Lecture 2 - Counterfactual Reasoning

counterfactual reasoning (CF)

-ability to reason about situations and events removed from, or contrary to, current reality

-includes the ability to think about the past, future and situations with multiple possible outcomes

cognition and time

-the ability to consider events or possibilities that are physically and temporally removed underpins many complex cognitive skills

fixation with present reality

-Piaget noted that young children frequently make egocentric realist errors

-studies have shown that three year olds have difficulty separating themselves from their own view of the world

-shown by their poor ability to reason counterfactually

features of counterfactual reasoning

-thinking of what might have been requires that we:

• separate ourselves from current reality

• mentally construct an alternative version of the world where one thing has changed

• track how this single change would have affected the rest of the world

separating from reality (development of counterfactual thinking)

-first step towards counterfactual thinking is separating ourselves from the world around us

-children first do this when they engage in pretend play from around 18 months

-pretence depends on the ability to represent absent objects and situations

pretending (separating from reality)

-pretending is not the same as acting in error

-pretence is when we act as if something is the case, while also correctly perceiving the current reality

-creating an imaginary world can be done without any reference to current reality

-pretending is not counterfactual thinking → CF involves considering specific ways that the current reality could be different

thinking of what might have been (development of counterfactual thinking)

-evidence for CF thinking emerges during preschool

-assessed by telling children stories and getting them to re-imagine the stories with one thing changed

-understand that at a specific point in the past, two possible worlds diverged because of a single causal event → understand the causal relation between a specific past event and its subsequent outcome

Harris - method (thinking of what might have been)

-told children that Sally walks across the floor with muddy shoes

-now the floor is dirty

then ask them if the floor would be dirty if Sally had taken her dirty shoes off

Harris - results (thinking of what might have been)

-three year olds were correct 75% of the time

-four year olds were correct 84% of the time

-suggests that preschool children can reason counterfactually

Riggs & Peterson - method (development of counterfactual thinking)

-told children that Peter the fireman is at home

-he gets a telephone call telling him that there is a fire at the post office

-so he goes to the post office to help put the fire out

-they are then asked where peter would be if there hadn’t been a fire?

Riggs & Peterson - results (development of counterfactual thinking)

-only 47% of 3 year olds answered correctly

-responded comparably poorly on a standard ToM task

Riggs & Peterson - interpretation (development of counterfactual thinking)

-can’t answer by pretending or using real world knowledge → have to think about the change in the story and how it would affect other events

-similar CF/ToM performance may be because both require children to consider more than one perspective on a situation

contradictory results for development of counterfactual thinking (Riggs vs Harris)

-children could answer the Harris question but not the Riggs’ question just by using their knowledge of the world works → did not have to think counterfactually

-do these tasks involve thinking counterfactually or do they just involve a simple heuristic

Rafetseder - method (counterfactuality in children)

-children told that mum makes a cake and puts it on the shelf

-the little girl sees the cake and takes it to her room

-asked where the cake would be if the boy had come along instead of the girl

-then told the mum puts the cake on the top shelf

-the boy takes the cake to his room (he can reach the top shelf)

-asked where the cake would be if the girl had come along instead of the boy (she cannot reach it)

Rafetseder - questions (counterfactuality in children)

-first question can be answered with simple reasoning

• cake would be in the boys room not girls room

-other question required thinking about how the counterfactual world would be different

• would she be able to reach the cake? if so the cake would be in her room, if not the cake would remain on the shelf

Rafetseder - results (counterfactuality in children)

-what if the boy had come along instead?

• 90% of 5-6 year olds were correct

-what if the girl had come along instead of the boy?

• 20% of 5-6 year olds were correct

• 39% of 9-11 year olds were correct

• 100% of 12-14 year olds were correct

-argued that adult-like CF reasoning does not typically emerge until early adolescence

evaluation of Rafetseder

-tasks may make additional reasoning demands in terms of informational complexity

-in middle childhood, children can relate the real and counterfactual possibilities

counterfactuality in children

-the age at which counterfactual thinking emerges depends on how you define it

-children become increasingly able to think about different aspects of counterfactuals

-can use real-world knowledge to answer simple CF questions by 3 years

-use basic conditional reasoning to answer CF questions by 6 years

-reason in a more adult-like CF was from around 12 years

Robinson & Beck - method (explaining children’s difficulties with CF)

-showed children a road with two garages, with one at either end

-a toy car drove into one of the garages, children were asked either:

• past → what if he’d driven the other way - where would the car be?

• future → what if next time he drives the other way - where would the car be?

Robinson & Beck - results (explaining children’s difficulties with CF)

-past question

• 30% of 3 year olds were correct

• 60% of 4 year olds were correct

• 95% of 5 year olds were correct

-future question:

• 95% of 3 year olds were correct

• 90% of 4 year olds were correct

conditional reasoning (Robinson & Beck)

-both questions require children to give an answer contrary to the current reality

-yet 3 year olds appear well able to respond correctly when asked about a future event

-thinking about the future seems to be easier than thinking about the past

relating imagined counterfactual world to the real world

-understand that at a specific point in the past, two possible worlds diverged because of a single causal event

-understand the causal relation between a specific past event and its subsequent outcome

-children find relating the two possible worlds challenging

Beck - mouse game method (explaining children’s difficulties with CF)

-children were shown a slide with two outlets

-asked to place a cotton wool to ensure the mouse landed safely

-after the mouse slid down, children were asked two questions about the future:

• what if next time he goes the other way?

• can you put out the cotton wool to make sure the mouse lands safely next time?

Beck - mouse game questions (explaining children’s difficulties with CF)

-to answer “what if next time he goes the other way?” → only have to think about a single possibility

-”can you put out the cotton wool so he lands safely?” → requires children to think about two possibilities

Beck - mouse game results (explaining children’s difficulties with CF)

-not until six years where children could answer the difficult question with multiple possibilities in mind

-children can think about a single possibility in the future → find it harder to think about multiple possibilities in the future

-children’s difficulties arise not from the past or future, but from whether they are required to consider multiple possibilities

processing demands and counterfactual reasoning

-children become better at dealing with more complex CFs with age, as their general cognition improves

-development is not a conceptual shift from inability to full ability

-incidental processing demands become easier to cope with as children get older

German & Nichols - story (processing demands)

-Mrs Rosie planed a flower in her garden

-she calls her husband to come out and look at her flower

-the dog escapes when her husband comes out and squashes the flower

-Mrs Rosie is very sad about this

German & Nichols - method (processing demands)

-told children a story using picture cards

-asked one of three questions about it:

• what if the dog hadn’t squashed the flower? → short causal chain

• what if the dog hadn’t escaped from the house? → medium causal chain

• what if Mrs Rosie hadn’t called her husband? → long causal chain

-varied the length of the causal chain involved in each question

German & Nichols - results (processing demands)

Type	Short	Medium	Long
3-year-olds	69%	20%	15%
4-year-olds	100%	86%	67%

-the shorter the causal chain, the better the children did

Beck, Riggs & Gorniak (processing demands)

-compared 3 and 4 year olds’ performance on CF tasks, EF tasks and language

-CF performance was predicted by both inhibitory control and by language ability

• supports the idea that difficulty with CF reasoning is ignoring what we know to be true

• working memory did not predict CF performance

regret and relief

-result from the comparison between actual and counterfactual possibilities

-requires us to consider a particular situation and how that situation compares to other possible alternatives

-more difficult to understand than happiness or sadness

regret

-reaction that arises when our actions give rise to a reality that is less desirable than it could have been

relief

-reaction that arises when our actions give rise to a reality that is more desirable than it could have been

Amsel - method (understanding others’ emotions)

-asked children to play a game where they chose between two cards and received a reward based on the card they chose

-sometimes, the other card would have given a much better reward

-sometimes it would have given a much worse reward

-were asked to judge how other people would feel in particular game situations

Amsel - results (understanding others’ emotions)

-5 year olds could correctly judge whether participants would be happy or sad if a particular outcome occurred

-but showed no understanding of how people would feel if they found out a better or worse alternative could have been chosen

-7 year olds correctly judged how people would feel if they found a better or worse alternative had been available

Kahneman & Tversky (understanding others’ emotions)

-adults feel worse when:

• a negative outcome arises through something they choose to do, rather than something they chose not to do

• a negative outcome arises following an unusual choice of action, rather than a habitual or usual choice of action

Guttentag & Ferrell - method (understanding others’ emotions)

-looked at children’s views of how other people would feel

-they gave children stories in which two people experienced negative emotions, but varied whether:

• the event was typical or atypical for protagonists

• the outcome arose through action or inaction

Guttentag & Ferrell - story (understanding others’ emotions)

-for lunch Mary always has chocolate pudding

-Susan always has vanilla pudding

-however, today Susan decided to have chocolate pudding

-there were germs in the chocolate pudding and everyone who at it got sick

• who would feel worse, Mary or Susan

Guttentag & Ferrell - results (understanding others’ emotions)

	Mary	Susan	same
Adults	9%	79%	12%
9-year-olds	8%	78%	14%
7-year-olds	14%	67%	19%
5-year-olds	19%	17%	64%

-5 year olds responses only reflected an understandings of the actual outcome, and no influence of considering alternative outcomes

-7 year olds approached adult levels of performance when judging complex emotions

self vs other judgements (understanding others’ emotions)

-there’s a slight lag between children experiencing counterfactual emotions and attributing them to others

-7 year olds reported feeling regret on a gambling task, felt worse when they discovered a better prize was available

• but when watching other people in the same situation, they reported those people would feel just as happy

anticipating counterfactual emotions

-adults experience counterfactual emotions when reflecting on what could have happened in the past

-but the relation of CF to reality is complicated by the fact that adults anticipate counterfactual emotions

adult counterfactual thinking

-susceptible to various biases

-engage in more counterfactual thinking when people act rather than fail to act

factors affecting counterfactual thinking

• executive functions

• language

• domain-specific knowledge

• conceptual change in children’s thinking about time

executive functions (factors affecting counterfactual thinking)

-to reason with information that is known to be false we need to:

• resist reasoning with what we know to be true → inhibition

• to relate both the counterfactual and the real world requires holding them in mind → working memory

• to make comparisons requires switching between them → attentional flexibility

language (factors affecting counterfactual thinking)

-language may be necessary for counterfactual thought

-may need to understand specific grammatical constructions to think counterfactually

domain-specific knowledge (factors affecting counterfactual thinking)

-a child’s ability to think about possibilities and counterfactuals may differ depending on the child’s causal knowledge in the relevant domain

conceptual change in thinking about time (factors affecting counterfactual thinking)

-children come to appreciate that different events could slot into a particular time point

-changes in temporal concepts and in EF may occur in parallel