Cognitive Psychology

Lecture 7: Reasoning
October 08
John Toner
 Reasoning
Studying the human memory system involves
questions about how we acquire and retain
knowledge

Problem Solving and Reasoning research

investigates what we do with this knowledge

Reasoning involves using knowledge within

systems of formal logic
 Reasoning
Reasoning can be defined as the mental processes
by which people derive conclusions from a given
set of premises.

E.g. Thursday is the day after Wednesday - premise

Today is Wednesday - premise

Tomorrow will be Thursday - conclusion

 Reasoning
There are two types of reasoning:

Inductive Reasoning: Involves deciding what is

probably the case based on one’s knowledge

E.g.
Every morning in the past the sun has risen in the east
Therefore the sun will rise in the east tomorrow
 Reasoning
There are two types of reasoning:

Inductive Reasoning: Involves deciding what is

probably the case based on one’s knowledge

E.g. of a turkey’s inductive reasoning

I have been fed every day up to today (23rd Dec)

Therefore I will be
fed tomorrow (24th Dec)
 Reasoning
Inductive Reasoning:

As this example illustrates, in inductive reasoning, when the

premises are true, the conclusion is not necessarily true.
The conclusion can only be judged true with a certain
degree of probability.
 Reasoning
Deductive Reasoning involves conclusions that
follow with certainty from the premises.

E.g.
If it is raining in Dublin there will be ripples in the Liffey
It is raining in Dublin
Therefore there are ripples in the Liffey
 Inductive Reasoning
We use inductive reasoning all the time to make
decisions about the world

It is getting cloudy and dark, its probably going to rain

If I flick the light switch, the light will come on

If I don’t eat something I’ll get hungry

 Inductive Reasoning
Inductive reasoning results in a hypothesis
Testing a hypothesis will result in either confirmation
or falsification

• Confirmation involves finding evidence is support of the

hypothesis.

• Falsification involves finding evidence that does not support

the conclusions.

• NB: A hypothesis cannot be proved

 Inductive Reasoning
Confirmation Bias:

People tend to test hypotheses by seeking confirming

evidence rather than by attempting falsification of the
hypothesis.

Confirmation bias is observed for both ordinary people and

professional scientists (Tweney, 1998) even though
falsification can be regarded as what distinguishes science
from unscientific activities (Popper, 1968)
 Inductive Reasoning
Confirmation Bias:
“The Earth is flat”

Confirmation bias leads to the following test:

It appears flat

Falsification leads to the following test:

If one sails westward for long enough
they will arrive back home from the east
 Inductive Reasoning
Confirmation bias is evident in peoples social thinking

Stereotyping: “All skinheads are violent”

People are very good at remembering instances that support

these judgements

People tend to neglect instances which do not support these

judgements
 The 2-4-6 task
Wason (1960) investigated strategies used by people when
testing hypotheses

Participants were told that there was a general rule for

grouping 3 numbers

As an example they were told that ‘2-4-6’ conforms to the rule

They had to suggest examples in order to discover what the

rule might be
 The 2-4-6 task
The actual rule was: Three numbers ascending in value

Therefore the following would all conform

4-6-8
1-3-7
100-150-200

People were generally bad at discovering the rule. 28% failed

to discover it at any stage
 The 2-4-6 task
What was happening?

People were coming up with a hypothesis:

“The rule is ascending in twos”

They tended to come up with suggestions that confirmed this

rule.
“What about 5-7-9. What about 20-22-24”

The problem was that these all conformed so they believed

their hypothesis to be true
 The 2-4-6 task
In fact, the best was to test a hypothesis is to try to falsify it
“What about 6-8-9”

Doing this leads to discovery of the rule

 The 2-4-6 task
Klayman & Ha (1987) argue that this experiment is flawed if
we try to generalise the findings to real life reasoning

They argue that the difficulty with the 2-4-6 task is that it
possesses the unusual characteristic that the correct rule is
much more general than any of the initial hypotheses that
participants are likely to form.

As a result, positive testing cannot lead to discovery of the

correct rule, and negative testing is required
 The 2-4-6 task
Tweney (1980) carried out tests on a variation of the 2-4-6
task. They were instructed to find two rules rather than just
one

One rule called DAX, was “three ascending numbers” (i.e.

Wason’s original rule)

The other rule, called MED, was any other triple (i.e. does not
obey the DAX rule).

Each time a triplet of numbers was suggested by participants,

they were told that it was either a DAX or a MED triplet
 The 2-4-6 task
Tweney (1980)

People were much better at discovering the DAX rule than in

Wason’s original study.

Tweney did not come up with an explanation of the effect

Nevertheless, it shows how the way a task is presented

effects how it is tackled, and thus must reveal something
about how are reasoning works
 The 2-4-6 task
• One explanation (proposed by Evans, 1989) is that people
have a positivity bias in their hypothesis testing strategy.

• The idea of positivity bias supposes that people are more

likely to make positive tests of their hypothesis than
negative tests.

• Since negative testing is required to find the rule in the

original 2-4-6 task, participants’ positivity bias makes this
task difficult.

• However, the dual goal paradigm allows participants to use

positive tests of their hypotheses about the MED rule in
order to gather information about the DAX rule.
 Deductive Reasoning
Deductive reasoning allows us to draw conclusions that are
definitely valid provided that the other statements are
assumed to be true

Conditional Reasoning:
If it is raining in Dublin then there are ripples in the Liffey
 Deductive Reasoning
Deductive reasoning allows us to draw conclusions that are
definitely valid provided that the other statements are
assumed to be true
Syllogistic reasoning

Peter Paul Paul Patrick

 Deductive Reasoning
Conditional Reasoning involves deciding something based
on knowledge about something else
Reasoning based on if and then

If it is raining in Dublin there will be ripples in the Liffey

It is raining in Dublin
Therefore there are ripples in the Liffey
 Deductive Reasoning
Conditional Reasoning
It is raining in Dublin (We will call this A)
There are ripples in the Liffey (We call this B)

We know if A, then B

This rule of inference is known as modus ponens

 Deductive Reasoning
Conditional Reasoning
It is raining in Dublin (We will call this A)
There are ripples in the Liffey (We call this B)

We also know

If B is false, then A is false

If there are no ripples in the Liffey then it is not raining

This rule of inference is known as modus tollens

 Deductive Reasoning
Conditional Reasoning
It is raining in Dublin (We will call this A)
There are ripples in the Liffey (We call this B)

What about

If A is false, then is B false?

Not Necessarily!
If it is not raining, there could still be ripples in the Liffey

This is known as denial of the antecedent

 Deductive Reasoning
Conditional Reasoning
It is raining in Dublin (We will call this A)
There are ripples in the Liffey (We call this B)

What about

If B is true, then is A true?

Not Necessarily!
If there are ripples in the Liffey, then it is not necessarily
raining

This is known as affirmation of the consequent

 Deductive Reasoning
Marcus & Rips (1979) The percentage of subjects endorsing
the various conditional inferences
 Deductive Reasoning
Syllogistic Reasoning:
Mayo is in Ireland
Ireland is in Europe
Therefore Mayo is in Europe
 Deductive Reasoning
Mistakes with Syllogistic Reasoning:
Biases: People accept believable conclusions and reject
unbelievable conclusions irrespective of their logical validity

All French people drink wine

Some wine drinkers enjoy cheese
Therefore some French people enjoy cheese
This conclusion does not follow from the premises
 Deductive Reasoning Theories
There are three major theories to be considered

• Abstract Rule Theory

• Mental Model Approach

• Probabilistic Approach
 Deductive Reasoning
Abstract Rule Theories

According to Braine, and others, in several publications, the

following processes occur when someone encounters a
deductive reasoning problem
 Deductive Reasoning
Abstract Rule Theories
1) The premises are comprehended and encoded into a
mental representation in working memory
2) Abstract-rule schema’s are applied to these premises in
order to derive a conclusion (e.g. modus ponens)
3) Feeder schemas are applied to produce intermediate
conclusions
4) Incompatibility rules examine the contents of working
memory for any incompatible references (e.g. inferring
both ‘A’ and ‘not A’)
 Deductive Reasoning
Abstract Rule Theories
1) The premises are comprehended and encoded into a mental
representation in working memory
Ireland are playing in a football match. Crowd in pub are watching
2) Abstract-rule schema’s are applied to these premises in order to
derive a conclusion (e.g. modus ponens)
If Ireland do well the people watching will be happy. Loud cheer!
3) Feeder schemas are applied to produce intermediate conclusions
I hear a loud cheer during the match
4) Incompatibility rules examine the contents of working memory for any
incompatible references (e.g. inferring both ‘A’ and ‘not A’)
Could there be Cyprus fans in the pub? Would the be that loud?
Could Ireland fans cheer a Cyprus goal because they want the
manager to get sacked?
 Deductive Reasoning
Abstract Rule Theories
Braine (1984) argued the reasons why people make errors in
reasoning:
1) Comprehension errors: The premises are interpreted
incorrectly. (e.g. If there are ripples in the Liffey then it
must be raining)
2) Heuristic inadequacy: The participants reasoning
processes fail to locate the correct line of reasoning.
(They fail to see the link between rain and ripples)
3) Processing errors: The participant fails to attend fully to
the task at hand or suffers from memory overload.
(Distraction, interruption, not thinking things through)
 Deductive Reasoning
Abstract Rule Theories
General assumption:

Normally people reason correctly provided they don’t

misunderstand the premises, get distracted etc.
 Deductive Reasoning
Abstract Rule Theories
Reasoning error: Affirmation of the consequent
“There are ripples in the Liffey” leads incorrectly to the
conclusion “It is raining”
According to Braine et al. (1984) this occurs because of a
conversion error
“If it rains there are ripples in the Liffey”
Is replaced by
“If there are ripples in the Liffey, it is raining”
 Deductive Reasoning
Abstract Rule Theories
Braine goes on to say that there is an assumption amongst
people that we are being given certain information for a
reason.
“If you mow the lawn, I will give you 5 euro”
Invites the inference
“If you don’t mow the lawn, I wont give you 5 euro”

How is this exploited in advertising?

 Deductive Reasoning
Abstract Rule Theories
Braine et al. reduced the likelihood of a conversion error by
providing an additional, clarifying premise:

e.g. If it is raining, then Alicia gets wet

If it is snowing, then Alicia gets wet
Alicia gets wet
Conclusion…
 Deductive Reasoning
Abstract Rule Theories
Limitations:
‘Comprehension’ component is under specified, so it is hard to
make predictions about how well a person will reason
The theory has only been applied to a limited range of
reasoning tasks
Individual differences are de-emphasised.
There is little convincing evidence that people use mental
logic when presented with deductive reasoning problems
 Deductive Reasoning
Mental Models
Proposed by Johnson-Laird (1983, 1999)
A mental model is a possibility for the way things are in the
world
E.g. Premises
The lamp is on the right of the pad
The book is on the left of the pad
The clock is in front of the book
The vase is in front of the lamp
Conclusion: The clock is to the left of the vase
 Deductive Reasoning
Mental Models
Here’s this particular model:
book pad lamp
clock vase

We can construct a model that is consistent with the premises

and the conclusion. This indicates that the conclusion is
valid
 Deductive Reasoning
Mental Models
Assumptions:
• A mental model describing the given situation is
constructed, and the conclusions following from this are
generated
• An attempt is made to construct alternative models that
will falsify the conclusion
• If a satisfactory alternative model is not found, the
conclusion is assumed to be valid
• The construction of mental models taxes the Working
Memory system
 Deductive Reasoning
Mental Models
Assumptions: …continued
• In order to save Working Memory resources people tend
to construct models that represent explicitly what is true
and not what is false. (Model based on “If it rains there
will be ripples in the Liffey”, not on “If there are no ripples
in the Liffey, it is not raining”. This is the Principle of
truth.
• Problems requiring the construction of mental models are
harder to solve that those requiring only one mental
model, because of the demands on Working Memory
 Deductive Reasoning
Mental Models
Johnson-Laird (1983) had participants arrive at a conclusion
based on premises

When only 1 model was required 78% drew valid conclusion

When 2 models were required, 29% drew valid conclusion

When 3 models were required 13% drew valid conclusion

 Deductive Reasoning
Mental Models
Laird & Goldvarg (1997) showed that participants over-
emphasis on the principal of truth led to 99% of
participants making the wrong conclusion in a hand of
cards task
Only one of the following premises is true about a particular
hand of cards
(1) There is a king in the hand or there is an ace, or both
(2) There is a queen in the hand or there is an ace, or both
(3) There is a jack in the hand or there is a 10, or both
Is it possible that there is an ace in the hand?
 Deductive Reasoning
Mental Models
Mental model theory accounts for a wide range of problems
and most of its predictions have been confirmed
experimentally
Mental models do not require the existence of ‘mental logic’,
but rather the theory requires nothing more than the
normal processes of comprehension
This leads to the argument that the tests used in reasoning
experiments can be used to draw conclusions on
everyday real life reasoning
 Deductive Reasoning
Mental Models
Limitations:
More detail needed as to how mental models are formed and
which knowledge we bring forth and use
Ford (1995) Identified spatial and verbal reasoners.
“Neither…could be said to provide evidence of
developing mental models that are structural analogues
of the world”
 Reading

Eyesenck & Keane Chapters 15 & 16

Sternberg Chapter 12

Article:
Schrovens, W. & Schaeken, W. (2003) A critique of Oaksford,
Chater, and Larkin's (2000) conditional probability model
of conditional reasoning. Journal of Experimental
Psychology: Learning, Memory, and Cognition. Vol 29(1),
pp. 140-149
 Animal Ethics APA

http://www.apa.org/science/animal2.html