20/02/2026
๐๐๐๐
https://www.facebook.com/share/p/1AddB9ZSSW/
Food is not just a โwageโ for behaviour.
Yes, we can use treats operantly. Behaviour happens, consequence follows, and behaviour becomes more likely in that context.
Yes, we can use treats respondently. Trigger appears, cheese appears, and over time we change what the stimulus signals and what the learnerโs body prepares for.
The two processes are not mutually exclusive.
But the reason food can be so useful is that it lets us design experience, and experience is what updates the learnerโs model of the world.
Letโs explore this a bit.
1) Operant is not just reinforcement. It is allocation.
In the real world, the learner is not choosing between โsitโ and โdo nothing.โ They are allocating behaviour across options:
- Orienting to the environment
- Scanning, sniffing, moving away
- Checking in, disengaging, reorienting
- Barking, lunging, freezing, fidgeting
- Choosing proximity or choosing distance
Food helps because it lets us build a reinforcement system where staying is possible, recovery is reinforced, engagement is valuable, and the learner has a predictable way to succeed.
2) Prediction error is the update signal
If you want one technical concept that explains most of this, it is prediction error. Prediction error is simply what the learner expected, minus what actually happened.
When โthat dogโ, โthat personโ, โthat noiseโ reliably predicts uncertainty or aversives, the learnerโs model updates in the wrong direction. Hypervigilance becomes sensible.
Food can be powerful because it can repeatedly create moments where the learner thinks: โI expected worse,โ โthat was safe,โ โsometimes it is even good!โ
3) Incentive salience: food changes what feels worth doing
Food is not just pleasant. It has motivational gravity. It can pull attention, support approach behaviour, and make engagement with their caregiver feel like a coherent option. That matters in triggering contexts because attention is not necessarily the rational choice when a dog is in a survival system. Food can help the learner remain organised enough to notice, assess, and recover, instead of spiralling into scanning and escalation.
4) Not all change is associative
Habituation is a reduced response to repeated exposure that stays within tolerance. Sensitisation is the opposite, when exposure is too intense or too long and the response grows. Treats are not the mechanism of habituation, but they can support the conditions that make habituation more likely. Exposures are brief. Intensity stays under threshold. Recovery happens quickly. The learner has something predictable to do. Opting out is easy and safe.
5) These systems stack, and stacking is the point
In a well-designed moment, several things can be true at once:
- Operant: behaviour is reinforced, especially recovery and disengagement
- Respondent: the triggerโs meaning shifts toward safety
- Motivational: attention and approach become easier
- Non-associative: repetition under threshold supports desensitisation and/or habituation
6) Learning is often context-bound
If the street corner has a history of unpredictability, it will carry emotional weight.
A reinforcement system that consistently creates safety and success can start to change not only the trigger, but the feel of the whole context.
7) Food is not automatically ethical just because it is โpositive!โ
Used well, it is often the least intrusive option available. I will absolutely use food to reinforce trained operant skills in order to create safe movement and clean choices in situations where the alternative might be tighter leashes, physical steering, or escalating management. Sometimes a treat is the gentlest lever in the room.
But use mindfully, because food can also slide into coercion when the learner cannot realistically opt out, the set-up is too close, too long, or too intense, when food is being used to keep them in a situation they would otherwise leave or when the โcost of noโ is high (no access to distance, no exit, no break)
Treats should usually be working WITH the functional reinforcer, not competing against it.
If the learner is barking and lunging because distance is what they need/want, then my food delivery should support behaviour that earns distance safely. If the learner is worried and trying to create relief, then food can help organise behaviour, but relief still has to show up in the system.
If you want lasting change, understanding function is not optional. Otherwise you are just trying to outbid the environment forever.
8.) If food is doing all the work, the system is fragile
If the behaviour only happens when food is visible, it usually means one of two things:
a) the behaviour has not been reinforced enough yet in that context, or
b) the behaviour is not contacting enough functional reinforcement in real life.
So the goal is to make sure the behaviour starts being reinforced by other outcomes too, including:
- Distance and relief (especially in trigger contexts)
- Access to sniffing, moving, social space, or rest
- Predictability and control
- Successful completion of the task itself
Food can be the bridge that gets you there, but it should not be the only load-bearing wall.
9) Treats should make the world bigger, not smaller
If my reinforcement system is working, the learnerโs options expand over time. There are more contexts where the skill is available, more flexibility in what counts as reinforcement, less friction around triggers, and more real-life reinforcement maintaining the behaviour.
If the options are shrinking, if we are paying more and more to get less and less, or if the learner looks like they are enduring rather than choosing, it means the set-up needs changing, not that the behaviour needs โmore proofingโ.
So yes, treats reinforce behaviour. But they also:
- Reduce uncertainty
- Support regulation
- Keep exposure within tolerance
- Update predictions
- Make safe choices genuinely available
But look, treats (or toys, or any other extrinsic reinforcement) are not the point. The point is the learnerโs updated expectation: what happens next here, what options are available, and what they can do that reliably works.
Want to learn more? Join me at the IAABC Foundation conference for my talk on โThe Reinforcement Economy.โ
(You can find the schedule here: https://iaabcfoundation.org/animal-behavior-conference/schedule-of-events/)