Do and I Do

Question 1

What is do as I do

Answer 1

Teaching dogs via social learning - I’ll show you the action

Question 2

What do we look at wt a new method?

Answer 2

We go back in time - the do as I do paper then go back to see what papers came before i.e. what the new paper is building on

Question 3

What date papers are we looking at?

Answer 3

2006
2009
2010
2013

Question 4

What is the 2006 paper looking at

Answer 4

m

Question 5

Who are the main players do as I do

Answer 5

Miklosi
Fugazza
Csanyi
Range
Topal

Question 6

Who are the main players - team anti imitation

Answer 6

Udell
Wayne
Kaminski
Dorey

Question 7

What is the 2006 paper about?

Answer 7

Philip the Terv

Question 8

How old was Philip the Terv?

Answer 8

4

Question 9

What was Terv’s background?

Answer 9

Service dog trained

Question 10

How did the researcher train Terv?

Answer 10

OC

Question 11

What happened for10 training sessions

Answer 11

They added a cue - Trainer doing the action is the cue for the behaviour

Question 12

What did they train?

Answer 12

Spin
Jump in the air
Bow
Lie down
Put bottle in the box
Take the bottle to the owner
Move stick
Jump over
Give a bark

Question 13

What questions would we immediately ask?

Answer 13

Isn’t this stimulus control not imitation

Question 14

What was involved in the testing phase of do as I do

Answer 14

1 Can he do these actions on cue (“imitate”)
2 Can he do these actions if someone else cues them?
3 Can he do untrained actions demo’d by the trainer of his owner? [imitation]

Question 15

What were the novel actions

Answer 15

Move shoe
Throw bottle (take bottle to another person)
Go round human
Pull arm
Open/close door
Search in sand
Take out and put in objects
Pull cloth
Push swing
Laying down on top of cupboard
Crawling into the cupboard
Put the tape on the chair
Go and turn
Drinking
Push childtoy

Question 16

What are dogs inclined to do?

Answer 16

Follow routes

Question 17

What is a prop cue?

Answer 17

Dogs with a lot of training who know that when there is an item they are expected to do something with it

Question 18

What are some problems with the 2006 paper with Philip

Answer 18

Some of the behaviours are not really novel to Philip. There are props and there is route taking

Question 19

What was the point of the first stage of the experiments with Philip?

Answer 19

Teaching dogs they should imitate - pay for that - then ask for true “imitation”

Question 20

How did philip do in the testing phase?

Answer 20

If the trainer cued B - 72.7% scored right
If new person cued B - 68.6% scored right

For untrained B
67% scored right

Question 21

How was Philip scored?

Answer 21

An observer’s opinion!

Question 22

What was experiment 2 with Philip

Answer 22

Can you put a bottle and put it in the same receptacle as the owner

Question 23

What did they find in experiment 2 with Philip?

Answer 23

Correct direction 26/60 trials
Correct receptacle 21/60 trials

Question 24

What were the flaws with experiment 2?

Answer 24

Philip was now 6 and had been practising experiment 1 for 2 years and had done demos

Question 25

What was the conclusion from the 2006 experiments with Philip?

Answer 25

His performance could be related to his service dog training and work
Prop cues
but “it seems that dogs have some imitative abilities”

Question 26

What did Huber, Range et al. say at the beginning of their paper?

Answer 26

“dog studies have provided evidence of action copying and the reproduction of results”

Question 27

Why is this discussion in the Huber, Range et al. study contentious?

Answer 27

It’s based on a study of 1!! Philip!

Question 28

What did Huber & Range acknowledge in there discussion

Answer 28

Philip had training, and familiarity of actions may have a major influence. Aka. we trained him to do it

Question 29

What study did Huber, Range et al. study?

Answer 29

1 dog - Joy Weimaraner! Trainer to do 8 behaviours in the same ways as Philip

Question 30

How did Huber test Joy?

Answer 30

They tested her on compound actions (behv chains in the right order) and “exotic” actions (not trained behvs)

Question 31

What are “exotic actions”

Answer 31

Imitation test

Question 32

What did the find wt Joy?

Answer 32

Pretrained - same as Philip
Chains - 1/3 correct, 2/3 wrong
Exotic actions - 0 !!!!

Question 33

What did the conclusion of the Huber Et al. paper say?

Answer 33

Showed some tendency to approximate the action in three trials with object-orientated actions”

Question 34

What is the problem with Huber et al. conclusion

Answer 34

They are trying to prove the hypothesis, they really want there to be something in it!

They make noise and then compare Joy & Philip to apes and are aggressive in their statements that these dogs are showing “imitations”

They refer to Csibra & Gergely 2007 paper but neglect to refer to the replication paper where the replication event was a dead fail.

Question 35

What did Range, Huber & Heyes examine?

Answer 35

Automatic imitation - not “pre-trained”
They tried to “pull” imitation out of dogs

Question 36

What happens in species with very good imitation software?

Answer 36

They don’t need to be taught to do it!

Question 37

What did Range, Huber & Heyes do?

Answer 37

10 dogs (agility/s&r or obedience trained)

Preliminary training - taught to open sliding door with paw & head

Then added a cue

Question 38

How many groups were the Range, Huber & Heyes dogs split into?

Answer 38

4

Question 39

What were the compatible group in the Range, Huber & heyes research?

Answer 39

2 dogs trained that owner using head = use your head
2 dogs trained that owner using hand = use your paw

Question 40

What were the noncompatible group in the Range, huber and Heyes research?

Answer 40

3 dogs trained that owner using head = use your paw
3 dogs trained that owner using hand = use your head

Question 41

What was Range, Huber and Heyes hypotheses?

Answer 41

Compatible group will learn cues faster

Question 42

What were the results in the Range, Huber and Heyes experiments?

Answer 42

After 350 trials

No dogs had achieved 85% correct responses if there was no command. i.e. they couldn’t do the B on cue!?! This was known as the non-command procedure.

If the owner gave a verbal cue when they demo’d “paw” or “head” and again when the dog approached, the dog did better. This was known as the command procedure.

The compatible group did better - 85% in 83 trials
The compatible group did worse - 85% in 265 trials

Question 43

What were Range, Huber and Heyes conclusions

Answer 43

The command procedure played some non-specific role in enabling the dogs subsequently to learn to use their owner’s behaviour as a discriminative cue

Our animals received hundreds of observation and test trials. (we did a lot of training!)

Then they say in conclusion the results of this study show the first evidence of automatic imitation and of automatic counter imitation in dogs.

Question 44

What is the problem with all these papers?

Answer 44

They ach. the weaknesses and then they say but it proves imitation! very aggressive interpretations

Question 45

What did Fugazza and Miklosi research?

Answer 45

8 dogs

A delay between cue and response rules out stimulus or local enhancement

Dogs already trained in “do as I do”

Question 46

Why do we need to drill down into original research when referred to in a paper?

Answer 46

what were the studies, what were they based on?!
“two independent studies”
Philip and Joy!! N=1 !!

Question 47

What happened in the F&Miklosi studies?

Answer 47

3 pre trained
3 trained in study

that owner doing B was the cue to copy

Dogs gradually trained to defer - 5 sec, 10 sec. up to 30 secs) sit … wait……“ok now” do they do the B?

Then tested familiar actions, novel actions and a “2 action test” - was actually stimulus discriminatio

Question 48

What did F&Miklosi find?

Answer 48

90% scored match across all conditions

Test = 18 trials

Question 49

What was the issue with the novel actions in the fugazza and mlikosi experiments on “do as I do”

Answer 49

The novel actions strongly resembled the pretrained actions!

Question 50

What did Fugazzia and Miklosi conclude?

Answer 50

Novel action conditions and in both 2 action conditions subjects imitated the novel behaviours without delay without any previous practice of these particular actions, so that their memory and recall could not have been based on reaccessing a motor habit, because none was formed.

The novel actions WERE very similar to the Trained actions so how can this be “true”

Without the ability to base their recall on the aid of previous motor patterns .. taken together these results suggest the presence of some form of declarative memory for dogs!

Question 51

What did Udell et al study?

Answer 51

Building on the paper regarding –> Dogs outperform chimps/wolves in the task of following human gestures (Hares etc)

The sensitivity of dogs to human cues may not be a function of domestication but rather learning throughout life experience that they should pay attention to human gestures because they have more opportunity to learn due to there comfort with proximity to humans.

Question 52

What did Udell et al think we should rule out?

Answer 52

OC - conditioning to follow human limbs

Question 53

What do we take away from the do as I do webinar?

Answer 53

Cognition researchers would do well to better understand OC, specifically stimulus control, if they are do convince those versed in OC

Novel behaviour needs defining and more rigor

Owners probably will still prefer saying sit or leave it than doing the action themselves as a cue for the behaviour!

Question 54

Would should a novel beh. incorporate

Answer 54

Physically able
NOT pre trained
Doesn’t involved route taking or a prop

Question 55

What did Udell conclude?

Answer 55

Don’t get to excited about imitation, we already have ways to explain this stuff!

Have you ruled out OC?

Question 56

What is Udells 2 stage hypothesis

Answer 56

Rather than saying dogs have this innate ability to imitate couldn’t it be:

1) social imprinting to humans during the socialisation period allows dogs to be less fearful of humans

2) Experiences with relevant human stimuli so that associations between specific stimuli and available R+ can be formed, are both important to the developments of sensitivity to human action.