How training actually works

Ravi says “sit.” Mabel sits. Ravi reaches into a pocket, struggles with a bag, drops a treat, and finally offers it as Mabel jumps toward his hand. The household concludes that Mabel knows how to sit but prefers jumping.

Mabel may be learning from a sequence quite different from the one Ravi intended. The useful moment and the useful consequence have become separated by several other events. Training improves when the person makes the relationship easier to detect.

This chapter explains the structure of a teaching attempt: the setting before the behavior, the behavior itself, and what follows. It introduces rewards, markers, shaping, cues, and generalization, then uses an original target exercise to connect them. The next chapter applies the same principles to everyday skills.

Define the behavior you want to see

“Be good” is not an observable task. “Keep four paws on the floor while I attach the leash” is more specific. “Stop being annoying” does not identify an alternative; “rest on this comfortable mat while I prepare food” does.

A useful definition includes the setting and the first achievable version. A beginner may be ready for one second of four paws on the floor in a quiet room, not a minute beside an open door with a visiting dog outside.

Write the goal in terms of what the dog will do, then check physical suitability. A sit may be uncomfortable for a particular dog. A standing station might serve the same household purpose. If reluctance or a new change suggests pain, obtain veterinary advice instead of treating the behavior as an obedience problem.

For Mabel, the first goal is a brief voluntary nose touch to a safe target presented within easy reach. It has a clear beginning and end, requires little movement, and lets Ravi practice timing without involving traffic, guests, or a door that might open onto danger.

Consequences change future behavior

Reinforcement means that a consequence increases the future likelihood of a behavior in the relevant conditions. Calling something a reward describes the person's intention; observing what happens helps establish whether it functions as reinforcement for this dog.

Food can reinforce behavior, but so can access to a sniff, a toy game, movement toward something interesting, or another valued outcome. Value changes with context. A dry-food piece may matter in a quiet kitchen and compete poorly with a lively environment outside.

Positive reinforcement means adding a consequence that increases behavior. “Positive” here refers to adding, not a moral label. For example, a nose touch followed by a suitable treat may become more likely. The course uses reward-based methods because teaching effectiveness and the dog's welfare both matter.

The AVSAB humane-training statement recommends reward-based methods and discourages aversive approaches. This does not mean allowing every action or ignoring safety. It means arranging the environment and teaching useful alternatives without relying on pain, fear, or intimidation.

The setting teaches before the reward arrives

An antecedent is something occurring before the behavior: the cue, an approaching person, the available surface, the location, or another part of the situation. Changing antecedents can make success much easier.

If Mabel jumps when the leash appears beside the exciting front door, Ravi can begin practice farther inside the home. If a dog repeatedly steals food from a low table, securing the food prevents rehearsal while a separate skill is taught. Management changes opportunity; training changes what the dog has learned to do.

Both are useful. A gate that prevents entry into a work area is not a failure to train. It can protect the dog while the household builds a more durable routine. Conversely, removing every gate without testing the learned skill safely does not prove confidence; it may simply restore the original opportunity.

Ask what the current arrangement makes easy. If the answer is “jump, grab, or rush,” redesign the beginning before demanding greater self-control at the end.

Timing identifies the relevant action

A consequence closely connected to the intended behavior is easier to interpret than one arriving after a long sequence. If the dog touches a target and then jumps before food appears, unclear timing may reinforce a different pattern from the one you wanted.

Prepare before the session. Put suitable rewards where you can deliver them smoothly, choose the target position, and know what counts as success. Practicing your hand movements without the dog can reveal awkwardness that would otherwise become the dog's problem.

The Dogs Trust reward-training guidance emphasizes timely consequences and rewards the individual values. An ordinary small success delivered clearly is more useful than an elaborate exercise with confusing feedback.

Reward placement also matters. If you want the dog to remain near a mat, repeatedly throwing food far away creates a different movement pattern. If you want a reset for another target attempt, delivering the reward slightly away from the target can help, provided the movement is comfortable and safe.

A marker makes one moment easier to notice

A marker is a brief consistent signal, such as a quiet word or a click, that has been associated with a reward. It identifies the moment that earned the consequence. It is not a command and does not replace the promised reward.

The Dogs Trust marker introduction describes pairing the signal with an immediate reward before using it to identify behavior. Choose a signal the dog can perceive comfortably. A loud click near a sound-sensitive dog's head would be a poor starting point.

For Mabel, Ravi says a short marker word, then promptly delivers a small suitable food piece. After the relationship becomes familiar, he can mark the instant her nose contacts the target, then deliver the reward. He does not say the marker repeatedly while searching an empty pocket.

A marker improves communication only if the timing is accurate and the consequence remains reliable. If Ravi accidentally marks the wrong moment, he still provides the promised reward and makes the next attempt clearer. One imperfect repetition is a reason to adjust, not to punish Mabel for believing the signal.

Capturing, luring, and shaping are different routes

Capturing means noticing and rewarding a behavior the dog offers naturally. If Mabel lies comfortably on a mat, Ravi can reinforce that moment. He does not have to produce every behavior through a hand motion.

Luring uses something the dog follows, often food, to guide a movement. It can help establish an initial action, but the visible lure should not become the only condition under which the behavior occurs. Move thoughtfully toward a clear cue and a reward delivered afterward, without abruptly making the task impossible.

Shaping reinforces successive approximations: small steps toward the final behavior. For a target, the first step might be looking toward it, then moving closer, then touching. The criterion changes gradually as the dog becomes successful.

These are tools, not competing identities for trainers. Choose the simplest suitable route and observe the result. Avoid physically forcing the dog's body into position. If the behavior requires restraint to make it happen, you are no longer observing the voluntary learning you intended to teach.

An original target lesson

Ravi chooses a large safe target that Mabel cannot swallow and holds it a short comfortable distance from her nose. It remains still. When she looks toward it, he marks and rewards. After several easy responses, he waits for a small movement toward it before marking.

If Mabel immediately touches it, the task can begin there. If she is hesitant, Ravi presents it farther away or rewards a smaller response. He does not push the target into her nose and call the contact a learned touch.

Once brief touches are readily offered, Ravi changes the target position slightly. A small move to the side is one change; moving it across the room is several changes in distance, movement, and perhaps posture. The next step should preserve a high chance of success.

Ravi keeps sessions short enough that Mabel remains engaged. He records the target position and the successful response. This lets the next session begin at a level she already understands, with only a modest new challenge.

Add a cue to an understood action

A cue signals that a particular behavior has a useful consequence in this setting. Add it when the dog can already perform the behavior readily, rather than repeating an unfamiliar word while hoping it creates the action.

For Mabel, Ravi says “touch” just before presenting the familiar target. Over practice, the word and presentation become part of the learned situation. Later, he can vary the presentation carefully to check what Mabel is actually responding to.

Use the cue once, give a reasonable opportunity, and make the task easier if the dog does not respond. Repeating “touch, touch, touch” may teach that the relevant cue is a long string of sounds or simply add noise. It does not clarify a task that is too difficult or poorly understood.

Different household members should use compatible cues and criteria. If one rewards a nose touch and another expects a paw strike after the same word, the dog receives a confusing lesson. Agree on the behavior before debating whether the dog is listening.

Generalization is learning across conditions

A skill learned in one room may not transfer immediately to another person, surface, distance, or distraction. Generalization develops through suitable practice across those changes.

Imagine Mabel touching the target reliably in the kitchen. Ravi then asks in the hallway. He begins with the target close and the setting quiet, making the behavior easier while the location is new. When that is comfortable, he can add another small change.

For stationary skills, duration, distance from the person, and distraction are useful separate dimensions. Increasing all three together is a common way to make a known behavior fail. Build one while temporarily making the others easier.

A successful repetition in a difficult setting is encouraging, but reliability requires a pattern, not one lucky result. Keep physical safety measures in place for skills such as recall. Training should not be tested by releasing a dog near a road to see whether the cue works.

What to do when learning stalls

Check the task before blaming motivation. Is the behavior physically comfortable? Is the dog able to notice the cue? Does it understand the current criterion? Is the reward useful here? Is delivery clear? Has the environment become more difficult? Is the session too long?

Suppose Mabel stops touching when the target moves lower. The new position may require an awkward bend or place the object near a noisy appliance. Restoring the previous position can reveal whether the difficulty came from that change. It does not require a theory that Mabel is testing boundaries.

If she leaves, allow the session to end and review the setup. If a normally easy behavior suddenly becomes difficult, consider health and context. Repeatedly escalating the reward may conceal an uncomfortable task rather than solve it.

A short log helps: criterion, conditions, what happened, and next adjustment. Five clear observations can be more useful than fifty confused repetitions. Training efficiency comes partly from deciding what the next attempt should teach.

Why welfare belongs in the method

A method can suppress visible behavior while leaving the underlying experience difficult. Quietness alone does not establish comfort, and immediate compliance is not the only outcome worth measuring.

Vieira de Castro and colleagues' 2020 study compared dogs attending schools using different proportions of aversive and reward-based methods. It assessed behavior, cortisol measures in a subset, and a judgment-bias task. The findings associated aversive methods with poorer welfare indicators. Dogs were recruited from existing schools rather than randomly assigned to identical experiences, so the study has limits on causal isolation. It nevertheless contributes evidence relevant to the choice of humane methods.

At home, a useful plan asks both whether the skill improves and how the dog experiences learning. Voluntary engagement, manageable difficulty, and recovery matter alongside performance. A dog should not have to endure escalating fear for the person to call a lesson successful.

Keep rewards meaningful as skills develop

Once a behavior is well learned, the consequence can vary appropriately with the task and setting. Everyday opportunities can become useful rewards: coming back may lead to another safe sniff, and standing calmly may lead to the door opening for a walk.

Do not remove support abruptly in the hardest setting. A new distraction can justify returning to easier criteria and more frequent reinforcement. Maintenance is part of training, especially for important safety-related skills.

A reward delivered after a cued behavior is not a bribe merely because food is involved. The useful distinction concerns the sequence and learning: constantly showing food to obtain a response differs from a clear cue followed by an earned consequence. Even that distinction should guide better teaching, not become a reason to withhold appropriate reinforcement.

A lesson should leave a clear next step

Ravi's revised session is unremarkable to watch. Mabel touches a target, hears a marker, receives a reward, and tries again. The simplicity is the achievement. Both participants can understand what changed and what to practice next.

Write tomorrow's starting point before putting the equipment away. It might be “same room, target near nose, three easy touches, then one small position change.” If today was difficult, tomorrow begins easier. Progress follows learning rather than the ambition of the plan.

Check your understanding: A dog succeeds on a mat at home but fails when asked to remain there far from its person in a busy café. Which parts of the task changed, and what should the next practice look like?

Expected answer: Location, distance, distraction, and possibly duration changed together. Return to a quieter manageable setting, keep the person closer, use a short achievable interval, and build one dimension at a time. The failure does not prove defiance or justify aversive correction.

Application

Write a five-step target progression for a fictional dog. Include the starting arrangement, exact behavior to mark, reward delivery, one criterion for advancing, and the response to hesitation. Then identify an accidental lesson in Ravi's original sit-and-jump sequence.

Model interpretation: Begin with a safe reachable target in a quiet room, mark voluntary orientation or contact at the dog's current level, deliver a suitable reward promptly, and progress to a slight position change after repeated comfortable success. If the dog hesitates, reduce the demand or end and reassess. In the original sequence, delayed reward delivery after jumping may have made the consequence ambiguous or favored jumping. Preparing food and marking the intended moment makes the next attempt clearer.

Evaluate your plan: Another person should be able to perform it without guessing what counts as success. The dog should have a feasible route to the reward, and the plan should not depend on force or prolonged failure.