Blog Coordinator

« How to Escape Kim’s Logical Headlock | Main | Arvan Wins 3QD "Charm" Prize »

12/22/2014

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

Peter, thanks for your many wonderful posts this month.

The roadblock that I’m running into regarding your theory of criterial causation, is about exerting control.

You stated above: “information is inherently decoder-specific”. I agree with you on that, and the result is: the decoder is the entity which exerts “control” based upon its interpretation of information/patterns.

But what is the definition of “control” that we’re using?

Just like the term “free will”, there seems to be controversy about the meaning of the word “control”. Perhaps we could reach agreement regarding the meanings of the phrases “control in the weak sense” and “control in the strong sense”.

In my opinion, here’s an example of “control in the weak sense”: A cue ball rolls directly into an 8 ball, thereby sinking the 8 ball. I could say that the cue ball exerted control on the 8 ball, and there wouldn’t be anything wrong with using language that way. But I’m simply referring to “control in the weak sense”, since all of the control was actually from the laws of physics (along with some randomness).

In order for “control in the strong sense” to exist, I’m thinking that an event must be controlled in a manner that’s not solely from the laws of physics (or solely random in nature). Hopefully, that’s a reasonable meaning for the phrase “control in the strong sense”.

So for me, your theory of criterial causation boils down to the following question: When you say that a group of neurons are configured in a certain manner (i.e., thereby creating criteria) and those neurons are only open to certain patterns of input (i.e., thereby forming a constraint), where is the event wherein “control in the strong sense” occurs? In your model, I’m only seeing events that are controlled solely by the laws of physics (or said events are random in nature), thereby causing me to believe your model is predeterministic in nature (or else random). That’s why I doubt criterial causation is the source of ambitious/strong sense free will. (I believe that FW in the “strong/ambitious sense” requires control in the “strong sense”.)

I’m heading off to Loblolly beach later today and I’m unsure if I’ll have Internet access (it’s out in the boonies). Happy Holidays.

Hi James, I have not talked about control in the strong sense yet this month, but do in my book. Strong control comes from executive circuits (the dorsolateral-posterior parietal mental workspace circuits; the anterior cingulate-opercular task-maintenance circuits) that can plan, imagine, deliberate and make decisions in light of highest level demands and needs. Strong control concretely comes from their capacity to 'rewire' circuits, so to speak, by changing synaptic weights, to fulfill their executive ends. So, for example, operations in the mental workspace (see Schlegel et al. 2013 for details) might reach a conclusion 'need to recall someone who has red hair,' which can them be implemented by lower level neural circuitry, such that, within half a second, Ron Weasely comes to mind. So the weak control operating at the level of, say, recall circuits in the hippocampus, serves the needs and commands of circuits that afford strong control, because the latter can set the former to fulfill whatever criteria they demand.

Does this meet your conditions for strong control?

Peter, if the physical events you've described aren't controlled solely by the laws of physics (and said events aren't random in nature), then I'd agree with you that there's control exerted in the strong sense. Otherwise, I'd say there's only control in the weak sense.

Do you believe that some of the physical events which occur in brains aren't controlled solely by the laws of physics (and said events aren't random in nature)? I believe that's the case, and I'd be thrilled if you agreed with me. (I'm typing this comment *very* slowly on my wife's small phone in a Houston airport.)

James, many possible outcomes are consistent with the laws of physics if quantum mechanical laws are correct. For example, whether a photon passes through a double slit and ends up at position 1 or position 2 on a photographic plate is not determined. Either outcome could happen, and if either did happen, both outcomes would be consistent with all the known laws of physics. Energy would be conserved in either case, for example.

Even though all events in the brain are physical events that obey physical laws, the fact that there are multiple possible outcomes at the particle level means that there are multiple possible outcomes open at the macroscopic level if microscopic randomness can be amplified to a level where it matters in brain function, which is a level of neural spike timing. In my book I zoom in on how exactly quantum domain randomness, say at the level of the single magnesium atom blocking one NMDA receptor, can be amplified up to a level of randomness in a neuron's spike timing.

The strong control you want is not afforded by some ghostly soul, or violation of the laws of physics, or information acting as a supernatural force. Information is causal by virtue of parameters that dictate which subclass of neural inputs can drive a neuron. This effectively delimits a subclass of physical causal chains from all those that are possible as the subset that meets these parameters. These parameters are not only physical parameters, they are set up as informational parameters (that is, the informational parameters or criteria for firing are realized in the physical parameters or physical criteria that have to be met for the neuron to fire). If that is the case, then informational parameters delimit informational causal chains as a special subclass of possible physical causal chains.

In a nutshell, information is not causal as a force. Its causal power lies in its capacity to constrain or delimit which subset of possible physical causal chains can become real. Any that do become real (i.e. that happen) are still physical causal chains that obey all the laws of physics. But they are also informational causal chains by virtue of having met the informational criteria realized in physical criteria for firing of neurons.

Nothing I am saying involves violations of the laws of physics. However, the neural system we have effectively 'sculpts' informational (though nonetheless still physical) causal chains out from among all the physical chains that are possible at the particle level.

The traditional view of causation is rooted in a kind of 'yang' conception, where things happen because of forces and the transfer of energy. Even the very word 'force' carries quite explicitly the meaning of force or coercion. This is an incomplete and for many situations, such as the one we find operating among neurons, even a wrong view of causation. In addition we need to add at least the 'yin' conception of causation via constraints placed on what can possibly happen.

Hope this helps, and enjoy the holidays.

The comments to this entry are closed.

Books about Agency


3QD Prize 2014: Marcus Arvan