Living and Learning with Animals - 3 days with Susan Friedman
When you're interested in training animals, when you work with animals with behavior problems, or when you suffer from behavior problems yourself, sooner or later, you'll find that life is messy. Now there are two kinds of training-and-behavior people. The ones who seem to say: life is messy. Avoid the mess (Turid Rugaas, Anne Lill Kvam etc.), and the ones who say: life is messy. Clean it up. (Behavioral analysists.) I, for one, am clearly a "Clean it up" kind of person: I want to pinpoint the mess, see what precedes it and what it results in, and then I want to organize it in whichever way I please. In other words: I love behavioral science. I want things to be explainable, quantifiable, mesurable, and I want to be able to change behavior rather than avoid the conditions that make it occur. For me, science equals control (over myself and my environment). Control equals empowerment. And I want to be empowered! (And I want my dogs and students to be empowered, too.)
Enter Susan Friedman. Susan Frieman who says just that: we can use behavioral science to empower our companion animals! Hell, yeah, I'm sold! I'm sold because of Susan's philosophy of empowerment that frames her talk. I'm sold because her appreciation of her audience is as tangible as appreciation can be, her smile is real, and her ear is open to our every question. Also, I'm sold because it's hard to organize a great seminar, and it's hard to give a great seminar. Susan Friedman and Happy-Fellow managed to do just that. And this is how.
Attending seminars with renowned speakers flying across an ocean in order to speak in Austria is expensive. It always is. Passion is what makes us happily pay the seminar fee. However, if we pay a lot of money, we expect something in return - not just from the speaker, but we also expect the general set-up to adhere to certain standards. We expect an environment that makes listening and learning fun and easy. We expect a spacious seminar room, a sufficient number of sockets for everyone who wants to plug in a computer, and a comfortable temperature. We expect lunch breaks with good food and an opportunity to connect with likeminded people. We expect free fresh water. We expect free W-LAN so we can tweet and post about the awesomeness of the speaker on Facebook in real-time.
While it's always expensive for the organizer to invite a renowned speaker, and it's connected with months of work. It's resource- and time-consuming to organize a conference or a seminar, and to organize it well. If you organize such an event professionally, pay your speaker a fair salary, and keep the fee affordable for the audience, you're practically a volunteer. You can't host a well-organized event like Susan Friedman's seminar for financial gain, because the hours and hours of work you'll put into preparing, planning and setting up, and the amount of money it costs to rent a decent seminar room and pay for a delicious coffe and lunch buffet for your audience doesn't leave much for yourself. Especially if you think of all the little things such as goodies for seminar bags or ruffle rewards and offer early-bird registration fees. I'm happy to report that the Seminar organized by Happy-Fellow's Nicole M. Pfaller and Gernot Sadovsky exceeded my expectations.
And now, let's get to the content!
http://www.youtube.com/watch?v=3YY7IhoebQo
"Like gravity, the laws of behavior apply to all living animals, whether or not we recognize them at work. These are natural processes in the wild, in captivce facilities, and in our homes. We didn't invent them, we just harness them!" (Susan Friedman)
Those of you who know me know that I'm addicted to writing seminar reviews. I feel like everyone should know how awesome certain speakers are! So here goes. These are some of the areas Susan discussed that I found most interesting.
- Operationalize your labels!
The problem with labels is that they give us information about how the speaker feels about the animal rather than how the animal behaves. In order to successfully change behavior, we first have to try and get rid of labels. "We can't teach animals what to be but we can teach them what to do and when." (Slides part 1, p. 10; my emphasis). I love how how Susan fine-tuned our awareness of our use of labels and insisted we operationalize them in order to come up with observable data: don't teach "friendly". Teach "recall" and "relaxed body"! Don't talk about "reactivity," but operationalize, i.e. describe what can be observed: does your dog lunge? Bark? Etc. "Lunging" and "barking" are descriptions of observable behavior that can be changed. "Aggressive" is a label - and may well be a self-fulfilling prophecy (4).
We can operationalize concepts like "aggression," "fear" etc. on a physiological, neurological, and overt behavioral level. The latter one is most interesting for designing a behavior modification program, since these are the reactions we can observe (e.g. freeze, fight or flight, duck, panting …) And yes, of course we can modify these reactions: "Although the basis of fear is innate, the antecedent stimuli that trigger it can be acquired through respondent learning." (slides part 2, p. 1)
Susan's wonderful pop culture example for labels was Gloria's encounter with the "skittish" horse. Those of you who are watching Modern Family will know!
Oh, yes! I love functional assessments.
Misconception: behavioral science based trainers don't believe that animals have emotions.
Rectification: behavioral science does not make such a claim. We just know that we cannot measure emotions. We can never know the experience inside an animal (not even if this animal is a fellow human being). Therefore, we rather focuse on behavior: behavior is "[w]hat an animal does, given certain conditions that can be observed". (slides part 1, p. 13).
Misconception: there is a dichotomy between nature vs. nurture. Either something is in our nature, or it is based on our environment.
Rectification: nope! It is our nature to learn; i.e. nature and nurture go hand in hand. E.g. our experiences activate or do not activate certain genes.
Misconception: lots of the behaviors you train are unnatural behaviors!
Rectification: there is no such thing as an unnatural behavior. If it were unnatural, it would not be possible for an environment or experience to enable such behavior. Animals learn lots of things in the wild, and they learn lots of things in captivity. These things may be different, but none of them are more or less "natural." Remember: nature and nurture go hand in hand.
Misconception: behavioral science is mechanistic, or simply a "carrot on a stick" approach!
Recrification: the "[b]ehavioral model has a high degree of scientific rigor, and social significance but a low profile in society at large" (slides part 1, p. 6). Our culture makes it hard for us to see behavior as a tool for a purpose (like, for example, our eyes are tools for seeing). Behavior is an evolved tool to affect the environment in ways that are valubale for us.
Why, however, do we tend to disregard the behavioral model as mechanistic or simplistic? Susan suggests it is because we tend to be biologically deterministic: we tend to assume that we are our genes and believe the behavior that goes with them is just, as Susan has it, "spurting out of us". However, we are not our genes: our genes are quiet until the environment pulls them (i.e. pulls behavior our genes make us capable of). Our genes are "turned off or on", depending on the environment.
One example commonly cited for the genetic base of our behavior used to be the fact that cowbirds growing up in a sparrows' nest speak cowbird, not sparrow, in the end of the day. However, studies showed that when a young cowbird in the nest makes a cowbird sound, the female cowbirds within range will answer and therefore reinforce the cowbird sound. (It is a reinforcer on the neurochemical level of analysis.) Our level of analysis is not the neurochemical level, but the level of the environment pulling behavior. What takes place with the young cowbirds is a beautiful example of operant learning. It is not mechanical or simplistic - it is effective, useful, and fascinating (if you ask me, anyways).
Misconception: "… but positive reinforcement doesn't work!"
Rectification: this is an oxymoron. It is impossible for positive reinforcement to not work: if it doesn't increase the frequency of the behavior, by definition, it isn't R+!
- The law of effect: we behave today based on the consequences our behavior had yesterday.
Once we understand the relation between what we do and the outcome, we can change observable consequences by changing conditions. Behvior is a function of its consequences.
- R+, or: putting money on the relationship account
If I share a history of positive reinforcement with an animal, she is more likely to comply than if I don't share any history with this animal: every time I use positive reinforcement, I put money into the relationship account. Relationship, of course, is a label. Let's have it mean: when an animal chooses to approach rather than remain distant.
This is to say: if I use lots of R+, I get an animal who will be much more likely to comply than if I just remain neutral, "just letting the animal be" rather than putting money on the relationship account. This is one of the reason I love R+: it creates a deeper relationship and an more active and vivid dialogue between two species than any other training method.
- "We need to empower the animals to be participants in our conversation."
I love Susan's philosophy. It reflects the way I see clicker training: I want my dog to train me as much as I train her. It's not necessarily about teaching perfect behaviors - the clicker becomes a mediator, a tool that bridges the language gap between two species (dog and human): I click when the dog does something I want. The dog can elicit the click (and treat) by means of figuring out what will make me click.
Susan emphasized that especially if a behavior is not under perfect stimulus control, there is a dialogue going on. Two-way communication! And that dialogue is one of the most fascinating, rewarding aspects of training!
What we do - says Susan - is not "Clockwork Orange". Rather, we want to influence one another. If people watch me and say that my dog is training me rather than the other way round (we've all been there, haven't we?), there's nothing wrong with that: of course we teach our dogs or our children that they can "manipulate" us. After all, that's what conversation, what dialogue is about!
I like that Susan brought up this point, since I have had disagreements with trainers who reject the clicker on the basis of its potential for causing addiction. An example commonly cited for the supposed perils of of the clicker is the client who manages to reinforce behaviors, but doesn't manage to put them under complete stimulus control. It is argued that this addiction to the click stresses the animal a lot by means of putting constant performance pressure on him.
I can follow the argumentation, but I don't agree. Of course, you can misuse any tool (including a clicker). However, I think in the majority of cases, even if a client doesn't manage to put perfect stimulus control on her dog's behaviors, the clicker empowers her dog and facilitates two-way communication between the species. The client will end up with a dog who offers behavior in certain situations (under certain circumstances where offering behavior has been reinforced in the past). In other situations (e.g. client falling asleep on the couch, client in bathtub (unless, that is, said client has read Peggy Tillman's book!), her dog is unlikely to offer behaviors because they have not been reinforced under these conditions in the past. The dog will not have lost down time and periods of relaxation. Rather, he will have gained periods of conversation that can be mutually initiated.
Furthermore, there is another powerful argument against the "addiction" theory: if every click is followed by a reinforcer, your clicker is a coke machine rather than a slot machine. And while we get addicted to putting money in slot machines, I've yet to meet someone addicted to feeding the coke machine.
- Every click should be backed up by a reinforcer.
Clicking should be like a soda machine, not like a slot machine. "If [the] click doesn't reliably predict [the] back up reinforcer, then it really becomes a bridge that leads to nowhere." (cf. slides part 2, p. 19) If it is not backed up with a well established reinforcer, the animal will find a more reliable marker - one that is 100% reliable that the treat will come (e.g. your hand reaching into your treat bag). The strength of the click as a secondary reinforcer depends on its reliability as a predictor of the treat!
I always treat when I click. However, some good clicker literature suggests that it's easiest to reinforce every click, but that it is also possible to only treat after some clicks, though this might be am approach suggested for expert rather than novice trainers.Backing up each click with a reinforcer doesn't equal treating every single time I give a cue. It just means that if I click, then there will reliably be a tangible back-up reinforcer (food or toy). If I say "Good girl!" instead, there won't necessarily be a back up reinforcer. If I say "Yes!", what follows will often be a Premack reward (go get that pigeon; say hi to that dog!).
I really liked Susan's illustration of why it makes more sense (and is not just easier) to always back up the click with a reinforcer:
The coke machine operates on a continuous reinforcement schedule (just like my click): this also makes it most reliable in predicting my actions: whenever I am thirsty, I will reliably go to the coke machine. It's not stressful; I am 100% sure that if and when I put in a coin, I'll get a soft drink (which reinforces my future behavior the next time I'll be thirsty and encounter a coke machine.) In case the coke machine is broken, I'll probably throw a tantrum or kick it, because I expected it to work, but then I'll walk away rather than throwing in more and more coins to no avail.
If I back up every click with a reinforcer, my dog will not be stressed by clicking and will not have to throw a tantrum. But if and when he's hungry, he'll most reliably work for treats, and when he's wide awake, he'll most reliably work for tennis balls. Score!
Slot machines work differently: they operate on an intermittent reinforcement schedule. That is to say, they are very unreliable predictors of reinforecers. If I throw in a coin, I don't necessarily expect it to give me a reinforcer. If it doesn't reinforce me, I won't throw a tantrum, but I will throw in another coin. And another. And another. Slot machines build persistence. They can be more frustrating and stressful than coke machines, and they do not reliably mean that whenever I'm hungry/thirsty, I'll throw money in a slot machine.
In animal training, what we generally need is quick learning and strong behavior - not persistence. Continuous reinforcement builds very STRONG behavior (whenever I'm thirsty, I WILL throw money into the coke machine) and facilitate FAST learning. This is what we want for most training goals. Continuous reinforcement does not build PERSISTENCE, but persistence is quite unnecessary for most training goals anyways. So: if you click, treat! (2)
- Empowerment and control
Remember how I (and, well, Susan, of course) said we need to empower animals to be particionats in our conversations? Then I gave you an example of the clicker as a facilitator of conversations. However, we haven't yet defined what "empowerment" means: It means that the animal has control over its environment. More precisely, we are talking about the learner being able to control its own outcome - i.e., to use behavior effectively. Control is central to behavioral health, and (maybe we should even say: because) it is a primary reinforcer.(1) This is, by the way, also the explanation for the fact that punishment is reinforcing to the punisher: she experiences control over her environment and is reinforced by this experience.
A lack of control over one's own outcomes leads to a lack of resilience. If we (or other animals) lack resilience, we're going to have a hard time in this world! We are born with resilience, but that resilience is a function of how much control we have over our outcomes. This doesn't mean that every animal must have control over every single outcome - this is simply not possible. Rather, it means that the "ratio of control:no control should be very high".
- Control via enriched environments
The notion of control relates very nicely to the effectiveness of enriched environments. Susan Friedman finally gave me the scientific explanation I've been missing in Anne Lill Kvam's course. Anne Lill keeps insisting on enriched environments as part of nearly every (?) problem behavior intervention program we define as trainers. Sometimes, she even suggests regular exposure to enriched environments as the sole solution to various problems clients might have with their dogs. I heard what she said, but I never understood why exactly enriched environments would be so effective.
Last weekend, Susan gave me the scientific explanation I need in order to embrace Anne Lill's training approach and recognize its scientific validity. "Empowered animals are brave learners", as Sarah Owings has it. That is to say, the more an animal experiences being able to control its environment, the bolder it will become towards new stimuli. "Bolder" means that the animal will curiously approach a novel object rather than engage in distance-increasing behaviors. If you want your young parrots to grow up to be bold adults, you'll let them grow up in enriched environments, i.e. with lots of stimulus diversity (i.e. change the cage environment every day). Birds are not wary by nature, but wary by experience! If your bird grows up with little or no stimulus diversity, it will grow up to be wary.
The same holds true for raising puppies. If you raise a puppy according to Ian Dunbar's advice, you're on the best way to have a bomb-proof adult dog, since Dunbar's focus is on the greatest possible stimulus diversity for your puppy.
This is why enriched environments may also help (adult as well as young) dogs to experience control over their own outcomes via interaction with the environment. Remember, control is a primary reinforcer. Getting lots of this primary reinforcer of control leads to a bolder animal. (An object in the environment is only considered enriching, btw., if the animal is interested in interacting with it. Otherwise, it is purely ornamental.)
- Resilience: other contributing factors
Having a rich bank account from R+ training also adds to the animal's resilience. This resilience is useful in the few situations (e.g. emergency) where we have to force an animal. For example, this may happen at the vet's. If we have built up resilience, the animal will "bounce back" quickly after a negative [negative used in its everyday sense] experience. Because the animal is empowered the rest of the time, it is okay to withdraw a little from that bank account when it can't be avoided. It's not about empowerment 100% of the time. Life on earth is messy and not always empowering, and that's okay. It's about the ratio of positive reinforcement/empowerment:punishment/force. There always has to be a lot more weight on the R+/empowerment side of that scale.
It's okay to be relaxed about controlling the environment! You don't need to keep it free from negative [everyday use of the word] stimuli. Never being exposed to even mild aversive stimuli might even shatter the subject: the person may lack the experiences that lead to resilience.
Susan's thoughts on that matter (and her telling of an anecdote involving her daughter Marnie and a roll of toilet paper!) were prompted by Christine Schragel's question. Christine had an experience with one of her dogs that strengthens the assumption that it may be counterproductive to shelter your puppy from all and even the mildest aversive stimuli.
However, there are renowned trainers out there who teach that, indeed, especially when growing up, a dog should - ideally - never be exposed to even mild aversive stimuli. That is to say, you should strive to control the environment as much as possible. There are more and less radical views on that matter out there, but I would venture that trainers like Turid Rugaas and Anne Lill Kvam would disagree with Susan's suggestion to relax about controlling the environment, especially when it comes to puppies. For example, Anne Lill suggests in her trainer education that the first one or two weeks your puppy is with you, she should not go out to meet other dogs or people, but be given time to adapt to his new home. In Austria, most puppies move into their new homes when they are 8 or 9 weeks old. This would mean that they not leave your house and garden until they are 10 or 11 weeks old!
Other renowned trainers such as Ian Dunbar suggest a radically different approach: "Your puppy must socialize with at least 100 different people before she is three months old. That's just twenty-five people a week, or four a day." (Dunbar, Before & After, p. 207) etc.
Personally, I side with Dunbar's and Susan's approach, because science shows that resilience grows with stimulus diversity. And I believe that even the most versatile environmental enrichment doesn't provide as much stimulus diversity as meeting and interacting with unfamiliar people and objects "out in the real world" (in addition to the own house and garden) does. However, if Turid Rugaas, Anne Lill Kvam can point me to a study and can show me data that proves that her approach is more successful in building resilience, I will be happy to change my mind.
One could argue that it is unncessary to build up the maximum amount of resilience that could be reached by lots of stimulus diversity in the puppy's home and out in the world: one might suggest that as long as the animal's environment is controlled by the handler all its life, the animal would never need the strong resilience it didn't build up as a puppy. I assume that you could keep a dog this way for all his life - you could make to always control his environment, i.e. keep it free from even mildly aversive stimuli like honking cars etc. (Which will in itself be reinforcing to you, because control is a primary reinforcer.)
However, for me, this argument is not a practical one. Personally, I believe that every animal (human as well as non-human) should "be able to enjoy as much freedom as she can handle," as Leonard Cecil nicely put it a while ago (Facebook status, June 17, 2014). "Freedom" includes being off-leash on walks - on walks in the city and on walks in the forest. It also includes the possibility to accompany me when I travel, or go to work. It means being okay on subways, and being okay in restaurants and other busy places, and being okay at friends' places, and in hotel rooms. (I am not saying I am forcing my animals to join me - rather, I want to give them the option to join me sometimes, and not other times, when they would rather stay home. And yes, we communicate: sometimes, they will follow me to the door with wagging tails, or lead the way and motivate me to venture out with them when I hadn't even planned it. Other times, they will stay on the couch instead. And that's a choice I respect.) It means it's okay for them to be dogsat by friends, even if this means being put in unfamiliar situations and meeting new people. It means being okay in shopping centers, being okay in busy dog parks, and being okay when meeting other dogs or other animals. In order for all these things to be okay, my dogs need resilience. And if I get a puppy, I will always try to build resilience to give her the greatest possible freedom to join me on my ventures, and to be off leash in as many places as possible. I've tried to give Phoebe lots of stimulus diversity as a puppy, but I would try to give my next puppy even more.What do others think? Do you agree or disagree with me? I would be interested in your opinions on building resilience, and on the greatest possible freedom your dog can handle. Let me know in the comments!
Contrafreeloading refers to the obervation that when given the choice between "eating for free" and "working for their food" (i.e. using their behavior effectively), animals tend to choose the food that requires effort over the free food. This is why clicker training is so much fun!
This theory is not new to me. However, Susan presented it in such an enticing way that I was compelled to try a little contrafreeloading experiment myself.
My Standard Poodle Phoebe knows puzzle feeders well, and as far as I can tell, she enjoys them (yes, I know that "enjoy" is a label rather than a behavior description. But let's say, that's okay for now). However, I have never offered Phoebe a choice between food in a bowl and food in a puzzle feeder - I usually serve her either the one or the other.
In my experiment, both the food bowl and the puzzle feeder hold the same amount of the same food. I expected Phoebe to choose the puzzle feeder - however, surprisingly to me, she chose the free food first and only started on the puzzle feeder once she had finished the food in the bowl. I repeated the experiment the next day in a different room and with the puzzle feeder on the left and the food bowl (a different food bowl) on the right, but once again, she finished the free food first and engaged in the puzzle feeder afterwards.
http://www.youtube.com/watch?v=DB-LQorRXUY
I repeated the same experiment with my foster Galga Luz. She has has only been with me for a short while, and has only recently been introduced to puzzle feeders. While she seems to enjoy them a lot, I expected her to choose the free food over the puzzle feeder, since she has little experience with it and tends to eat very fast. This is exactly what she did: she finished the food in the bowl first and engaged in the Wobbler afterwards.
http://www.youtube.com/watch?v=x1nYce4iw_U
It's interesting that my dogs chose to eat for free! I'll be sure to think about this some more and devise a follow-up experiment!
- An ethical guideline: least intrusive, effective alternative.
Why? Because effectiveness alone is not enough. And because science agrees with the least intrusive, effective alternative (cf. slides part 1, p. 42).
- Which strategy do I employ in order to change an animal's behavior?
A hierarchie of interventions, from least intrusive and most socially acceptable (1) to most intrusive and least socially acceptable (6):
The way I understand it, this graphic reflects the continuum from least to most intrusive for the average animal. However, as Susan repeatedly pointed out, it is important to remember that it is always the study of one: we are not dealing with an entire species, we are not dealing with the average dog. We are dealing with the individual, with the animal in front of us. So while she suggests to always start attempting to revise behavior with the least intrusive approach and then, if it doesn't work, move down the pyramid, the order of the hierarchy may be slightly different for any individual animal.
So we [people bound by the "least intrusive, effective" ethics] start at the least intrusive level and move down the hierarchy if the (correctly applied) approaches don't work. Does that mean if level 1, 2 and 3 fail to work, we will move into more intrusive territory? Are there instances where, if we apply the "least intrusive, effective alternative" ethics, we will use extinction/negative reinforcement or even positive punishment? The simple answer is: yes. However, first of all we will review our less intrusive approach carefully, and get feedback from others. Only if and when we are sure that nothing else works and that we applied the less intrusive approaches expertly will we consider using positive punishment. Over the course of our training experience, this will happen extremely rarely or not at all, since most problem behaviors can effectively be changed with level 1, 2 or 3. Positive reinforcement is an extremely powerful tool, and almost everything can be solved with it. For example:
http://www.youtube.com/watch?v=-IUeroC4g28
- Why I agree with the "least intrusive, effective alternative" ethics
For me personally, the "least intrusive, most effective" ethics resonates more with me than the "I will ONLY use R+" credo of some colleagues. A "least intrusive, effective approach" allows for full use of all the tools in the tool box, if necessary. Therefore, this approach is able to change a bigger range of problem behaviors than the "I will ONLY use R+" approach.
Another important message to take home from Susan is that there is no obvious relationship between extremeness of a problem behavior and the strategy that succeeds in changing it. A more severe problem behavior does not require a more intrusive approach (which is basically what Cesar Milan wrongly claims as justification for his use of P+). Very severe problem behaviors can be changed by means of R+! However, the more severe the problem behavior, the more in-depth should be its assessment.
If P+ isn't working, it hasn't been punishment. Punishment is the mirror image of reinforcement: just as, by definition, a consequence is only reinforcement if it increases the frequency of a behavior, it is only punishment if it decreases the frequency of a behavior. If you think you have been positively punishing your child/spouse/dog/parrot etc., but their behavior hasn't become less frequent, you have not really been punishing them. Simply because a consequence produces escape behavior (is perceived as uncomfortable) does not make it punishment. It is only called punishment if it weakens the behavior that came before. If it doesn't weaken the behavior, but still produces escape behavior (e.g. taking a step back), what will happen is adaption: the animal gets used to the stimulus and will slowly tolerate a higher and higher stimulus intensity.
Susan described an experiment where rats were put in a cage with a feeder. Whenever they fed, they received a mild shock. The shock was uncomfortable engouh to produce escape behavior (the rats jumped back), but not strong enough to weaken the behavior itself (they still returned for more food). Hence, it was, by the scientific definition, not punishment.
The intensity of the shock was slowly increased, and eventually, the rats were able to tolerate a rather strong shock.
Another group of rats was placed into a cage and immediately shocked with the strong shock the very first time they went to the feeder. Without gradual increase, there was no opportunity for the rats to adapt to the intensity of the shock. This was punishment: it weakened the behavior. The rats did not go back to the feeder; they would starve to death rather than getting more food.
This is something anyone working with P+ has to bear in mind. And this is the reason why parents who start talking loudly when unhappy with their child will often end up screaming at their children sooner or later - all to no avail, since the children have adapted.
As Susan says she likes to point out to her graduate students: her use of positive reinforcement is a choice, not a weakness. If she wants to, she can punish with the best of them. It's useful to bear that in mind yourself: if you ever want to really punish your spouse, for example, do it right and prevent adaption! ;)
- Motivating (establishing) motivations:
We're in operant territory here. We can "change the frequency of a behavior by temporarily altering the effectiveness of the consequence". (slides part 2, p. 8). It's ALWAYS possible to change the strength of a reinforcer.
Yey! This is great news! Even when a reinforcer is extremely strong - say, intrinsically motivating, such as getting to chase someone or something - we can temporarily weaken its strength by means of letting the animal have a good run before training. We can temporarily change the strength of a reinforcer by moving the animal forther right or left (depending on whether we want to decrease or increase the strength of the reinforcer) on the continuum of
hunger ………………………………………… satiation
fatigue ………………………………………… rested
activity deficit ……………………………… activity excess
For example, for me, chocolate is a strong reinforcer. However, when I'm satiated, it's considerably less strong. On the other hand, a rare treat like Lindt or Zotter chocolate is even more reinforcing for me than an everyday treat like Milka.
Also, finally, I've got a scientific source for the fact that small and frequent reinforcers build behavior more quickly than large, occasional ones! Susan mentioned Schneider (1973), and Todorov, Hanna, & Bittercourt de Sa' (1973). Will have to look into these in more detail.
- Cues (discriminative stimuli)
Give stronger reinforcers rather than bigger/louder cues! Cues can be very small - they need to be clear, but not big. The stronger the reinforcer, the stronger the cue will become. "To get positive reinforcers, learners often enthusiastically exceed the minimum effort necessary to gain them." (slides part 2, p. 14) For R-, on the other hand, animals will often only do the minimum effort necessary for escape.
- Antecedent arrangement to make behavior more likely
I loved Susan's example: Watch the video "The piano staricase" on http://www.thefuntheory.com. The antecedents (the stairs) are tunred into a giant piano to make it more likely that people will take the stairs rather than the escalator. (The other videos on this website are also lots of fun to watch, by the way!) Antecedent arrangements are quite common in everyday life, actually. At least in my life. Usually, when I want to seduce or entice someone, I'll carefully arrange my apartment in a way that might make it more likely that they fall for me. I'll select certain kinds of music, clean my apartment, maybe get my visitor's favorite snack - I might casually mention that I just happen to have made Guglhupf the other night -, and I might even move certain books or DVDs to a place where they can be easily seen by my visitor, and hide others.
- Non-contingent reinforcement
if you freely give the reinforcer the animal used to get by means of crying/biting etc., the frequency of the problem behavior will decrease. (3) I found a nice outline of NCR online that shows how NCR can be used to change attention-seeking problem behaviors in children.
- The cultural fog, or: a philosophy of teaching
Susan shared her philosophy of teaching with us, and I am very grateful. She has found perfect metaphors for concepts I've been struggling to get a grip on. Listening to her, it was obvious that for Susan - unlike for a surprisingly high number of "positive" trainers! - ethics doesn't only apply to non-human animals. It is equally important to her to treat her human clients with the utmost respect.
She reminded us that she finds it helpful to think of clients in the following way: "The client is me, just with less information." This is not only a useful way of looking at it if we want to build healthy and productive relationships with our clients, it is also true. Clients aren't "stupid" or "ignorant"; they really just are us with less information.
Simply burying them under an avalanche of new information, however, is not the way to go. If we want to convince someone of the effectiveness of R+ and the unnecessariness of P+, we are not only asking her to change the way she treats her dog, but the way she sees the world. Approach your clients gently, because you are asking a lot! Keeping this in mind keeps us compassionate teachers.
What makes it so hard to accept new information is what Susan calls cultural fog. Cultural fog refers to the conventional wisdom that can become an obstacle to learning. "… despite new information to the contrary, conventional wisdom has a property analoguous to inertia that opposes the introduction of contrary belief." (slides part 1, p. 2)
Cesar Milan (who we should be thankful to for being such a convenient negative example to so many things) uses cultural knowledge rather than scientific knowledge: he operates within the cultural fog, which is why his followship is so big. He caters to beliefs already held within the cultural fog and refines them.
No one is immune to the cultural fog. While I might be "out of the cultural fog" when it comes to canine behavior modification, I am sure there are other areas of life where I'm comftably holding my position in the cultural fog without even being aware of it. So it's not okay to "blame" others for not being able to see clearly: it's hard to see through the fog. Just try driving over the Wechsel in November, and you will know what I mean.
I enjoyed Susan's cultural fog metaphor not only because it applies to clients with dogs, but also because it applies to the human students in the adult German classes I teach for immigrants in Austria. Feminist that I am, I regularly discuss gay equality, adoption rights etc. in class. I usually come out as a gay woman when there is a homophobic comment, when someone criticizes the pride parade or complains about the life ball poster …
My wish is to lead them out of the cultural fog. I ask them whether it's okay if we discuss this topic a little further, and if they agree, I tell them stories. Personal stories: stories about falling in love with a woman. Political stories: stories about the 40 differences that still exist between domestic partnerships and marriages in Austria. Stories about countries where being gay gets you sentenced to death. I try not to preach, but just share. I don't take it personally if they disagree, and respect their opinions, their cultural background: they are just me with less information.
I encourage them to ask questions; political ones, life-style related ones, and personal ones. I ask questions myself: I ask who can think of reasons why a pride parade may be considered necessary or useful by the gay community. I ask them what we (the gay community) might want to archieve with public visibility. I want them to come to the realization that not having to care what others think is a privilege rather than a choice. I try not to tell them things, but to listen to their opinions, validate them, and challenge them in ways that let them practice their German - and maybe, just maybe, take a tiny little step out of the cultural fog. Maybe now that I have Susan's metaphor, I can lead them out of the fog even more effectively? We'll see! I'm already looking forward to the next time the topic crops up in class.
As for clients with dogs, that's a piece of cake for me. If you deal well with something as personal as gay rights in the cultural fog, it's easy to deal with dog training clients in the cultural fog. If you happen to be gay and a trainer, I encourage you to come out to conservative strangers more often - it'll make your encounters with clients just seem so much easier.
- Don't underestimate your clients!
I also liked that Susan encourages us to not underestimate our clients. I completely agree: Of course, our clients are smart enough to undestand and implement behavior modification programs if we just explain them well. Even (or especially) children can be great teachers for their animals if we just manage to be decent teachers for them. (Compare the fantastic video with Noah's parrot below!) We are no more intelligent than our clients; it's just that our respective expertise lies in different areas.
I believe that, if a client or student doesn't manage to do well, it's because we have failed them as a teacher, not because they have failed us as a student. It may be easier to complain about our clients than to take a critical look at our teaching approach, but the latter is just so much more effective - and so much more rewarding! After all, happy clients make for positive feedback, and positive feedback is a strong reinforcer for us! (This is why I love teaching: I'm addicted to the positive feedback! And if there sometimes is critical feedback - and there is - well, I love a good challenge as well. I will work even harder on my relationship with that particular student and try to implement his wishes in the classroom.)
- A-*B-C assessments and case studies
I particularly enjoyed the last part of day 3, when we employed what we had learned and did A(ntecedent) - *B(ehavior) - C(consequence) analyses ourselves. First we practiced with videos with different animals, then we paired up and did an A-B-C assesment and found ideas how to modify antecedents and consequences in order to modify behavior for a case one of us dealt with. Christine Schragel and I worked on Nala's case, one of the dogs she works with at Wiener Tierschutzhaus. Trying to describe Nala's behavior without using labels, and identifying the antecedents and consequences to her problem behavior was quite challenging! We came up with a number of ideas Christine will try to implement, and I'm looking forward to hearing about Nala's progress! I keep my fingers crossed for her!
When doing an A-*B-C assessment, Susan suggests we always start with the *B, and then fill in the respective A and C. We would first do A-B-C for the problem behavior. Then, we would think about what we want the animal to do instead of the problem behavior, and how we could arrange antecedents and consequences to make this replacement behavior more likely to occur than the problem behavior. However, in addition, we would also preserve the reinforcer for the problem behavior (the original C): this is a reinforcer valued by the animal, and Susan suggests we always also come up with an acceptable alternative way for the animal to earn the reinforcer that was originally gained by the problem behavior. It seems a very good idea to preserve all reinforcers!
… and of course, there was just SO MUCH MORE I could talk about. But then I'd never finish this post. So even though it's hard to stop, I leave you with a final thought: if you get the opportunity to attend one of Susan's seminars, do so. It's a wonderful opportunity, it's a lot of fun, and even if you have worked with animals for a long time, there's still going to be lots you can learn from Susan.
Wheee … and that's all of us posing with our "certificates of excellence" on day 3:
Susan's website: behaviorworks.org
Behavior toolkit by Susan Friedman and James Fritzler
(1) Examples for control as a primary reinforcer, as cited by Susan: Watson, 1967, 1971 (human babies); Joffe, Rawson, & Muliak, 1973 (rats); Mineka & Henderson, 1985 (Rhesus macaques).
(2) Susan mentions 5 sources for the growing consensus on this approach among experts: Fernandez (2001), Martin (2011), Ramirez (1999), Bailey (personal communication with Susan, 2011), Pryor (personal communication with Susan, 2011).
(4) Cf. Rosenthal, Pygmalion effect.
