Elsevier

Consciousness and Cognition

Volume 13, Issue 4, December 2004, Pages 789-820
Consciousness and Cognition

The consequences of encoding information on the maintenance of internally generated images and thoughts: The role of meaning complexes

https://doi.org/10.1016/j.concog.2004.07.004Get rights and content

Abstract

Three experiments investigated the hypothesis that internally generated images and thoughts were driven by meaning complexes, a construct which reflects a synthesis of semantic meaning and personal salience (Klinger, 1999). Experiments 1 and 2 contrasted the mutual inhibition between encoding words and non-words on: (i) the frequency that thoughts and images unrelated to the task (task unrelated thought, TUT) were experienced (Experiment 1) and (ii) on the intensity of images generated from long-term memory and maintained under dual task conditions, which whilst familiar were not of particular personal salience (Experiment 2). Experiment 3 examined the physiological arousal associated with the experience of TUT in a semantic encoding task. Evidence suggested that, in general, internally generated images and thoughts, irrespective of the personal salience, were suppressed by the co-ordination of information in working memory. In addition, only the experience of spontaneous images and thoughts of personal salience (TUT, Experiments 1 and 3) interfered reliably with the encoding/retrieval of semantic information from memory. Finally, in Experiment 3, physiological arousal, as indexed by mean heart rate, was associated with a high frequency of TUT. The results of all three experiments support the notion that the maintenance of spontaneously occurring images and thoughts is simultaneously influenced by both the semantic content and the personal salience of the information held in working memory.

Introduction

When engaged on a task, one’s attention can be directed externally towards task relevant material or alternatively can be directed internally, such as when one is involved in a daydream. Several authors have suggested that an internalised focus to attention, consistent with daydreaming, may be associated with important functional consequences, such as facilitating problem solving via the conceptual manipulation of semantic information (Baddeley, 1999; Binder et al., 1999). This position is exemplified in the following quotation: “By storing and manipulating internal information we organise what could not be organised during stimulus presentation, solve problems that require computation over long periods of time, and create effective plans governing behaviour in the future. These capabilities have surely made no small contribution to human survival and the invention of technology” (Binder et al., 1999, p. 85). One possible advantage of these internal processes is that they facilitate problem solving in the less hazardous environment of the neural workspace (Cleeremans & Jiménez, 2002; Dehaene & Naccache, 2001).

The subjective experience of internally generated images and thoughts can be understood in terms of a competitive process within the attention of the individual with both internal influences, such as current concerns (Klinger, 1999), dysphoria (Smallwood et al., 2004b) or current mood (Seibert & Ellis, 1991) and external influences, such as presentation rate (Antrobus, 1968), task (Teasdale, Lloyd, Proctor, & Baddeley, 1993), length of testing session (Smallwood, Obonsawin, & Reid, 2003c) or demand characteristics, dominating at various times (Antrobus et al., 1999). However, much of the mental content is derived internally without any overt cue (Klinger, 1999). Estimates suggest that approximately 20% of thoughts recorded during a laboratory task were unrelated to the current setting (Klinger, 1978), although this proportion is probably higher than in an everyday setting (Klinger & Cox, 1987). Hence, under laboratory conditions internally focused attention can be described as a situation in which “the attention of the individual becomes to some extent decoupled from the processing of external perceptual information” (Smallwood, Baraciaia, Lowe, & Obonsawin, 2003b). To account for the internal, stimulus-independent influences on attention it has been suggested that they can be understood in terms of the notion of meaning complexes: a higher order construct which can be seen as a synthesis of the notion of semantic meaning and the individual’s intentions or current concerns (Klinger, 1999). In the subsequent sections of this paper, we review evidence which supports the importance of both components of meaning complexes in internally generated information: (i) the personal salience of the information and (ii) the semantic nature of the information.

Perhaps the most obvious link between internalised attention and meaning complexes comes from the association between daydreaming and the individual’s current concerns. Current concerns can be considered to reflect the “hypothetical process active during the time that one has a goal” (Klinger, 1999, p. 439). In short, current concerns reflect the notion that many human goals extend beyond any given situation and that these goals need to be maintained in the mind of the individual, possibly in an unconscious form (Klinger, 1999). Thus one perspective on daydreams is that they represent the “unfinished business” of the individual (Singer, 1966).

Evidence from both naturalistic (Klinger, Barta, & Maxeiner, 1980) and laboratory settings (Klinger, 1978) suggest that when an individual’s attention is directed internally they are likely to be processing their current concerns (Klinger, 1999). First, when thinking was sampled on a day-to-day basis through the use of a pager the content of the participant’s thinking was often associated with the contents of their current concerns as sampled by a questionnaire. These concerns can be associated with the individual’s present life (67%), past or future (12%) or no particular time period (23%; Klinger & Cox, 1987). Second, an experimental induction which induced a personal salient concern with extensive implications for an individual, in this case a broadcast indicating that China had entered the Vietnam war, increased the likelihood of task unrelated thinking (0.32–0.45) and the frequency of errors (.045–.055), relative to a neutral control broadcast (Antrobus, Singer, & Greenberg, 1966).

Similarities can also be seen in the overlap between the content of one’s daydreams and the coping strategies that one employs on a day-to-day basis (Greenwald and Harder, 1995, Greenwald and Harder, 1997). These studies indicated that a consistent and reliable association could be observed between the content of daydreams experienced on a day-to-day basis and the content of sustained fantasies evoked in situations of distress as a source of comfort to the individual (Greenwald & Harder, 1995). For example, a high frequency of daydreams with a hostile content was associated with sustained fantasies at a time of crisis with a similar content. In a follow-up study, this finding was broadly replicated and the authors demonstrated additional overlap with the coping strategies employed by individuals. In this context, a high frequency of daydreams was associated with the employment of a coping strategy in which the individual “thinks endlessly” about a problem (Greenwald & Harder, 1997). These results suggest that, in addition to a process by which they refresh one’s awareness of one’s current concerns (Klinger, 1999), the experience of a daydream may allow the individual to mentally rehearse the coping strategies that they employ on a day-to-day basis (Greenwald and Harder, 1995, Greenwald and Harder, 1997). In either case, therefore, from a theoretical perspective, daydreams are likely to entail a strong motivational component because they reflect the processing of information of personal salience to the individual.

Evidence from the investigation of thinking via the use of thought probes has implicated the co-ordination of information in working memory in the experience of a daydream (Smallwood et al., 2002, Smallwood et al., 2003a, Smallwood et al., 2003b, Smallwood et al., 2003c; Teasdale et al., 1995). In these experiments a shift in attention is measured through the use of a thought probe which terminates the current block of the task Teasdale et al., 1993, Teasdale et al., 1995, Smallwood et al., 2003a, Smallwood et al., 2003b, Smallwood et al., 2003c. Using this paradigm, thoughts are recorded verbatim and classified in terms of whether they reflect attention to matters unrelated to the task in hand or the current situation (Task unrelated thought2; TUT, see Smallwood et al., 2003c for detailed criteria on making these judgements).

As the notion of meaning complexes implies the experience of TUT can be considered a situation in which the content of awareness departs from task relevant information processing (see Smallwood et al., 2003a, Smallwood et al., 2003b) and instead working memory resources becomes directed towards internally generated information of personal salience (see previous section). Evidence for this assumption comes from three sources: (i) tasks which require the co-ordination of information in working memory interrupt the genesis of TUT (see Smallwood et al., 2003a, Smallwood et al., 2003b, Smallwood et al., 2003c; Teasdale et al., 1993, Teasdale et al., 1995) and (ii) when the attention of the individual strays from the task and the individual experiences TUT their performance on the secondary task should suffer (Smallwood et al., 2002, Smallwood et al., 2003b; Teasdale et al., 1995), and (iii) perhaps the most direct source of support for the role of working memory resources in the experience of daydreaming comes from a series of studies in which daydreaming impaired text comprehension (Schooler, 2002; Schooler, Reichle, & Halpern, in press).

Consistent with the perspective outlined in this paper, tasks which require the co-ordination of information in working memory suppress the experience of TUT. For example, when presented with information for subsequent retrieval TUT is experienced at a lower frequency than when shadowing the same information, irrespective of the length of the digit string (Teasdale et al., 1993). In addition, articulatory suppression and rehearsal of a 5-digit load (Experiment 1) and a spatial motor task (Experiment 2) significantly reduced the frequency of TUT relative to a control task (Teasdale et al., 1995). Similarly, recalling a list of verbal stimuli yields lower frequencies of TUT than encoding the same stimuli, irrespective of the method of measurement of subjective experience (Smallwood et al., 2003a, Smallwood et al., 2003b). Moreover, as the central executive is thought to play a progressively smaller role as time on task increases (Baddeley, 1999) it is relevant that TUT increased with prior practice on both a pursuit rotor task and a memory load task (Teasdale et al., 1995, Experiment 3). Consistent with this finding the frequency of TUT was unaffected by the length of the block when generating verbal information (Smallwood et al., 2003c), whilst longer block lengths led to measurable increases in the frequency of TUT in both encoding and vigilance tasks. Overall, therefore, evidence supports the assertion that when the external task requires the co-ordination of information in working memory, the frequency with which TUT occurs is low (Smallwood et al., 2003a, Smallwood et al., 2003b, Smallwood et al., 2003c; Teasdale et al., 1993, Teasdale et al., 1995).

A second prediction, consistent with the pattern of mutual inhibition between daydreaming and external task performance, is that when the individual’s attention strays from the task, and they experience TUT, their performance on the secondary task should suffer (see Baddeley, 1993). Consistent with this prediction, experiencing TUT whilst engaged on a random generation task was associated with impairment in the individual’s ability to generate random sequences (Teasdale et al., 1995, Experiment 4). In addition, experiencing TUT whilst encoding verbal information, led to: (i) a shift in retrieval response towards retrieval of information on the basis of familiarity, rather than recollection and (ii) an increased reaction time during study (Smallwood et al., 2003b). These impairments in task performance are in contrast to a large body of evidence that suggests that TUT during vigilance has little or no implications for task performance (e.g., Antrobus, 1968; Giambra, 1995; Grodsky & Giambra, 1989; see also Teasdale et al., 1993). Moreover, whilst recent work has implicated the experience of TUT in vigilance performance (Smallwood et al., 2004a), this detriment was limited to successive vigilance. As successive vigilance tasks require the participant to “discriminate between a currently viewed stimulus and a standard representation of a specific stimulus held in working memory” (Desmond, Matthews, & Bush, 2001, p. 1386) whilst in simultaneous vigilance “all of the necessary information to make the discrimination is presented in the current field” (Desmond et al., 2001, p. 1386). This finding provides further support for the notion that daydreaming monopolises working memory resources and prevents the co-ordination of task relevant information towards task completion.

The final source of evidence to indicate that daydreaming interferes with the co-ordination of information in working memory comes from the analysis of daydreaming during text comprehension (Schooler, 2002; Schooler et al., in press). The co-ordination of information in working memory has been implicated in the comprehension of text (Daneman & Carpenter, 1980; Gathercole & Baddeley, 1993; Just & Carpenter, 1992; Oakhill et al., 1986, Oakhill et al., 1988) and the role has been described as follows: “Processing is needed to recognise the lexical items represented by the surface forms of language, access their syntactic and semantic specifications and interpret the meanings of sentences” (Gathercole & Baddeley, 1993, p. 222). In this context it is relevant that work by Schooler et al. (in press) identified that daydreaming during text comprehension was associated with periods of low comprehension. In the words of Schooler et al. (in press), during reading comprehension“Your eyes may continue moving across the page, the phonology of the words may continue sounding in your head, yet fundamentally your mind may be elsewhere.”

Not only, therefore, is the experience of TUT less likely when the task requires the co-ordination of information towards task completion, but when TUT is experienced it can lead to a measurable impairment in the co-ordination of task relevant material in working memory, either in terms of: (i) our ability to generate random sequences (Teasdale et al., 1995, Experiment 4), (ii) our ability to encode the stimulus in the first instance (Smallwood et al., 2002, Smallwood et al., 2003a, Smallwood et al., 2003b) and (iii) our ability to extract higher-order meaning from the text (Schooler et al., in press). Thus empirical evidence suggests “ when one is paying attention to external events one cannot let ones mind wander; or conversely if ones mind wanders one is not paying attention (Antrobus, 1999, p. 8).

It is clear that existing evidence supports the notion that “meaning complexes” (Klinger, 1999) can be considered to play an important role in the internal regulation of the information in working memory. First, evidence suggests that the content of daydreams tend to be highly salient to the individual (Greenwald & Harder, 1997; Klinger et al., 1980). Second, the experience of TUT competes for working memory resources and therefore interferes with our ability to co-ordinate task relevant information (Antrobus et al., 1966; Teasdale et al., 1995). Recent work has demonstrated that this detriment extends beyond simple sensorimotor tasks such as that employed by Antrobus (1968) and TUT has been shown to interfere with an individuals ability to derive complex semantic information from a stimulus (Smallwood et al., 2002; Smallwood et al., 2003b; Schooler et al., in press).

The experiments presented in this paper are concerned with directly evaluating the role that “meaning complexes” (Klinger, 1999) play in the maintenance of internally generated streams of information. As the content of awareness of an individual can be conceptualised as a competitive process between internal and external stimuli (Antrobus, 1999) we can derive a series of hypotheses about the conditions that promote the experience of TUT and the consequences of this state of mind. First, if internal experience does involve a semantic component, then the task of co-ordinating semantically rich task relevant information in working memory should interfere with the experience of TUT. This relationship is qualified, however, because experimental evidence suggests that the relationship is one that may represent an affordance (Smallwood et al., 2003b, Smallwood et al., 2003c). That is, merely experiencing semantic information is not likely to disrupt the experience of TUT, instead the participant needs to engage with the stimulus materials (see Smallwood et al., 2003b, Experiment 2; see also Teasdale et al., 1993). In all three experiments presented in this paper, therefore, we examine the mutual inhibition between the encoding of semantic information from the external environment on the experience of internally generated images. In these experiments, this mutual inhibition can be operationalised in two dimensions as: (i) a lower frequency of TUT (Experiment 1)/less vivid internally generated images (Experiment 2) and (ii) the experience of the internally generated images should yield an impairment in the ability to encode stimuli from the environment.

Second, because the spontaneous streams of thought, such as TUT, reflects the processing of one’s current concerns, the internal information should reflect the processing of personally salient information. A large body of evidence suggests that emotive or highly salient material tends to attract the attention of the individual (e.g., Ohman, Flykt, & Esteves, 2001) and may do so in an automatic fashion (MacLeod, 1991; Martin, Williams, & Clark, 1991; McKenna & Sharma, 1995). It follows, therefore, that “if the emotional arousing stimuli are central to the task and are therefore at the focus of processing—they facilitate perceptual and attentional responses. If on the other hand, incidental stimuli are emotionally arousing, they distract from the target task” (Klinger, 1999, p. 36). The previously mentioned study by Antrobus et al. (1966) provides support for this notion because the experimental induction which induced personal concerns produced both: (i) elevations in TUT and (ii) increases in errors. On this basis, it is probable that some of the interference between TUT and an external task, as described in the literature (Antrobus et al., 1966; Schooler et al., in press; Smallwood et al., 2002; Smallwood et al., 2003a, Smallwood et al., 2003b; Teasdale et al., 1995, Experiment 4), is a result of the personal salience of the information maintained in working memory.

We addressed the role of personal salience in a different manner in each of the three experiments presented in this paper. In Experiment 1, we investigated the role that dysphoria plays in the generation of TUT. Dysphoria can be considered to be a state in which the current self departs from the ideal self (Pyszczynski & Greenberg, 1987), and consistent with this notion, research has demonstrated that dysphoria but not rumination is associated with high levels of TUT (Smallwood et al., 2004b; see Smallwood, O’Connor, & Heim, submitted for a replication; see also Seibert & Ellis, 1991). On this basis we hypothesised that consistent with previous research, the experience of TUT would be associated with higher levels of dysphoria (Experiment 1). In the second experiment, we examined the consequences of processing internally generated information on the vividness of images that are not personally salient. If, as suggested by Klinger (1999) and others, daydreams have an important influence on information processing because they are emotionally significant to the individual, one could hypothesise lower levels of mutual inhibition between internally generated images in Experiment 2, because whilst they are familiar they are not personally salient to the individual. Finally, in Experiment 3 we measured the physiological arousal that occurs during a semantic encoding task. Previous research has suggested that in circumstances in which the report of TUT was expected to be less frequent was associated with reductions in physiological arousal (Antrobus, 1973; see also Antrobus, 1999). We hypothesised that, if TUT was associated with the individual’s current concerns, we would be able to detect a positive association between TUT and measures of physiological arousal.

The study of the maintenance of internal streams of information requires the investigation of subjectively reported information. With such a study, it is important to validate the subjective information reported by the individuals during task completion by reference to measurable aspects of task processing (Baars, 1988). In particular, two criteria have been provided to validate the verbal reports provided during the investigation of TUT in the context of laboratory tasks (Smallwood et al., 2003b). First, when classifying thinking, in addition to high levels of inter-rater consistency, it is important to validate the coding employed by the judges through comparison with retrospective measures completed by the participants themselves. Reasonable agreement between these two sources of information would suggest that there was agreement between experimenters and the participants on the contents of thinking. In Experiment 1, we demonstrate reasonable correspondence between these two forms of measurement of TUT. Second, it is important to identify the consequences of the subjective state on concurrent information processing, such as a meaningful correspondence between experiencing TUT during encoding and the subsequent retrieval of information from memory presented in the literature (Smallwood et al., 2002, Smallwood et al., 2003b). Such a correspondence between verbal response and behaviour provides evidence that the experience of TUT is more than a verbal label applied by experimenters—it has actual consequences for ongoing task completion.3 Finally, in Experiment 3, we demonstrate that high levels of physiological arousal, as measured by heart rate, is associated with high frequencies of TUT. This provides an important alternative source of validation because the activity of the autonomic nervous system is less available to conscious control than either measures of retrospective self-report or task performance.

Section snippets

Aim

The first aim of Experiment 1 was to investigate whether the experience of TUT would be contingent upon the semantic content of the external task. To achieve this aim we compared TUT across three conditions: (i) encoding words, (ii) encoding non-words, and (iii) during vigilance. If TUT reflects the co-ordination of semantic information an implication of the notion of meaning complexes, the fewest examples of TUT should be reported whilst encoding words. In addition to the comparison of the

Aims

The aim of Experiment 2 was to examine whether instructing participants to hold images in their head which are not personally salient yields: (i) a similar pattern of inhibition between external semantic stimuli and internal images and (ii) the same pattern of mutual inhibition between externally encoded information and internally maintained information. To this end, we employed the paradigm by Baddeley and Andrade (2000).

In their paper, Baddeley and Andrade (2000)

Aims

The main aim of Experiment 3 was to assess whether the hypothetical personal significance of the information processed during the experience of TUT was observable in an objective physiological measure. Little is known about the physiological correlates of TUT, although previous research has identified an association between body temperature and TUT (Giambra, Rosenberg, Kasper, Yee, & Sack, 19881989). Similarly, physiological measures, including heart rate and Galvanic Skin Response (GSR) were

Methodological limitations

Before dealing with the implications of these experiments, it is worth dealing with several experimental limitations. First, it is clear that large differences can be seen between the two paradigms employed in this paper. These differences include methodological differences such as task length, stimulus presentation rate and the dependent measures themselves (frequency in Experiment 1 and vividness in Experiment 2). In addition, the results differ in terms of the consequences on performance

Speculative conclusions

It is worth noting the implications that these results have for our understanding of the mechanisms that contribute to conscious awareness. Whilst not directly measured in either experiment, the effort directed towards task completion is likely to be larger when encoding non-words for subsequent retrieval. However, in the non-word condition of Experiment 1 there was no advantage to the individual by concentrating on the task, either in terms of reaction time or the number of words correctly

Acknowledgments

Thanks to all the individuals who took part in this series of experiments. The experiments presented in this paper were partly funded by a grant from the Research & Development Fund from the Psychology Department of Glasgow Caledonian University made to the first author. Thanks to Dr. Marc Obonsawin for his help in the recording of the physiological data, Gordon McAlpine for writing the computer programmes for Experiments 1 and 2 and to the comments of John Antrobus and an anonymous reviewer on

References (64)

  • J.S. Antrobus

    Eye movements and non-visual tasks

  • J. Antrobus

    Toward a neurocognitive processing model of imaginal thought

  • B. Baars

    A cognitive theory of consciousness

    (1988)
  • A. Baddeley

    Human memory

    (1999)
  • A. Baddeley

    Working memory and conscious awareness

  • A. Baddeley et al.

    Working memory and the vividness of imagery

    Journal of Experimental Psychology: General

    (2000)
  • J.R. Binder et al.

    Conceptual processing during the conscious resting state: A functional MRI study

    Journal of Cognitive Neurosciences

    (1999)
  • S.A. Bird et al.

    The effect of bimodal input on implicit and explicit memory: An investigation into the benefits of within-language subtitling

    Applied psycholinguists

    (2002)
  • Bradley, M. M., Lang, P. G., 1999. Affective Norms for English Words (ANEW). Gainesville FL. The NIMH Center for the...
  • A. Cleeremans et al.

    Implicit learning and consciousness: A graded, dynamic perspective

  • A. Damasio

    The feeling of what happens: Body emotion and the making of consciousness

    (2000)
  • Desmond, P. A., Matthews, G., & Bush, J. (2001). Sustained vigilance during simultaneous and successive vigilance...
  • D.C. Fowles et al.

    Publication recommendations for electrodermal measurements

    Psychophysiology

    (1981)
  • S.E. Gathercole et al.

    Working memory and language

    (1993)
  • A. Grodsky et al.

    Task-unrelated-thought frequency as a function of age: A laboratory study

    Psychology and Aging

    (1989)
  • L.M. Giambra et al.

    A circadian rhythm in the frequency of spontaneous task-unrelated images and thoughts

    Imagination, Cognition, and Personality

    (1988–1989)
  • A. Grodsky et al.

    Task unrelated images and thoughts whilst reading

  • A. Grodsky et al.

    The consistency across vigilance and reading tasks of individual differences in the occurrence of task unrelated and task related images and words

    Imagination, Cognition and Personality

    (1990)
  • D.F. Greenwald et al.

    Sustaining fantasies, daydreams and psychopathology

    Journal of Clinical Psycholology

    (1995)
  • D.F. Greenwald et al.

    Fantasies, coping behaviour and pychopathology

    Journal of Clinical Psychology

    (1997)
  • J.F. Golding

    Phasic skin conductance activity and motion sickness

    Aviation, Space and Environmental Medicine

    (1992)
  • L.L. Jacoby

    Invariance in automatic influences of memory: Towards a users guide for the process-dissociation procedure

    Journal of Experimental Psychology: Learning Memory and Cognition

    (1998)
  • Cited by (78)

    • A wandering mind is a forgetful mind: A systematic review on the influence of mind wandering on episodic memory encoding

      2022, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      First, mind wandering during episodic memory encoding had a significant hindering effect on later retrieval performances in 20 studies. This effect was very consistent as it impacted the retrieval of words (Garlitch and Wahlheim, 2020; Maillet and Rajah, 2013; Peterson and Wissman, 2020; Seibert and Ellis, 1991; Smallwood et al., 2004; Smallwood et al., 2003a; Smallwood et al., 2003b; Thomson et al., 2014b; Xu and Metcalfe, 2016), images (Blondé et al., 2021; Chinchanachokchai et al., 2020a; deBettencourt et al., 2018; Krasich et al., 2018; Zhang et al., 2021), source (Blondé et al., 2021; Maillet and Rajah, 2014, 2016), audio messages (Chinchanachokchai et al., 2020a, 2020b), virtual reality scenes (Blondé et al., 2020) and one-shot stimulus-control learning (Whitehead et al., 2021). Among these studies, 15 demonstrated this effect using a voluntary encoding paradigm, and 5 found it by using an incidental encoding.

    View all citing articles on Scopus
    1

    Present address: University of British Columbia, Vancouver, BC.

    View full text