Are We Silently Suppressing "Opt Outs"?

A thorn in my brain that has been stuck there for some time has resurfaced recently. Matching Law describes a behavioral principle linking the probability of a particular behavioral option being performed to its relative rate of reinforcement. For example, if:

Behavior 1 is reinforced 75% of the time,

Behavior 2 is reinforced 25% of the time,

Behavior 1 will be approximately 3x more likely to be chosen by the learner than Behavior 2.

Now imagine that a dog is learning to rest their chin on a chair as the opt-in signal for cooperative nail trimming. The dog rests their head, and the human begins the procedure when they lift their head the procedure stops. Perhaps the cessation of the activity provides some concurrent negative reinforcement for the opt-out behavior. Perhaps a treat is also layered in to more equitably reinforce the two competing options. However, as the procedure progresses the handler begins to sense the threshold at which the animal will opt out and ceases the interaction first, preventing the practice of escape and reinforcing liberally for increasing participation in the activity. Soon the animal is no longer opting out, and as such, the relative rate of reinforcement for that option ceases or dips to such a low rate that it calls into question the communicative value of the choices made by the animal in that context. If we imagine this in the context of the above formula:

Opt-in is reinforced 75% of the time,

Opt-out is reinforced 25% of the time,

Opt-in will be approximately 3x more likely to be chosen by the learner than Opt-out.

Now of course this is an overly simplistic representation of a deeply nuanced topic, but I find myself frequently pondering:

• How we can actively bias against our assumptions to determine whether the value in the behavior is communicative, rather than merely another shaped and fleeting piece of their behavioral repertoire shaped by humans for humans?

• Can we pick a better criterion than compliance to measure the degree of comfort or discomfort that the learner experiences?

• Are we leveraging the availability of "choices" only as a stepping stone to reach our own personal goals, and stripping agency from the learner once it has served our purpose?

• How does the recency effect contribute to the perceived availability of an exit strategy to the learner? If it hasn't been practiced recently enough, does it still exist in their mind?

• Does consent offered in response to prompts from the trainer hold the same value as in learner-led interactions?

To me, these questions embody the paradigm shift from a cooperative (instructor-centered) framework toward a collaborative (learner-centered) approach for which we are only just scratching the surface. Do you find yourself pondering on elements of our current practices on this or any other topic? How can we critically reflect on our application of training and behavior modification strategies to find potential areas that may warrant innovation as we reach for the next iteration of training standards?