How to Get What You Want From Highlighting • Zettelkasten Method
How to Get What You Want From Highlighting • Zettelkasten Method
Saving highlights from books doesn’t make you recall them when you need them. A highlight is a writing prompt to yourself: process it in your own words, and then the idea takes root in your Zettelkasten.
Howdy, Stranger!
Categories
- 3K All Categories
- 153 Research & Reading
- 709 The Zettelkasten Method
- 13 Knowledge Work
- 102 Writing
- 475 Software & Gadgets
- 161 Workflows
- 735 The Archive
- 16 Plug-In Showcase
- 89 Resolved Issues
- 225 Projects Logs and Journals
- 86 Project: Zettelkasten.de
- 53 Critique my Zettel
- 182 Random
- 376 Introduce Yourselves!
Comments
Is there any scientific literature available for this technique?
My current understanding of dog communication is somewhat different.
First, traffic lights make sense when describing the behavior of a dog interacting with another dog or a human. How does a dog behave who is at ease (green)? How does a dog behave who is stressed (yellow)? How does a dog fight (red)?
As humans we might interpret a growl as a "warning". But I'm wondering if dogs think the same way. Maybe they're just saying "you're too close, go away"? Can dogs really understand a "No" as a warning as in "Heed my warning or else…"?
Second, there's a lot of psychological research available on operant conditioning1. I can understand why some dog trainers use fear and force when interacting with their dogs (P+).
But I can also understand why others prefer reward-based dog training2. It uses primarily positive reinforcement (R+). You teach the dog what he should do by rewarding desirable behavior. You try to use as little punishment, force and fear as possible. The difficult part is that you have to be really clear of what you consider desirable behavior.
I'm not sure where a "warning" fits in. Can you recommend an article or book that explains the technique in more detail?
* * *
BTW, here's a tidbit from my zettelkasten. In S11E20 of the TV series NCIS the hacker and former kindergarden teacher Heidi is being interrogated by two agents.3 Heidi uses three index cards with emojis to control the interview:
Psychologically, I find the "uh-oh face" interesting. Is it a warning? Does it mark a boundary?
For example: https://www.simplypsychology.org/operant-conditioning.html ;↩︎
For example: https://avsab.org/what-are-reward-based-training-methods-for-dogs-and-cats/; https://www.psychologytoday.com/us/blog/decoding-your-pet/201809/positive-dog-training-is-not-always-what-it-seems; https://avsab.org/wp-content/uploads/2021/08/AVSAB-Humane-Dog-Training-Position-Statement-2021.pdf ↩︎
Screenshot: https://ayearofncis.home.blog/2022/06/29/a-year-of-ncis-day-254-page-not-found-episode-11-20/ ↩︎
Just to clarify the terminology, when you write "entrainmemt" do you mean "training"?
Whatever the case for the growling dog, a puppy that has been harassing one reacts as if it understands the intention by rolling onto its back and presenting its belly. I've seen an annoyed adult put its snout with bared teeth right up to the puppy's belly. It doesn't actually bite because that seems to be innately inhibited. The puppies know this and I've seen them intentionally harass a dog until they get it to threaten them, intending to go belly up if needed.
In the case of baboons, they have a cry or call that people usually interpret as "I'm lost. Help me!" Other baboons may call out, and this lets the lonely one know where the pack is. But in a fascinating book about the work of a team of researchers in the field, the call is said to merely indicate that the baboon is feeling lonely or isolated. Another baboon hearing the call may notice that it is also far from the pack and emit one itself. But the response is not an intentional signal "Here I am". It only signals "Hey, I'm lonely too". Sorry, I can't dredge up the citation just now (try your friendly chatbot, I suppose).
Why are you putting these two ideas together like that? Operant conditioning has nothing specifically to do with force and fear. Operant conditioning is about pairing a stimulus with a behavior, such that the behavior ends up getting triggered by the stimulus. Most practical training does not use operant conditioning in a classical sense (I'm speaking as a one-time apprentice porpoise trainer; I knew Karen Pryor, who developed clicker training). It is more specific and targeted. BTW, it is possible to get trained without even being conscious of it. I know because it happened to me (a little joke by another trainer).
You always have to be clear about the behavior being reinforced, but in practice punishment is hardly ever precise. Usually it serves to make the animal fearful of the punisher without knowing just what brought on the punishment. Negatives are hard to convey. "Stop whatever it is you are doing" is not a precise instruction. It would be better to decide, e.g., that "No!" means "Stop where you are" or "Stand still", or even "Look at me", and use that as a definite behavior that can be reinforced.
That sounds more like classical conditioning (Pavlov).
In operant conditioning (Skinner) we have an antecedent, a behavior in response to the antecedent and a consequence that follows the behavior. In dog training those consequences can be either a pleasant (rewarding) stimulus, an unpleasant (aversive) stimulus or a lack of stimulus. Pleasant consequences tend to increase the likelyhood of a behavior, unpleasant consequences tend to decrease the likelyhood of a behavior (see link in first footnote).
In my zettelkasten I use this cheatsheet:
Dog training is mostly operant conditioning. The dog's behavior is followed by a human consequence. It can be unpleasant like yelling, yanking the leash or forcing a dog into a position. It can be pleasant like treats, praise and attention. And sometimes the key is to do nothing.
I know the feeling of getting trained unconsciously. I sometimes wonder who conditions who in a human-dog relationship. Somehow the dog often gets me to do what he wants. ;-) I'd say the main part of a dog trainer's job is not to train dogs but to train their humans. :-)
BTW, unconscious training is a topic in another TV show. In The Big Bang Theory (S3E3) we see how Sheldon is training Penny with operant conditioning. He uses chocolate as reward and water as aversive stimulus. https://bigbangtheory.fandom.com/wiki/The_Gothowitz_Deviation
Yes. That's the reason why reward-based dog training gets better results (according to the American Veterinary Society of Animal Behavior, see links in the second footnote).
Yes, I wrote operant but was thinking more of classical.
I have often thought the same.
It is not a technique but a communication model.
The standard book in general is this: https://www.amazon.de/Hundepsychologie-Sozialverhalten-Wesen-Emotionen-Individualität/dp/3440097803
The traffic light color coding is a nice simplification.
At least my dog and any dog that I know understand (human) warnings. To me that is enough.
If you want to simplify it, a warning could be seen as a cue for an incoming punishment.
Regarding your image: Negative reinforcement is the removal of a negative stimulus. It is all about change (removal) and not state (abscence).
I am a Zettler
A good Zettelkasten does not merely store what its human has learnt. It changes what this user will be able to learn in the future.
Yes, that's an excellent book. I have access to the 5th edition (2013).
Hmmm. I have difficulties to map it to Feddersen's descriptions. Maybe it's based on a different source?
I've had different experiences with the dogs in my life. ;-)
Maybe we understand the word "warning" differently?
Do you mean agonistic behavior ("Drohverhalten"), eg when growling indicates a readiness to bite?
Do you mean alerting others ("Warnbellen"), eg when someone approaches a dog's territory?
Do you mean announcing a conditional response, eg "if you do this, I'll do that"? Not sure that a canine brain can process such an announcement. Feddersen writes on page 107 (emphasis added):
"Caniden verfügen also über ein sehr komplexes, flexibles und differenziertes Kommunikationssystem. Was sie wohl nicht können, ist mit Hilfe ihrer Kommunikation über das unmittelbare Hier und Jetzt hinauszugehen. Dieses ist allein dem Menschen und seiner Fähigkeit zu sprechen möglich – die Kommunikation über vergangene und zukünftige Ereignisse und über Einflüsse, die nicht unmittelbar wahrnehmbar sind."
(Machine translation: "Canids thus possess a highly complex, flexible, and nuanced communication system. What they probably cannot do, however, is use their communication to go beyond the immediate here and now. This is possible only for humans, through their capacity for language: communicating about past and future events and about influences that are not directly perceptible.")
Hmmm. Why would you need a cue for an incoming punishment?
Not sure how to reconcile this with Feddersen's recommendations, in particular section "Bedingte Hemmung" (p. 383-386). If I understand her correctly, the "no" itself works as an aversive stimulus. It it's not enough, it would make more sense to combine it with a stronger stimulus rather than following it up with a more severe punishment (emphasis added):
"Die zugleich mit der »Verfehlung« verabfolgte Strafe hat, auch wenn sie nur leicht ist, Erfolg, weit mehr als schwere Strafen nach dem unerwünschten Verhalten (Hassenstein 2001). Denn nun verknüpft sich eine Verhaltenstendenz mit der unangenehmen Erfahrung. So wird eine »gelernte Hemmung« etabliert."
(Machine translation: "The punishment administered simultaneously with the ‘transgression’ is effective, even if it is mild—far more so than severe punishments administered after the unwanted behavior (Hassenstein 2001). This is because a behavioral tendency now becomes associated with the unpleasant experience. In this way, a ‘learned inhibition’ is established.”)
What am I missing? :-(
Thanks for the feedback! I updated the cheat sheet:
I used to play a game with my dog in which we faced each other and growled at each other. Then we would bark at each other, as simultaneously as possible. Then we would repeat the cycle. He loved this game and I was always the one who got tired of it first.
A fun way to repurpose a warning!
Lucky dog! This sounds like a safe relationship between you two. You both trust each other to know the difference between real fighting and playfighting, between actually threatening each other and playing with behavioral patterns. Nice. :-)
I linked to the general standard book. I never meant to imply that the traffic light color coding model is based on something in this book. It was just an explanation by this one trainer.
They share both the characteristic of announcing an incoming negative event.
This would imply a dog doesn't understand that growling means f off or else?
To move the aversive stimulus from the punishment to its announcement via conditioning.
I am a Zettler
OK, thanks.
I don't think that dogs are announcing a future event. The signal makes also sense in the here and now. Translated to human speech: "You are in my/our territory. Get out!"
Yes. If I understand Feddersen correctly, dogs can't consciously process "…or else" in the way we humans do.
However, there's dog trainer Ed Frawley who writes in The Theory of Corrections in Dog Training (PDF): "I am a firm believer in giving a dog the chance to think and consciously refuse to do what I have asked him before I correct him." (I like that he makes his bias explicit. He tells us about the products he sells and what he thinks about vets.)
This explanation is compatible with my understanding of dog psychology, because it doesn't require the dog to consciously choose between scenarios. The NO itself becomes more aversive. If obedience is rewarded ("good dog") the combination of light positive punishment and strong positive reinforcement should work well.
I found this article interesting about the role of science in dog training: https://undark.org/2022/09/12/punishment-puppies-and-science-bringing-dog-training-to-heel/
I just read the linked article and it's not very satisfactory, is it, at least as to resolving some of the issues "scientifically".
No work that I have read takes into account how dogs discipline puppies and other dogs. It's clear that they do so, using mixtures of growls, posture and expression, nips, and bites (rarely serious). Puppies don't become timid or frightened personalities by these occasional episodes, in my experience. And this makes sense because there has to be some means for older dogs to communicate displeasure and warnings to youngsters, and have the youngsters learn from them.
I conclude that mild actions like a whack on the rump can be effective without causing trauma or fearfulness, especially when they are somewhat similar to normal dog actions. Problems come when the actions are excessive, too frequent, too frightening, or come for no reason apparent to the dog. This is not much different than for human children, really.
Just like people, different dogs have different sensitivities. One might be terrified by something that another one barely notices.
That's the reason why I mentioned it. Apparently there are no simple answers; but there are some answers that are too simple. :-)
The book @Sascha mentioned describes it, for example on page 260. It comments on agonistic behavior of the mother dog (machine translation):
"This agonistic behavior on the part of the mother dog thus occurs at a point in time when behaviors aimed at reducing distance — that is, contact-seeking behaviors between mother and puppies — determine their relationship to one another. The puppies do not react to the bitch with fear, however. They gradually come to associate where she sets limits for them. And they learn to respect this behavior. (…) And this distance-creating takes place during the socialization phase, in which important imprinting-like bonding processes between mother and puppies occur. Nevertheless, the bond between mother and puppies does not suffer in any way from this. The sole consequence of the temporary aggressiveness is that the mother, in this way, builds up a social status within the small group — that she establishes and secures her high rank position, her special rights vis-à-vis the puppies."
(Original: "Dieses agonistische Verhalten seitens der Mutterhündin tritt also bereits zu einem Zeitpunkt auf, wo Verhaltensweisen zur Distanzverminderung, also Kontaktverhaltensweisen zwischen Mutter und Welpen deren Verhältnis zueinander bestimmen. Die Welpen reagieren nun nicht etwa ängstlich auf die Hündin. Sie assoziieren sukzessive, wo diese ihnen Grenzen setzt. Und sie lernen, dieses Verhalten zu respektieren. (…) Und dieses Distanzschaffen findet statt in der Sozialisierungsphase, in der ja wichtige prägungsähnliche Bindungsprozesse zwischen Mutter und Welpen ablaufen. Dennoch leidet die Bindung zwischen Mutter und Welpen keineswegs darunter. Folge der zeitweisen Aggressivität ist einzig und allein, dass sich die Mutter auf diese Weise einen sozialen Status innerhalb der kleinen Gruppe aufbaut, dass sie ihre hohe Rangposition, ihre besonderen Rechte den Welpen gegenüber etabliert und sichert.")
On page 263 (machine translation):
"The punishing bitch grabs the puppies, knocks them over, and also pinches them quite roughly. A human hand can imitate all of this. It can press the young dog down and push it away — which ends a play sequence when the puppy plays too wildly and bites. The puppy will remember this, because the interruption of play is a distinctly negative experience."
(Original: "Die strafende Hündin packt die Welpen, stößt sie um, zwickt sie auch recht unsanft. Das alles kann eine Menschenhand nachahmen. Sie kann den Junghund runterdrücken und wegschieben – was eine Spielsequenz beendet, wenn der Welpe zu wild spielt und zubeißt. Er wird es sich merken, denn ein Spielabbruch ist ein ausgeprägt negatives Erlebnis.")
Some positive and negative punishment seems quite natural.
Yes. That matches what Dorit Urd Feddersen-Petersen writes in her book.
I should have said "in the context of animal training".
Very interesting, thanks. My own dog always seemed to have the attitude towards my disciplinary actions of "Oh, that's what you want? OK". Very matter of fact, he was.