So many words have been poured unto this subject whatever more can be said? Give humans a topic and over time they will build magnificent formal towers of differentiation, comparatives, back-door arguments and such like, making ever more sophisticated cases for and against derivatives hanging off the merest of bases. Yet simple relevant truths are in danger of being overlooked in such engagements, lost to hubris.
Of course I am well aware of the converse trap, the misguiding unreliability of intuitive judgement and the lulling comfort of straightforward simplicity. Yet two vivid examples of domain masters demonstrating the power of simple truths, by calling upon common sense, spring to my mind: The first is John Searle’s acting demonstration, repeated in so many of his videoed lectures on free will, of wishing his arm to go up and – voila! It does indeed! The second, in the Q&A that followed Basil Hailey’s lecture in TSC2015 he was asked what he thought of the Everettian Many-Worlds quantum mechanics interpretation. Prof. Hailey seemed to be mildly irked by the question, dismissing it with a succinct response: “Where does the energy come from?”.
So what is the problem with the notion of causal closure as a line of reasoning favouring epiphenomenalism? There is an auto-contradictory facet to the arguments, but the problem deserves to be presented in a tangible so-evident form. Here is a hint: Given that there is no colour in Philosophical Zombie world, it is extremely unlikely that zombie culture (if there could be such a thing) would conjure up names for different colours; at least not before they evolve a mature zombie physical science (or neuroscience). Why would they? How could they?
Lets take a close look at a bunch of zombie neurophysiologists studying an anesthetized zombie brain while poor Z’s retina is exposed to red or green shape patterns on variously distinctly coloured backgrounds. The zombie scientists do not quite z-understand what is distinctly unique through and about their stimuli set; they had stumbled upon it’s discovery. Now they have got several multi-electrode probes stuck in their subject’s brain, and from the readings they can infer that there are some shape specific reactive neural populations (in V4), some wavelength specific reactive neural populations (in V1 and higher areas in the visual processing hierarchy), and some spatially-averaged wave-length correlated delta measure (aka “colour” to some humans) reactive populations (in the colour centre, wherever exactly that turns out to be once the debate settles). And if through genetic tuning or training and Baysean learning, specific shape-colour combinations have become associated with particular behavioural reactions then they could get lucky and find traces of these associations in the supplementary motor cortex areas.
Eliciting a reaction A to shape-colour pattern Sh1-Cx vs a reaction B to shape-colour pattern Sh1-Cy would be akin to zombie blind-sight, an incidental discovery which indeed seems to be a trait exhibited by all so-trained (non-anesthetized (?!)) zombies – they are disposed to react with particular affinities. yet nothing in the zombies’ accessible account of the so-called external world, i.e. in an account of their world model as constructible from that which is directly sensed by them, gives them anything to latch-on to, in order to z-explain such preferences. To begin to develop such an account zombie physicists must first figure out what prisms do, or, alternatively, their scientists must study retinal and CNS micro-anatomy and physiology.
The excessive amount of jesting above notwithstanding, I think the point should be clear: This ‘Mary’ theme inspired story illustrates how we lean on qualia (simples, perhaps referents?) as terminal surrogates for the reality out there, regardless of behavioural impact. The irrevocable immediacy of the proxies underpins and catalyses the crystallization of our world model, anchoring raw non-patterned multi-dimensionality in a way that enables us to refer back to it. We can and do tag qualia and we wouldn’t be doing that if consciousness was epi-phenomenal. Conversely – qualia are not tagged on, it is inherent in the (conscious) representation.
Is an evolutionary advantage conferred on agents having qualia? The conscious agent benefits in ways that are readily apparent and in ways that are yet to be understood. Apparent advantages include:
- Accessible information advantage (e.g. as highlighted by the above ‘Mary in the operating theatre’ story).
- Stackable and combine-able, yet inherently differentiated, reflexively multi-dimensional world model. Grounding by phenomenal surrogate does not rely solely on inferred correlations and label tagging.
- Reduces computational demands associated with sensory driven world modelling (as applies e.g. to object detection or inter-pattern correlation for classification purposes; although it remains a challenge to model or express this in verifiable formal notation).
- The reflexive Red quality requires no inference, and retains it’s particular (distinct and resiliant) subjective nature that may never be confused with other coincident concurrent sensations (e.g. the smell of a red rose, the prick of it’s thorn).
- The underlying capacity supporting ‘pop-out’ perceptive capabilities. For some qualities – the first aspect that emerges from pre-attentive processing
(As an aside: Could it be the case that experience is focused by the computational process of cognition, that raw qualia are experienced alongside and prior to final interpretation of patterning? It seems that simple subjective qualities are experienced before it becomes apparent what they are associated with – a pattern, an object, a shape, a fitting into a context; Further-more it seems learning modulates the ability to discern rather than enhance the potential to experience).
In other words, beyond obvious information conservation advantages (i.e. qualia plus labelled lines should convey more information than just labelled lines), in the framework of IMT the manifestation of qualia drives a-priori inclusion of intrinsic information that is packaged and wrapped as qualia; whereas zombie learning of same 1st-base properties must rely on posterior correlation-based clustering and coupled behavior analysis.
- An enabling disposition with regards to the possibility of an inverted iconic computational model, as proposed and outlined in the main pages of this site, that has inter-alia the following characteristics:
- A capacity for true distribution of information, in a way that ‘makes a difference’.
- In combination with (and by) an iconic representation tracing computational process, underpins ‘Binding’ in a physical volume-occupying brain-like processing device (to clarify – the question how does the the brain produce binding, is resolved by the combination of the manifestation of spatiotemporally featureless proto-phenomenality and iconic representation).
- Decouples sensation from overt action, facilitated by the creation of a raw information-poor proto-self.
- Facilitation of the ability to evaluate conjectured scenarios.
- Promotes early percept sensation with evolving accuracy and specificity over (computational processing) time. Supports early (re)action initiation and better control.
- Temporal alignment of perceptual dynamics; bound percepts are synchronous (on a behaviourally relevant temporal order of scale).