« Home

Home » Psychology » Rose: Consciousness » Student resources » Updates » December 2006 » Chapter 11

Rose: Consciousness

Chapter 11

It’s easy to draw up lists of putative types of representation, but two major attempts to add a novel type are worth knowing about. These make an interesting comparison with one another as well as providing technical ways of justifying the classification scheme.1 Both start with Anscombe’s (1957) point that a shopping list could be a description of what has been bought at the shops or it could be a prescription or directive as to what should be bought. The same representation can have two different types of intentionality, depending on it usage. Are there such entities in the mind? Jacob and Jeannerod (2003) argue that ‘visuomotor’ representations exist which have just such hybrid functionality. Millikan (2004) does too, calling them ‘pushmi-pullyu’ representations (see the web update to Chapter 5 for the evolutionary argument behind her view). These at once describe the world as it is here and now and guide current action within it. They may be related to neural processing along the dorsal stream (section 7.2.2.3, Side-box 8.2; Jacob and Jeannerod, 1993, ch. 6; but cf. Millikan, 2004, pp. 176-179).

Millikan differentiates her hybrids from the two ‘pure’ types of representation, descriptives and directives, which in advanced creatures necessarily also exist alongside the older type. These two have contrasting functions which are well known. For example, recall Miller et al.’s (1960) division between images of how the world is and plans as to how to change it (Consciousness, Side-box 8.1). The distinction between beliefs and desires is similarly intuitive, and Jeannerod (1997) drew a distinction between semantic and pragmatic processing. So far these dichotomies all seem consistent, giving us three types (two thoroughbreds and one hybrid).

Further thought however reveals that there must be numerous types of pure breed. First, consider Searle’s (1983) characterisation of beliefs as entities that are supposed to conform to the way the world is. To be true they should somehow ‘fit’ what they represent, i.e. their content. Desires however are the opposite: in these, the world is supposed to fit with the content of the mental representation. (The hybrids thus have simultaneously dual direction of fit: Jacob and Jeannerod, 2003, p. 41.) But second, he noted that the causation between world and mental state may or may not be immediate. Indeed, in the case of beliefs and desires, there need be no such connection. To make this clearer, contrast these states with ongoing perception and intent (motor guidance) functions, which also differ from each other in their respective directions of fit but which are more intimately related to reality. Thus perception (and memory) entail that the presence of the mental state is caused by whatever it is that the state represents (in its own contents). For instance the percept of a red triangle should be caused by a red triangle, whereas the belief you are looking at a red triangle need not be — it is still a belief even if there is no red triangle present. The same applies for causal mind-to-world fitting. Thus intents and motor guidance require that the action be caused by the mental state representing that action, whereas desires do not have to lead to action or fulfilment in order to exist and to be desires.

Searle thus formulated a 2x2 classification scheme, in which direction of fit determines one axis (‘cognition’ versus ‘volition’, i.e. world-to-mind fit versus mind-to-world) and causation the other axis (present or not). For a summary see Searle (2004, pp. 118-120). But where are the hybrids? Jacob and Jeannerod (2003, esp. pp. 193-208) argue these exist (‘visuomotor representations’) alongside pure motor intentions, percepts, thoughts/concepts, desires and beliefs — making 6 types in all!2 Their scheme does not admit so readily of a systematic classification by criteria as does Searle’s but they provide parameters by which some of the types may be dichotomised, such as their content coding coordinates (egocentric versus allocentric), informational encapsulation (present or not) and the ability to re-identify or track the same object over time and space.3

Section 11.3. Is the self special?

At the end of 2005, the journal Consciousness and Cognition published a special issue on ‘The Brain and its Self’. Seven interesting articles are preceded by a useful overview of the field by the editors, Boyer, Robbins and Jack. In this, they focus on self-representation rather than the ontology of the self, though in the end they hint at a constructivist solution, with self emerging from interactions between scientifically-derived meta-knowledge, its social dispersal as revised 'folk psychology', and our individual understanding of the concept of self. Meanwhile, they interpret the self as comprising a multitude of aspects or domains (cf. bundle theory: Consciousness, section 11.3.4.4). Indeed, their analysis can be seen as an example of homuncular functionalism (Chapter 4): the self is a hierarchy, comprising (at level n-1) modules such as cognitive, social, physical and moral 'selves'. Each of these can be broken down further: for example social skills include empathy and mind-reading, cognitive awareness includes autobiographical memory which itself consists of episodic and semantic (self-related) facts (see also section 9.2), and the sense of agency and motor control depends on mechanisms for both forward modelling and internal feedback (see also section 8.4). These modules are not encapsulated however (cf. section 4.3), but overlap at least partially with other mechanisms whose function is to represent other people in general. Finally, an evolutionary justification is given for this piecemeal basis of self-awareness (cf. section 4.2.6).

Several reviews on the brain regions involved in self-awareness have also appeared lately. In a major survey of brain imaging studies, Northoff et al. (2006) derive conclusions that, like Boyer et al.’s (2005), can be summarized in hierarchical terms. First, they suggest we can simplify discussion of the multifarious aspects of the self (narrative, bodily, autobiographical, emotional, spatial, facial and social) as having a common theme, namely that of self-related stimulus processing. Their paper therefore concerns which brain areas are relevant to the subjective experience of stimuli associated with the self, leaving aside the metaphysical nature of that self (although the latter is discussed in their paper). The self-related functions are closedly tied in with the emotional aspects of experience, including feelings of possessiveness and familiarity, in contrast to other more rationalistic cognitive functions.

Northoff et al.’s argument is that medial cortical structures are central to these self-relating functions. These can be contrasted firstly with sensory and subcortical centres for extero- and interoception, which generate a basic ‘proto-self’, and secondly with lateral cortex, which deals with ‘higher order’ cognition.

Within the medial cortical structures are three main clusters of relevant activity. The first is in the ventro-medial prefrontal cortex, which connects with subcortical (mesolimbic) interoceptive regions and with exteroceptive sensory cortices. It perhaps “codes” or represents just how related stimuli are to the self. The second cluster is in dorso-medial prefrontal cortex and deals with the social context of the stimulus. This region connects to the lateral prefrontal lobe, where self-relevant information, and only such information, is subject to further, ‘higher-order’ analysis and executive usages.4 Third is posterior medial cortex, including medial parietal lobe, which processes the spatio-temporal context. This area mediates the self-relatedness of stimulus and body location, and the egocentric point of view, connecting as it does with lateral parietal lobe. It also deals with autobiographical memory via links with the hippocampal region.5

Critique

Despite being repetitive, unclearly written and speculative, Northoff et al.’s paper shows appropriate concern at the lack of crucial evidence on many of the points raised and suggests where further research is needed. For instance it is not clear how their scheme relates to other dimensions. First, having argued that the self-representation is related to the affective component of experience, they briefly mention that there are hemisphere differences in affect. (See Gillihan and Farah, 2005, for a review of how inconsistent the findings are about hemisphere differences in self-representation. See also sections 7.2.2.3, 11.2.3 and 11.3.2.2.) Second, an anterior-posterior dichotomy has been proposed between externally-focussed and internally-focussed processing (Goldberg et al., 2006; Malach and Josipovic, 2006).5

Indeed, Gillihan and Farah (2005) published a detailed methodological critique of a large number of brain imaging studies of self-representation. The data presented so far are reprehensible in their inconsistency. This lamentable situation results because the different studies used a diversity of definitions of ‘self’, omitted controls for differentiating self from non-self processing, and used a variety of response measures, modalities of stimulus presentation and task demands. Gillihan and Farah conclude that defining the self as one’s own body is probably the only valid notion. On the other hand they admit they are investigating the hypothesis that the self is both psychologically and neurally isolable. They accept that the concept of a unitary and special self may be valid at a subjective level; however they reveal their bias by citing the mysterian philosopher Nagel (see Consciousness section 2.4.3.2) and by continuing with the blatantly dualist statements that ‘Our vivid awareness of a self, like awareness more generally, may not be explicable in terms of the mechanistic workings of the brain.’ but that the cognitive and brain sciences can be useful ‘insofar as self-representation is interpreted in information-processing terms and not phenomenological terms.’ (p. 95).

Section 11.3.2.3 Multi-level models

The role of ‘mirror neurons’ continues to be discussed (see also section 6.2.2.1), with Gazzola et al. (2006) arguing for their importance in interpersonal empathy, and Arbib (2006) and Aziz-Zadeh et al. (2006) supporting their role in the evolution of language (but for a review of alternative models see Aboitiz et al., 2006).

Section 11.5 Conclusion

The view I was moving towards in this section of Consciousness resembles in some ways the theory of ‘higher-order global states’ developed by Van Gulick (2004). In his theory, consciousness arises from the combination of lower-order states into a unified holistic system. This avoids the several problems still besetting the HOP and HOT theories (such as why consciousness of an external object should arise when and only when one mental state represents the contents of another state that represents the object).

Van Gulick argues that the phenomenology we are trying to explain involves awareness of (for example) an object from our own point of view, and as such the self is implicitly built into all our conscious experiences. An explicit model of the self does not need to exist (cf. section 11.3 and its web update above) and instead Van Gulick adopts a more Dennettian stance towards the self as a ‘virtual entity’ (section 11.3.4.3). Conscious awareness of an external object arises when the representation of that object is integrated into an organized, dynamic pattern of mutually reinforcing representational states, physically distributed around the brain. The content of the lower-order state is defined by its functional effects (section 2.5) and is therefore changed when it becomes part of the global state (cf. top-down causation: see Side-box 4.4, p. 110), since it now has a wider influence (section 11.4). Thus the lower-order state’s properties are altered by its new context — it now exists within a wider system — and its content becomes part of consciousness in virtue of that enlargenment of its functional role. But the change in its properties involves not just an increase in efficacy and intentionality but also a more complex and structured (“thicker”) phenomenology — simultaneously of oneself and the world at a particular place and time — engendered by the unification of multiple lower-order functional states (cf. section 10.2.3.2).

Van Gulick does not explicitly describe higher-order global states as ontological emergents. Nor does he generalise to systems composed of more than one level of integration: for example he doesn’t include a social/linguistic level on top of the raw experiential level, merely pointing out that non-linguistic organisms such as animals and infants can have sensory experiences (cf. section 7.2.2.3); so the higher-order state need not be entirely global. Van Gulick’s argument nevertheless has the precision and detail necessary to make it a useful development in the project to explain the origins of consciousness.

Recommended Reading

Two collections of essays on the higher-order perception versus thought debate have appeared (Carruthers, 2005; Rosenthal, 2005).

A comprehensive and readable review of theories of schizophrenia (cf. Side-box 11.1) has been provided by Williamson (2006).

New references

Aboitiz, F., Garcia, R.R., Bosman, C. and Brunetti, E. (2006) Cortical memory mechanisms and language origins. Brain and Language 98, 40-56.

Amodio, D.M. and Frith, C.D. (2006) Meeting of minds: the medial frontal cortex and social cognition. Nature Reviews Neuroscience 7, 268-277.

Anscombe, G.E. (1957) Intentionality. Blackwell, Oxford.

Arbib, M.A. (2006) ed. Action to Language via the Mirror Neuron System. Cambridge University Press, Cambridge.

Aziz-Zadeh, L., Wilson, S.M., Rizzolatti, G. and Iacobini, M. (2006) Congruent embodied representations for visually presented actions and linguistic phrases describing actions. Current Biology 16, 1818-1823.

Boyer, P., Robbins, P. and Jack, A.I. (2005) Varieties of self-systems worth having. Consciousness and Cognition 14, 647-660.

Carruthers, P. (2005) Consciousness: Essays from a higer-order perspective. Oxford University Press, Oxford.

Gazzola, V., Aziz-Zadeh, L. and Keysers, C. (2006) Empathy and the somatotopic auditory mirror system in humans Current Biology 16, 1824-1829.

Gillihan, S.J. and Farah, M.J. (2005) Is self special? A critical review of evidence from experimental psychology and cognitive neuroscience. Psychological Bulletin 131, 76-97.

Goldberg, I., Harel, M. and Malach, R. (2006) When the brain loses its self: prefrontal inactivation during sensorimotor processing. Neuron 50, 329-339.

Jeannerod (1997) The Cognitive Neuroscience of Action. Blackwell, Oxford.

Kriegel, U. and Williford, K. (2006) eds. Self-representational Approaches to Consciousness. MIT Press, Cambridge MA.

Lieberman, M.D., Jarcho, J.M. and Satpute, A.B. (2004) Evidence-based and intuition-based self-knowledge: an FMRI study. Journal of Personality and Social Psychology 87, 421-435.

Malach, R. and Josipovic, Z. (2006) Perception without a perceiver. Journal of Consciousness Studies 13 (9), 57-66.

Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Dobrowolny, H. and Panksepp, J. (2006) Self-referential processing in our brain — A meta-analysis of imaging studies on the self. Neuroimage 31, 440-457.

Rosenthal, D.M. (2005) Consciousness and Mind. Oxford University Press, New York.

Searle, J. (1983) Intentionality. Cambridge University Press, Cambridge.

Searle, J.R. (2004) Mind: a brief introduction. Oxford University Press, New York.

Van Gulick, R. (2004) Higher-order global states (HOGS). An alternative higher-order model of consciousness. In Gennaro, R.J. (ed.) Higher-order Theories of Consciousness. Benjamins, Amsterdam, pp. 67-92.

Williamson, P. (2006) Mind, Brain, and Schizophrenia. Oxford University Press, Oxford.

Footnotes

1. They cite each other, so contain mutual criticisms too.

2. Both Jacob and Jeannerod (2003) and Millikan (2004) also add indexicals (e.g. ‘here/there’, ‘now/then’, ‘I/you/he/she’ and verb tenses), and demonstratives (‘this/that’), but I will pass those over for now! Millikan (ch. 15) also discusses time and space representation.

3. This last quality is explained with more generality by Millikan (2004, chs. 3 and 4). The intentionality of a representation derived from a natural sign depends on the ability to track that sign as it recurs within a particular ‘domain’ (range of spatio-temporal contexts, or channel) where its relationship to what is signified remains constant. You have to know the limits of the domain as well as the sign-to-signified relationship, since meaning is context-relative and can change if the situation alters. (She contrasts this with Dretske’s definition of information-bearing signs which, she says, makes them context-free and thus meaningless. This discussion revisits their disagreement which I introduced in Consciousness section 5.2.5.)

4. But Amodio and Frith (2006) place the social context area further forward towards the anterior pole, and instead link Northoff et al.’s second area to decision making and the choice of actions (see also section 8.4, p. 282).

5. In an interesting aside, Northoff et al. (2006) compare their ideas with those of Lieberman et al. (2004), who contrast knowledge of the self based on ‘intuition’ (involving a system composed of ventro-medial prefrontal cortex, amygdala and nucleus accumbens) with self-knowledge based on ‘evidence’ (involving lateral prefrontal and posterior parietal cortices and hippocampus). This duality of input to the self perhaps maps onto Northoff et al.’s model, in which proto-self and higher-order cognition sandwich the self-representation which is an interface between them.

6. This latter work is particularly interesting because it emphasises that when people are focussed on external stimuli they actually lose self-awareness, thus calling into question theories that consciousness arises only when the self-representation is activated (see Consciousness sections 6.2.2.1, 8.3.2 and 11.2.2.3 pp. 372-373, as well as 11.3). For further reading see Kriegel and Williford (2006).