22/12/2025
Food for thought — SHARE & LIKE
I spend a lot of time reading scientific literature on dog behaviour, cognition, and learning theory.
Recently, I had a “debate” with a well-known e-collar advocate that ended when he blocked me, not because of insults (although he was being personal), but because the logic being challenged couldn’t be answered.
This highlights a bigger issue that keeps appearing in both training discussions and some scientific interpretations.
For a dog to be said to have learned avoidance of a stimulus (for example, sheep), all other controlling variables must be removed.
Yet in e-collar livestock arguments, this almost never happens.
👉 The reinforcers are never removed
👉 The handler is never removed
👉 The context is never removed
If a dog has truly learned to avoid sheep, then the behaviour should persist after the permanent removal of the collar, the handler, and the original training context.
That standard is has never been demonstrated — would they say because it would be unethical? If so, then the whole e-collar over long line debate would end abruptly.
If we accept the claim that learning does not require the continued presence of reinforcement, then the same logic must apply in reverse:
I should be able to remove myself and food entirely and still get reliable behaviour — which we all know is false. Learning is always expressed under stimulus control.
When we apply the same standard consistently, the conclusion becomes unavoidable:
E-collars do not demonstrate true avoidance learning.
What they demonstrate is supervised inhibition under threat.
This matters because people like Danny, Ivan, and James argue that e-collars are necessary because sheep are killed when dogs escape and owners aren’t present.
But without evidence that behaviour persists without the collar, without the handler, and outside the original context, that claim is unfalsified.
If those elements must remain in place for “success,” then the argument for an e-collar over a long line collapses.
At that point, what’s being sold as rehabilitation is actually risk management via permanent threat availability — not learning.
This isn’t about emotion, ideology, or personal attacks.
It’s about applying the same scientific standard to punishment that we already apply to reinforcement.
Suppression is not the same thing as learning.
