This is the fourth part of the talk ‘Free Will / Free Won’t’. Notes in the text in the form [ABC] can be found by following the above link.
17. Tamar Gendler’s concept of ‘Alief’
‘Alief’ is a word concocted by Tamar Gendler for our low-level (subconscious) beliefs. She provides one particularly striking example of alief in action. Since 2007, people have been able to visit the Skywalk at the Grand Canyon, a horseshoe-shaped viewing platform that extends over the edge of the canyon, with a 4000-foot drop below. What’s more:
- The sides are made of toughened glass, and
- The floor is made of toughened glass!
It takes many people quite some willpower to walk on it. Even though they believe the walkway to be strong enough and not a risk, their lower-level ‘aliefs’ are screaming ‘don’t go there!’ Some people just freeze, unable to move forward or back. But there are some people who find themselves in the same situation in far less extreme situations. Their aliefs and beliefs are in competition but their beliefs are unable to override their aliefs.
18. Marvin Minsky’s ‘B-Brain’
A significant inspiration for Rodney Brook’s ‘fast, cheap and out of control’ robots came from the work of Marvin Minsky. In the very accessible book ‘The Society of Mind’, Minsky builds up a picture of intelligence from the interaction of lots of simple ‘agents’ – hence the term ‘Society of Mind’ (For Minsky, ‘mind’ is just what the brain does). One of the 270 essays in the book is about ‘B-brains’.
Imagine a brain that consists of two parts, A and B. What he calls the ‘A-brain’ has inputs and outputs connected up to the real world. This can react to events in the real world; it ‘thinks’ in the real world. The ‘B-brain’, by contrast, has no direct connections to the real world; it only has connections into the A-brain. It ‘thinks’ in the domain of the A-brain – the A-brain is the B-brain’s world! The B-brain’s job is to ‘correct’ behaviour in the A-brain, without any actual `understanding’ of the A-brain’s thinking. For example, if it finds the A-brain is stuck repeatedly trying to solve a problem in the same way, it might give it a jolt to get it to try something different. The B-brain is providing the organism with the freedom to avoid getting stuck in the A-brain’s mental ruts. The B-brain notion is obviously similar to that of ‘free won’t’: The B-brain is monitoring the A-brain and able to intervene (‘veto’) when it thinks it is appropriate.
Minsky splits the mind into 6 ‘levels of reflection’, similar to the (arbitrarily chosen) 5 levels shown in Figure 3 in Part II of the talk. Minsky’s levels are:
- instinctive reactions
- learned reactions
- deliberative thinking
- reflective reactions
- self-reflective reactions
- self-conscious reactions
Higher levels add to the intelligence of the overall mind but in themselves are not necessarily more complex. We have traditionally underestimated the complexity of the ‘base’ animal behaviour of being engaged in the environment and overestimated our higher-level reasoning capabilities. Unlike Minsky’s examples, a particular part of my ‘free won’t’ argument is that the higher levels can be more complex (and hence slower) than the lower-level functions and it is this variation in response times of the different levels that provide us with freedom at different timescales.
As far as I know, Gendler’s notion of alief and belief has made no connection with Minsky’s A- and B-brains. But there is an obvious parallel to be made here. The A-brain has aliefs; the B-brain has beliefs.
At the end of Minsky’s essay, he makes the obvious speculation that the B-brain could also have a higher-level monitor – ‘the C-brain’.
19. C-lief: Free-Won’tifying Ourselves
Recap: The basic concept of ‘Free Won’t’ from an evolutionary perspective is to gain an extra layer of ‘thinking’ hardware. This extra layer is typically larger than the existing layers but, as a consequence, slower. It’s ‘output’ action may override those of the lower levels but, because it is slower, this may be seen as the ‘exercise of a veto’.
Can we create a ‘thinking layer’ higher than that in our brains? Obviously, evolution works on a very large timeframe, much greater than our lifespans. But can we ‘free-won’t-ify’ ourselves? If we crudely map our unconcious aliefs to Minsky’s A-brain and our conscious beliefs to Minsky’s B-brain, what might the C-brain be? And what are its ‘C-liefs’? (If Tamar Gendler can concoct new words in the English language then so can I! ‘C-lief’ and ‘free-won’t-ify’ are deliberatively ungainly terms but are created only half in jest.)
I previously talked about building an extra layer of intelligence around ourselves in ‘The Extension of Mind’ which covered the ‘Extended Mind’ thesis of Andy Clark. This thesis was particularly persuasive about restoring the faculties of the disabled – external eyes for the blind; external memory for those with Alzheimers. But, unlike higher levels of Free Won’t, there is no veto here. It is still a conscious choice to use these extensions rather than having external objects overriding our actions. This is ‘Intelligence Amplification’ (and ‘Action Amplification’) rather than an extension of intelligence (and therefore, freedom).
Examples where machines can veto our actions are rare and are confined to safety mechanisms that prevent us from doing things, for example:
- Automatic braking systems now being introduced into new cars.
- Anti-stall technology on airliners. A pilot might pull back the stick in an emergency but the autopilot overrides the climb in order to prevent a stall.
Such safety mechanisms can lead to ‘automation complacency’ on the part of the driver/pilot – a willingness to disregard the need for attention or take risks, which then relies on the (non-perfect) ability of the machine to get us out of trouble.
An everyday example of machines interrupting our action is peoples’ enslavement to their mobile phones – they must answer if it rings. But it is also another example of our willingness to give up control to machines.
An alternative take on the C-brain is to equate it with society (‘C’ for Community?). We can view the ‘collective thoughts’ of society organization (a company, a family, a society, society as a whole?) as an additional layer. We may tend to think of an organization (particularly, companies) as being one where high-level commands emanate from the CEO and trickle down through the organization. But such a top-down organization where most action originated from on-high would be a stifling bureaucracy (one where the CEO was a sort of homunculus of the organization, and sole bearer of responsibility). A more dynamic company (and more typical in my experience) would be one organized on bottom-up ‘free won’t’ principles: the low levels of the organizations are `empowered’ to make their own decisions but occasionally there is a need for higher levels to step in and veto actions. Senior management are largely involved after the event, operating generally in a reactive manner. But this isn’t really an example of ‘free-won’tification’ in the sense of building a higher level on top of a single brain. It is extending intelligence by creating a ‘Society of Minds’ rather than a (Minskian) ‘Society of Mind’ [SUR].
20. An Alternative ‘Conscious Will’
What is the point I am trying to make with these ideas of aliefs and C-brains? It is that our conscious beliefs are in competition with both higher and lower level forces. From our standpoint, action is a consequence of:
- ‘UNC’: unconscious aliefs,
- ‘CON’: conscious beliefs and
- ‘SUP’: super-conscious C-liefs.
as shown in Figure 5. Actions resulting from our conscious consideration are in a battle of wills with our subconscious and the rest of society. Note: ‘super-conscious’ here just means ‘above consciousness’. I’m not making any claims here to higher levels of consciousness.
Figure 5: Alief, Belief and C-lief: a Battle of Wills
Note that ideas at one level can move to another:
- Our brains have enormous plasticity. Beliefs can seep into our aliefs: the conscious learning of a new skill eventually becomes instinctive after enough practice.
- Similarly, C-liefs can seep into beliefs: we can become indoctrinated by society.
Those who are unable to override their aliefs (the strength of will of their aliefs are stronger than that of their beliefs) will not be able to walk the Grand Canyon Skywalk or perform other ‘mind over matter’ personal victories. Similarly, those who are unable to use the strength of will of their beliefs to override the strength of will of society will find themselves powerless. It would seem natural to call this strength of will of beliefs ‘conscious will’: that will that arises at a conscious level (‘CW’ in the figure) as opposed to unconscious (‘UW’) or superconscious (‘SW’).
Wegner’s term ‘conscious will’ [WEG] is good for its emphasis on consciousness and may be a good account of our experience, but it does not seem to involve ‘will’ – a force or mental strength, in the way that ‘conscious will’ as presented above does.
[This is the end of Part IV. Follow the link at the top of this blog entry for the next part.]