for age, most theory of mind abilities showed no interrelations, and own and other’s belief understandings in deceptive appearance tasks were solely related to source memory, but not to free recall, temporal ordering, or working memory. Moreover, even when age and verbal intelligence were controlled, the association between representational change and source memory was highly significant in 6-year-olds but not in 4- and 5-year-olds. Results suggest that during development only a particular kind of theory of mind ability is integrated with episodic memory.

In much recent literature on both adult memory and child cognition, theories of episodic memory have been proposed, especially from the perspective of subjective awareness or consciousness involved in memory phenomena (Perner, 1990 and Perner, 2000; Suddendorf & Corballis, 1997; Wheeler, Stuss, & Tulving, 1997). As originally proposed by Tulving (1985), there is a distinction between remembering and knowing that accompanies conscious recollections of past events. The remembering, or episodic memory, of an event depends on autonoetic (self-knowing) consciousness that we have subjectively thought or experienced it before; whereas knowing, or semantic memory, of an event entails noetic (knowing) consciousness that we know or objectively think about it without the additional recollective awareness. In an elaboration of Tulving’s framework, Wheeler et al. (1997) considered episodic memory as a highly evolved human capacity not only to mentally travel back in one’s personal past but also to mentally project anticipated events into one’s subjective future (see for a similar discussion, Suddendorf & Corballis, 1997). On the basis of Tulving’s (1985) distinction of consciousness involved in memory, Perner, 1990, Perner, 1991 and Perner, 2000 hypothesized that episodic memory development was dependent on children’s developing theories of mind (i.e., abilities to understand mental states such as thoughts and beliefs and to impute them to people). Perner postulated that before the age of 4 or 5 years when children do not yet reflect on the perceptual origin of their own knowledge, they cannot encode events as personally experienced (see also, Welch-Ross, 1995). Episodic memories emerge only with the development of what Perner (1990) labeled “experiential awareness”—a state of consciousness similar to autonoetic consciousness. In support of this hypothesis, Perner and Ruffman (1995) found that 3–6-year-olds’ ability to understand the informational relation between knowledge and sensory experiences (as reflected in performance on theory of mind tasks) was related to free recall, even when cued recall and verbal intelligence were controlled. The possibility of a link between theory of mind and memory has recently attracted researchers’ attention, especially in the field of children’s suggestibility (i.e., biasing effects of incorrect, postevent information on their memory of the original events). Welch-Ross, Diecidue, and Miller (1997) demonstrated that 3–5-year-olds’ theory of mind ability predicted their suggestibility (see Welch-Ross, 1999 and Welch-Ross, 2000). Templeton and Wilcox (2000) also found that the amount of suggestibility in 6-year-olds was influenced by their theory of mind ability, whereas that in 3- and 4-year-olds was not. These results strongly indicate a link between theory of mind and children’s mnemonic ability to resist being misled by incorrect information. However, the existing evidence from suggestibility effects has not been concerned with recollective awareness that accompanies the memory phenomena. There has been virtually no study except Perner and Ruffman’s (1995) that directly examined the relation between theory of mind and memory from the aspect of autonoetic consciousness. The present study examined the relation focusing on this particular kind of recollective awareness. In conducting such a study, it is important to include a range of measures of theory of mind. In the past, children’s theory of mind abilities have been assessed by a wide variety of tasks. The most widely used are two types of traditional false belief tasks that measure the ability to understand representations. One is an unexpected transfer task (Wimmer & Perner, 1983) in which an object put by a protagonist in one place is transferred to another place without the protagonist’s awareness and a child who has observed the whole event is asked the protagonist’s false belief about the object’s whereabouts. The other type task is a deceptive appearance task (e.g., Naito, Komatsu, & Fuke, 1994; Perner, Leekam, & Wimmer, 1987). In this task, children are shown an object (e.g., “a sweets box”) whose appearance has a deceptive nature (e.g., it apparently contains sweets) and after discovering the true nature of the object (e.g., it really contains a pencil), children are asked to predict another person’s false belief about its nature. Furthermore, in the deceptive appearance tasks children are often asked to remember the false belief they once held when they were first shown the object. The understanding or remembering of one’s own past false belief is also referred to as representational change (Gopnik & Astington, 1988). Using the two kinds of tasks, the theory of mind literature has repeatedly reported that whereas most 4-year-olds successfully understand false beliefs, most 3-year-olds do not and has thereby treated the different false belief tasks as interchangeable. For example, Wellman, Cross, and Watson’s (2001) comprehensive meta-analysis revealed that developmental changes in performance did not differ between the various false belief tasks. However, Holmes, Black, and Miller (1996) pointed out that there had been only a few studies that directly compared children’s performance on different versions of the belief task and that these studies actually showed inconsistencies in their performance. Holmes et al. also found that children from a Head Start population found the transfer task easier than the appearance task and that fewer than half of the children performed consistently across the tasks. Their findings suggest that different false belief tasks do not necessarily measure an essentially unitary cognitive ability. The differences and inconsistencies in belief understandings found in some studies (e.g., Holmes et al., 1996) could be explained in terms of the autonoetic consciousness (Tulving, 1985), or experiential awareness (Perner, 1990), involved in the various false belief tasks. In the deceptive appearance tasks, children experience the false belief themselves before having to attribute it to another, an experience that is not part of the transfer task. Compared to the transfer task, the false belief measured by the deceptive appearance tasks would thus more likely involve explicit memory of past episodes (i.e., the context where children acquired their own belief). If this were the case, then it is expected that the extent to which false belief understandings are related to episodic memory performance would vary depending on the degree of autonoetic consciousness involved in each of the belief tasks. The present study examined this hypothesis. Another well-documented category of theory of mind abilities involves the understanding of one’s own knowledge and its origins (e.g., Gopnik & Graf, 1988; O’Neill, Astington, & Flavell, 1992; Povinelli & deBlois, 1992; Taylor, Esbensen, & Bennett, 1994; Whitcombe & Robinson, 2000; Wimmer, Hogrefe, & Perner, 1988). Some researchers (Drummey & Newcombe, 2002; Leichtman, Morse, Dixon, & Spiegel, 2000; Robinson, 2000) have also categorized this ability as source-monitoring skills (i.e., an ability to identify the origins of one’s own knowledge, beliefs, and memories), which are among the factors known to affect children’s suggestibility (e.g., Ceci & Bruck, 1993; Leichtman et al., 2000; Thierry & Spence, 2002). One task that measures this ability is O’Neill et al.’s (1992) “aspectuality” task, which assesses an understanding that our senses inform us only about certain aspects of the perceived objects. In their study, 3–5-year-olds were presented a pair of objects that either looked the same but felt different, or that felt the same but looked different. After hiding one of the objects inside a tunnel, children were asked whether they need to see the object or feel it to determine its identity. O’Neill et al. demonstrated that typically until 4 years of age, children could not understand which perceptual aspect was linked to their knowledge about the object. It seems that by 5 years, most children in the western world have acquired the ability to understand informational sources of their knowledge, as well as false beliefs. However, a study with Japanese children showed that their understanding of aspectuality lagged behind their Western counterparts’ (Ruffman, Perner, Naito, Parkin, & Clements, 1998, Experiment 4). The present study attempts to replicate this effect with a broader range of tasks that include the aspectuality task and the false belief tasks of unexpected transfer and deceptive appearance. The present study also adapted a task developed by Taylor et al. (1994), who investigated 4- and 5-year-old’s attention to learning episodes of their knowledge. Specifically Taylor et al. taught children a set of facts. Later, although the children remembered these facts, they could not accurately recall when or how they had acquired them. The Taylor et al. design was adapted in the present study to assess children’ s source monitoring. This task measures one of the defining features of episodic memory, according to Tulving and colleagues’ (Wheeler et al., 1997) most current theory; that is, it assesses the subjective aspect that one can travel back in time and recollect the initial experience and its specificity in locating the event in terms of time and space. In addition to the source test, other memory tasks were included in the present study; these included free and cued recall, temporal ordering, and working memory. Free and cued recall were administered because Tulving (1985) has claimed that free recall requires episodic memory whereas cued recall requires semantic memory knowledge; the former has also been related to children’s understanding of knowledge origins as assessed by the aspectuality task (Perner & Ruffman, 1995). Especially for young children, however, traditional free recall tests that require recall of just the content of word or picture lists might be accomplished without keen awareness of recollective experience as Tulving postulated for adults. Moreover, in the Perner and Ruffman study, although the partial correlations between free recall and aspectuality were significant after controlling for cued recall and verbal intelligence, the most potent developmental factor, age, was not controlled. The present study examined the relation between free recall and theory of mind abilities controlling for not only cued recall and verbal intelligence but also age. A temporal order memory task was included because of previous findings that adults’ performance on this task was similar to source memory monitoring in the involvement of particular brain functions (Janowsky, Shimamura, & Squire, 1989; Schacter, Kagan, & Leichtman, 1995; Shimamura, Janowsky, & Squire, 1990). As far as autonoetic consciousness is concerned, however, the temporal ordering would not be a genuine episodic memory test (see Wheeler et al., 1997; for a similar discussion), and thus the relation of temporal memory to theory of mind abilities was expected to be weak. Finally, the study included a listening span test developed by Ishio and Osaka (1994) to assess working memory capacity in Japanese preschool children. Researchers have recently paid special attention to relevance of executive function (which includes working memory) to the theory of mind development, and studies on this issue are accumulating (e.g., Perner & Lang, 2000; Russell, 1996). However, findings have not been consistent, particularly on the relation between working memory and theory of mind. Some studies showed reliable associations of children’s working memory with some of the theory of mind tasks (Davis & Pratt, 1995; Gordon & Olson, 1998); others reported that the association was not significant once age-related effects were controlled (Hughes, 1998). It has been discussed that the strong association between working memory and theory of mind is due in part to the “dual” nature of working memory tasks that require children simultaneously two different cognitive activities (Gordon & Olson, 1998). The present study hence used the listening span test that is considered to be a dual task (i.e., while listening to the sentences, children have to memorize the target words embedded in them). In summary, two different kinds of traditional false belief tasks and the aspectuality task were used as theory of mind tasks; in addition, tests of source memory, free recall, memory of temporal order, and listening span were used to assess children’s memory. The Japanese children who participated in the study were between ages of 4 and early 7 years, rather than 3 and 5 years reported predominantly in most theory of mind literature. This age range was determined according to the previous findings that Japanese 5-year-olds found the aspectuality task quite difficult (Ruffman et al., 1998) and that some source monitoring skills were not mastered by 5 years of age (e.g., Foley, Harris, & Hermann, 1994; Lindsay, Johnson, & Kwon, 1991; Taylor et al., 1994; for a review see Roberts & Blades, 2000). The relation between theory of mind and memory was examined with age, verbal intelligence, and when necessary, cued recall controlled.