Consciousness and Memory

Consciousness and Memory Penelope Rowlatt University College, London University. I am grateful for the support and advice of Professor Paul Snowdon and for comments from Chris Nunn, Geraint Rees and Neville Moray 12 December 2008 1. Introduction Thomas Nagel famously suggested that: “an organism has conscious mental states if and only if there is something that it is like to be that organism – something it is like for the organism” (Nagel, 1974, p. 436). However, it is difficult to imagine a creature being conscious of anything if it has no memory. As Baars and McGovern (1996, p. 66) pointed out, the only conscious events we can study are those that are “reportable”, that is, remembered for a short time at least. In this paper I examine the idea that Nagel-type consciousness might be what it is like to have certain types of memory. If this were the case, then different types of consciousness, ‘phenomenal’ or ‘access’ for example, could be defined in terms of the respective relevant memory stores. In his important 1995 paper, “Facing up to the problem of consciousness”, David Chalmers refers to the “information contents” that are “accessible to central systems” and calls them the ‘contents of awareness’ (Chalmers, 1995, p 212). There is coherence, he asserts, between “the structure of consciousness and the structure of awareness”. If we interpret Chalmers’ ‘information contents’ as being that which is contained in short-term memory stores, then his assertion implies coherence between the structure of consciousness and the structure of short-term memories – it is that coherence that I explore in this paper. Consciousness has frequently been linked with memory in academic writing over the years. For example, William James, writing in 1890 and quoting Richet (1886), stated: "Without memory no conscious sensation, without memory no consciousness" (James, 1890, Chapter XVI). Earlier, Richet had written (1884, p. 32): “ … for a conscious sensation … to occur, there must be a present of a certain duration, of a few seconds at least”; and: "… to suffer for only a hundredth of a second is not to suffer at all; and for my part I would readily agree to undergo a pain, however acute and intense it might be, provided it should last only a hundredth of a second, and leave after it neither reverberation nor recall". Among more recent authors, Ray Jackendoff writes of “awareness” as being “supported by the contents of short-term memory” (1987, p 280). Cowan (1997, p 77) sees “Short-term memory taken as a whole, including both sensory and nonsensory aspects” as representing “the subject’s present mind”, while Baddeley (2003, p 837) refers to a “link between memory and conscious awareness.” In the main part of this paper I distinguish three different classes of the “conscious of” type of experience, broadly familiar from the philosophy literature, and identify each with a specific type of short-term memory that has been discovered and studied by psychologists. This is followed by a brief discussion of the theoretical basis for identifying Nagel-type consciousness with memory. 2. Three notions of consciousness In this section I describe three notions of consciousness which are broadly consistent with three types of short-term memory store that have been identified by psychologists. These memory stores differ from the long-term memory in that information remains in them for a while, and while it is there it is available to consciousness, but once gone it cannot be recalled from this source It therefore may well make sense to postulate that these three types of consciousness are “what it is like” to have these particular types of memory. The three types of consciousness are defined in terms of three different mental activities of which we are conscious. I discuss them at greater length in the following sections presenting only a brief sketch here. The first two concern consciousness of information that can be drawn upon and used in cognitive processes. The third is consciousness of the happening of these cognitive processes. The figure below shows a possible schematic view of the relevant memory stores. External information Figure 1: A Hypothetical Structure of Consciousness and Memory A possible structure of consciousness and memory is illustrated in this figure. The sensory memory stores correspond to phenomenal consciousness, the short-term memory stores (perhaps Baddeley’s phonological loop, visuospatial skechbook and episodic buffer) to access consciousness and the conscious cognitive processing to thought consciousness. First, consider consciousness of information received through perceptual experience, which may be about our environment (sight etc.) or our bodies (proprioception, the physical effects of emotions, and so on) Many of the items in Chalmers’ list of things we experience (1995, p 210) come into this category and which is generally referred to as ‘phenomenal consciousness’ or p-consciousness. This term has been used in a wide variety of ways in the literature Psychologists have identified memory stores that are associated with the receipt of sensory information. Most of the experiments relating to these involve visual and auditory experiences, and they indicate that a sensory experience persists for a few hundred milliseconds after the external signal has ceased (see, for example, Efron, 1970). This is known as ‘sensory persistence’. It seems that all the information in the sensory memories is potentially available to consciousness, but what one is actually conscious of at any moment in time is determined by ‘attention’. A choice is made of which features shall receive conscious attention from the visual scene before one, from the sounds one can hear, from the complicated range of emotions one feels at any particular moment, and so on. Alan Baddeley’s work (1996, p 8, for example) suggests that the choice of what to attend to might be made in the ‘central executive’ part of the ‘working memory’ (see below), which may act as attention controller along with its other roles. Second, consider consciousness of other information used in cognitive processing. Much of the information consciously used in cognitive processing will not relate to the current experience of one’s external and internal senses. It may have been drawn from long-term memory, for example. The “… mental images that have been conjured up internally …” referred to by Chalmers (1995, p. 201) would be included in this category as they are perceived as being seen with the “mind’s eye” rather than the eye Ned Block has suggested the term ‘access consciousness’, or ‘a-consciousness’ for consciousness of information available for cognitive processing, which he describes as information that is available “for use in reasoning and rationally guiding speech and action” (1995, p. 227). I draw a distinction in this paper between awareness of perceptual and non-perceptual information, which I call p-consciousness and a-consciousness respectively (see figure 1). I suggest that the information stores that psychologists associate with Baddeley’s ‘working memory’ (Baddeley and Hitch, 1974), or the ‘short-term memory store’ of Cowan (1988), correspond to a-consciousness. Broadly, this is a short-term memory of ‘activated’ information that may be held in various forms, examples being Baddeley’s phonological loop and visuospatial sketchpad. It may be that only a subset of this short-term store is the ‘focus of attention’ at any given time (Baars, 1997). The third mental activity we need to consider is consciousness of cognitive processing, that is, consciousness of thinking (imagining, reasoning, etc.) and of the implementation of decisions (the triggering of speech, muscular activity and so on). This type of consciousness is less well represented in the philosophical literature than phenomenal consciousness and access consciousness (but see Peacocke, 2007 and Prinz, 2007). David Chalmers refers to it (1995, p 201), citing: “… the experience of a stream of conscious thought …” as an example of ‘experience’. One might call this type of consciousness ‘thought consciousness’ or t-consciousness. It would not be possible to create something in one’s imagination, or follow a line of reasoning, or speak a sentence if one were unable to remember the early parts of the relevant train of thought. All cognitive activities must therefore involve some sort of short-term memory. The ‘central executive’ part of the working memory structure initiated originally by Baddeley and Hitch (1974) is the obvious candidate. 3. Phenomenal consciousness and sensory memories David Chalmers uses the word ’experience’ in the context of the term consciousness (1995, p 201). He cites a number of different types of phenomenal consciousness saying: “When we see, for example, we experience visual sensations: the felt quality of redness … the sound of a clarinet, the smell of mothballs. Then there are bodily sensations, from pain to orgasms; … the felt quality of emotion, …”. The wide variety of subjective experiences associated with sensory perception are the familiar bundles of sensory experience that we understand as phenomenal consciousness, which is, I suggest, “what it is like” to have short-term sensory memories of the type the psychologists call ‘sensory persistence’. The visual form of sensory memory is called ‘iconic memory’, the auditory form ‘echoic memory’ and the olfactory form ‘odor memory’ (see Neath and Surprenant, 2003 for a review of evidence relating to them). Not surprisingly, no-one seems to have investigated the likelihood of similar types of memory store for the internal senses, but it is difficult to see how we could be aware of pain, the physical effects of emotions, proprioception etc. if there were no corresponding sensory memory. As William Burnham (1888, p 575) said: “Each sense … of the body may be said to have its memory …” Sensory persistence was demonstrated originally by Segner in 1740. According to Mollon, see p 182 of his unpublished PhD thesis (Baddeley, 1997, p 10) He attached a red hot coal to a cartwheel and rotated the cartwheel at a speed just fast enough for the light to form a complete circle to the eye in order to estimate its duration, which he found to be around a hundred milliseconds. Later experiments confirmed that duration for both the iconic and the echoic memories. For example, Efron (1970) compared the offset of a signal as perceived by a subject with its actual offset: he arranged for a reported simultaneity of the perceived offset with the onset of another signal and found a minimum perceptual duration of around 100-200ms for both visual and auditory perceptions. In a masking experiment, Massaro (1970) found that subjects’ ability to identify the pitch (high or low) of a 20ms tone improved as the period between the end of the test-tone and the start of the masking tone increased, reaching a plateau at around 250ms. He concluded that the auditory image persists for a few hundred milliseconds after the tone is terminated, and that the subject needs that extra time to extract the information relating to the pitch. If an image of everything we see persists in a sensory memory for around a tenth of a second after the signal has gone, then a question arises as to why our vision is not a great confusion of smudgy impressions. The answer has to be that the brain has learned to sort out the smudginess and present us with a sharp and clear impression of the actual information content out of the smudgy mess. Another question then arises: why does Segner’s hot coal experiment result in a circle of light being seen (that is, why has the brain not sorted it out and replaced the ‘tail’ by the visual field behind the light)? A possible answer to this is that if the contrast between the signal and the background is sufficiently pronounced then the persistence of the signal successfully attracts the attention; with a lesser contrast the persistence would be corrected out automatically. Many have speculated about the whereabouts of the iconic memory in the visual neurosystem. One possibility is that the iconic memory takes place at the retina, and is really no more than a form of after-image (Sakitt, 1976), another is that it is located in the primary visual cortex (V1). There are many reasons to reject these hypotheses. For example, after-images of coloured objects are seen in the complementary colour, and this is not the case with the images in the iconic memory (Banks & Barber, 1977). Also, if a subject is shown an image in one eye, and then their attention is drawn to a pattern seen with the other eye, this interferes with the iconic memory associated with the original image (Turvey, 1973) suggesting that the iconic memory is located at a higher level in the visual neurosystem than the point at which the information from the two eyes is brought together. If we knew the precise location of the iconic memory, and understood the chemistry of memory as Kandel does for Aplysia, then it might be possible to explore the relation between phenomenal consciousness and sensory persistence by direct intervention. Jackendorf (1987) and Crick (1994) report evidence suggesting that visual consciousness may be located at an intermediate level in the visual perception hierarchy (see also Rees et al., 2002 and Prinz, 2003). It seems, however, that the information content of a sensory experience can be stored in another short-term memory, described below, in a form different from that of sensory perception. This is shown by by the “partial response” experiments, performed first by Sperling (1960) and originally thought to relate to information in the iconic memory. A similar experiment on the acoustic memory is reported in Moray, 1965 Here, a subject was presented with a random, 3x4 array of letters and numbers for 50ms followed, from 0 to 1 second later, by a high, middling or low tone. Subjects were required to list the digits in the top, middle or bottom row according to the tone following their presentation. Information other than the visual images one would expect to find in iconic memory was found to affect the subjects’ responses: similar sounding letters tended to be mistaken for one another; people sometimes remembered the presence of a particular letter or digit while being unaware of its position; and people could distinguish sets of symbols by category (vowels, numerals) equally well as by position (Merickle, 1980). This suggests the subjects may have been accessing information from the ‘working memory store’, where information had been subject to greater processing and a wider range of information was available. The relevant information is thought often to be stored in acoustic form (in Baddeley’s phonological loop, for example), rather than the visual images of iconic memory. 4. Access consciousness and working memory stores In this paper I use Ned Block’s term ‘access consciousness’ to describe awareness of information other than the images that remain for a brief period in the sensory memory stores. Much of this information is likely to have been recalled from long-term memory, some may be the record of information from a recent sensory experience held in a different form, Alan Baddeley has suggested that visual information may be stored in a speech-based code after the image has faded (Baddeley, 1992) and a wealth of experimental evidence supports this (see Neath and Surprenant, 2003) and some may have been generated by a recent cognitive process, such as a learning session. Psychologists view this accessible information as being held in short-term non-sensory memory stores that are often called ‘working memory stores’, reflecting the fact that the information has to be accessible if cognitive processing is to use it. It is not yet established whether these memories have a ‘buffer’ of available, activated, information not currently “in the spotlight of attention” (as assumed by Baars, 1997). The capacity of the working memory store that comes within the focus of attention seems to be very limited. Psychologists have found that it can hold only seven plus/minus two ‘items’ of information at any point in time (in an experimental setting these may be numerals, “chunked” numerals – 4-digit dates, for example – or short syllables); it seems to be less for visual items where 3 or 4 only is the usual finding. In order to hold information in this memory you need to repeat it, either aloud or silently; this is illustrated by the effort needed to remember the directions someone gives you to find a location, or the difficulty of remembering a telephone number as you cross the room. Without such reinforcement an item of information will remain in this memory for less than a minute at most (see the classic paper, Miller, 1956). It is not yet clear to what extent the information in it is replaced by new information or decays over time. There is a wealth of evidence that suggests that information originally received in visual form is often held in speech-like code in this memory store Perhaps raising a question about differences between humans and other animals (see Neath and Surprenant, 2003, Chapters 3 and 4 for a summary). Baddeley’s ‘working memory’ model is the best known model of this type of memory. Its latest version (Baddeley, 2003) has four parts. It has three storage systems and an organisational element called the ‘central executive’ (see below). Thestorage systems are: the ‘phonological loop’ which holds processed acoustic-type information; the ‘visuospatial sketchpad’ which holds processed visual-type information; and, the episodic buffer which provides temporary storage, that is, a temporary interface between the long-term memory and the other storage systems (this is a relatively new addition). The content of the episodic buffer is available for consiousness, but some of the information in it may not attract the focus of attention. It is the content of the phonological loop and the visuospatial sketchpad that Baddeley now views as being actually attended to (that is, as being in the ‘focus of attention’, in Baars’ terms). 5. Thought consciousness and the central executive Psychologists see the ‘central executive’ bit of the working memory as the part of the mind that is involved in cognitive processing, that is, in the manipulation of information. Alan Baddeley envisages this as a system that “facilitates a range of cognitive activities, such as reasoning, learning and comprehension” (2003, p 829) as well as providing an interface between perception, long-term memory and action (see also Baddeley & Hitch, 1974). If consciousness is “what it is like” to have a certain sort of short-term memory, then the most appropriate candidate to correspond to thought consciousness in the psychologists portfolio of short-term memory stores is Baddeley’s central executive. Baddeley has identified two capacities that are explicitly required by the central executive in the context of the working memory (2003, p 835). These are: the ability to focus attention (that is, to select items for attention from the range of information available); and, the ability to manipulate information between the short-term non-sensory store and the long-term memory. But the processing involved in thinking extends far beyond these limited, but necessary, functions. For example, it requires the ability to identify logical relationships between the items that are focused upon, and it requires the ability to use logical processes to draw conclusions from a collection of related items. If the central executive is to play this key role in cognitive functions, then as well as obtaining information from our external senses and from the short-term non-sensory memory, it needs proprioceptive information (so it can initiate action and movement) and emotional information (so it can assess whether some course of action is likely to increase or decrease welfare) – as Alan Baddeley says: “… it is useful to conceptualise action as ultimately steered by emotion” (2003, p 837). 6. Theoretical considerations In this section I consider the proposition that consciousness is “what it is like” to have a certain sort of short-term memory from a theoretical point of view. A question arises as to whether a creature could be conscious of sensory experiences if it had no memory whatsoever. The creature’s senses would receive information and send it in messages to the brain. But if the creature had no memory for storing sensory information – even for a very brief period – each bit of information would be gone the moment it arrived and no record would be available for the creature’s p-consciousness to access and connect to the information received the following instant. In this case, it’s hard to see that the creature could possibly experience p-consciousness. Massaro’s experiment described above, along with many others (see for example Marcel, 1983) supports this intuition, since interrupting the formation of the record of a signal in the sensory memory (using a ‘mask’) disrupts p-consciousness, resulting in the subjects being unaware of the content of the signal. If this intuition is correct, then some form of memory for sensory information is necessary for p-consciousness. This raises the question: could p-consciousness and a sensory memory be, in some sense, one and the same thing? Could the experience of p-consciousness be “what it is like” to have a sensory memory? But if that were the case then a sensory memory, as well as being necessary for p-consciousness, would also be sufficient for p-consciousness – that is, it would be impossible for a creature to have a sensory memory without also experiencing p-consciousness. To explore the question of sufficiency, consider a hypothetical situation in which a creature has a sensory memory which is not accessible to p-consciousness. Is there any reason why this should not be the case? Could sensory experiences, stored in some sensory memory that is not accessible to any form of consciousness, induce learning and result in the sensory information being stored in an unconscious long-term ‘skills-type’ memory such as that which has been identified in humans by psychologists and neuroscientists (Scoville et al., 1957)? This would both affect a creature’s behaviour and enhance its ability to survive. So we cannot rule out the possibility that a creature might have a sensory memory that is not accessible to p-consciousness. So although the intuitively acceptable necessity of memory for consciousness lends plausibility to the idea that consciousness is “what it is like” to have a certain type of memory, we cannot justify that suggestion by the use of logic alone. Instead, we have to rely on Chalmers’ principle of structural cohesion (1995) combined with the observation that there is a correspondence between the different forms of consciousness and the different types of short-term memory identified by psychologists. The best we can say is that an “inference to the best explanation” (Lipton, 1991), based on these, lends credence to our proposition. References Baars, Bernard J. (1997), ‘In the theatre of consciousness, Global Workspace Theory, a rigorous scientific theory of consciousness’, Journal of Consciousness Studies, 4 (4), pp 292-309. Baars, Bernard J. and Katharine McGovern (1996). ‘Cognitive Views of Consciousness.’ In The Science of Consciousness. Ed: Max Velmans. London and New York. Routledge. Baddeley, Alan D. (1992), ‘Working memory’, Science, 255, pp. 556-59 Baddeley, Alan D. (1996), ‘Exploring the central executive’, The Quarterly Journal of Experimental Psychology, 49A (1), pp. 5-28. Baddeley, Alan D. (1997), Human memory: theory and practice, Psychology Press: Hove and New York. Baddeley, Alan D. (2003), ‘Working memory: looking back and looking forward’, Nature Reviews: Neuroscience, 4, pp. 829-39. Baddeley, Alan D. and Graham J. Hitch (1974), ‘Working memory’, in G. A. Bower (ed.), The Psychology of Learning and Motivation: advances in research and theory, Vol. 8, pp. 47-89, New York: Academic Press. Banks, William, P. and Grayson Barber (1977), ‘Color information in iconic memory’, Psychological Review, 84, pp. 536-46. Block, Ned (1995), ‘On a confusion about a function of consciousness’, Behavioral and Brain Sciences, 18 (2), pp. 227-87. Block, Ned (2005), ‘Two neural correlates of consciousness’, Trends in Cognitive Sciences, 9 (2), pp. 46-52. Burnham, William H. (1888), ‘Memory, historically and experimentally considered’, American Journal of Psychology, 2, 568-622. Chalmers, David J. (1995), ‘Facing up to the problem of consciousness’, Journal of Consciousness Studies, 2, pp 200-19. Cowan, Nelson (1988), ‘Evolving conceptions of memory storage, selective attention, and their mutual constraints within the human information-processing system’, Psychological Bulletin, 104 (2), pp. 163-91 Cowan, Nelson (1997), Attention and Memory, Oxford University Press: Oxford. Crick, Francis (1994), The Astonishing Hypothesis, Simon & Schuster Ltd. Efron, Robert (1970), ‘The minimum duration of a perception’, Neuropsychologia, 8, pp. 57-63 Humphrey, Nicholas (2006), Seeing red: a study in consciousness, Harvard College. Jackendoff, Ray (1987), Consciousness and the Computational Mind, MA: Bradford Books, MIT Press. James, William (1890), Principles of Psychology, (2 vols.), New York: Henry Holt, Chapter XVI “Memory”. Lipton, Peter (1991), Inference to the Best Explanation, Routledge: Oxford. Marcel, A. (1983), ‘Conscious and unconscious perception: experiments on visual masking and word recognition’, Cognitive Psychology, pp197-237 Massaro, Dominic W. (1970), ‘Perceptual auditory images’, Journal of Experimental Psychology, 85 (3), pp. 411-17. Merickle, Philip, M. (1980), ‘Selection form visual persistence by perceptual groups and category membership’, Journal of Experimental Psychology: General, 109 (3), pp. 279-95. Miller, George A. (1956), ‘The magical number seven, plus or minus two: some limits on our capacity for processing information’, Psychological Review, 63, pp. 81-97. Moray, Neville, A. Bates and T. Barnett (1965), ‘Experiments on the four-eared man’, The Journal of the Acoustical Society of America, 38 (2), pp. 196-201. Nagel, Thomas (1974), ‘What is it like to be a bat?’, The Philosophical Review, 88, pp. 435-50. Neath, Ian and Aimée M. Surprenant, (2003), Human Memory (Thomson: Wadsworth). Peacocke, Christopher (2007), ‘Mental action and self-awareness (1)’, in Contemporary Debates in Philosophy of Mind, (eds. Brian P. McCloughlan and Jonathan Cohen), Blackwell Publishing, pp. 358-74. Prinz, Jesse J. (2003), ‘A neurofunctional theory of consciousness’, Philosophy and Neuroscience, A Brook and K Atkins (eds.), Cambridge: Cambridge University Press. Prinz, Jesse J. (2007), ‘All consciousness is perceptual’, , in Contemporary Debates in Philosophy of Mind, (eds. Brian P. McCloughlan and Jonathan Cohen), Blackwell Publishing, pp. 336-57. Rees, Geraint, Gabriel Kreiman and Christof Koch (2002), ‘Neural correlates of consciousness in humans’, Nature Reviews: Neuroscience, 3, pp 261-70. Richet, Charles R. (1884), L’homme et l’intelligence. Richet, Charles R. (1886), ‘Les origines et les modalités de la memoire’, Revue Philosophique, XXI, p. 570. Sakitt, Barbara (1976), ‘Iconic memory’, Psychological Review, 83 (4), pp. 257-76. Scoville, William Beecher and Brenda Milner (1957), ‘Loss of recent memory after bilateral hippocampal lesions’, Journal of Neuropsychiatry and Clinical Neurosciences, 20, pp. 11-21. Sperling, George (1960), ‘The information available in brief visual presentations”, Psychological Monographs, 74, pp. 1-30. Turvey, M T (1973), ‘On peripheral and central processes in vision: inferences from an information-processing analysis of masking with patterned stimuli’, Psychological Review, 80 (1), pp. 1-52. PAGE 5