Comparing intelligence

Question 1

the debate

Answer 1

2 schools of thought
o Gradual (quantitative) difference between humans and other animals
o Sharp (qualitative) distinction between humans and other animals

Question 2

misapprehensions about evolution

Answer 2

That some species are “more evolved” than others: all species have been evolving for exactly the same amount of time

Evolution necessarily implies “improvement”: clearly complexity has come about through evolution, but evolution doesn’t always produce greater complexity

Evolution leads in the direction of being human-like

Question 3

consequences of misapprehensions

Answer 3

Anthropomorphism: we assume that animal cognition is like human cognition (but not as good)

Anthropocentrism: we interpret “advanced” as meaning “more like us”

Social/political baggage: we place the kind of human we are at the top of the “ladder of nature” – historically had influence

Similar errors made before theory of evolution formulated

Question 4

scala naturae

Answer 4

Cannot assume there is a “scala naturae”, a ladder of nature w/ Amoeba at bottom (least intelligent) and humans at top (most intelligent)

What are alternatives?

What do we know?

Why is concept of ladder of intelligence so appealing? – start with brain size

Question 5

brain comparisons

Answer 5

Bigger brain  higher intelligence?
o Some ev that works for some parts of brain, e.g. hippocampus and spatial memory – e.g. taxi drivers and the knowledge

Bigger animals have bigger brains anyway – doesn’t mean they’re more clever

“(En)cephalisation coefficient”
o EC = brain mass/body mass

Jerison: look for deviations from plot of brain v body mass

New World Monkeys, Old Word Monkeys and Apes follow diff regression line

Ev supports view that there are qual diffs in intelligence between diff species

Ev should be viewed w/ caution – factors determining brain/body ratio complex

Purple = average animal line – average to be able to survive

Lower than average = less intelligence

Amphibians spend lots of time in water – supports body weight – can be bigger for given brain size – may not have anything to do w/ IQ

see notes

Question 6

behavioural correlates of intelligence

Answer 6

Typical suggestion: learning rate – w/fixed task, animals that learn faster must be more intelligent – but contextual variables issue

Commonly used example: Hebb-Williams maze (sequence of T-mazes)

More sophisticated example: not one task but “learning how to learn”
o Successive reversal
o Learning set
o Probability learning

Better = battery of tests, looking for patterns (Bitterman, 1965) – more like approach to IQ

Question 7

examples: serial reversal learning (Mackintosh, 1974)

Answer 7

see notes

On later reversals, animal makes less errors in acquiring discrimination

Rate at which this occurs correlated with intelligence?

Maddingly sheep – use spatial version of task – down tunnel and can turn left/right – reversed – learns and changes direction – learn v. quickly – task may just particularly suit them

Question 8

examples: learning sets (Harlow, 1949)

Answer 8

see notes

On later problems, animal makes less errors in acquiring discrim

In extreme cases makes just 1 error

Can we use rate of acquisition as index of intelligence?

Question 9

learning sets across species

Answer 9

see notes

Case for some correlation between brain size and intelligence quite strong

Dangers

Comparisons across species difficult because hard to specify what optimal conditions for testing given species would be

Problem of contextual variables – e.g. sheep – just task could do well – monkeys v. similar to us

Question 10

Macphail’s null hyp

Answer 10

E.g. 1982, 2000: there are no cog diffs between non-human animals – more refined version: no diffs amongst non-human vertebrates

Only imp diff is emergence of human language – language training may confer special abilities

If we find more sophisticated cognition in e.g. apes than other species, may be because can understand better how to test cognition in species like us  “contextual variables” could be responsible for observed diffs in perf rather than genuine diffs in intelligence

Question 11

the role of contextual variables

Answer 11

Goldfish discrim 1: they fail to learn

Goldfish discrim 2: they learn
o Need reward within certain time window

see notes

Question 12

contextual variables and learning sets

Answer 12

Herman and Arbeit (1973) found that dolphins had difficulty forming learning sets with visual stim but could with auditory stim
How well animal forms learning set may depend on type of stim used to test them

Only valid test would be to compare animals with sensory and effector capabilities

Question 13

ev in support of Macphail

Answer 13

Simple forms of learning (e.g. classical and operant conditioning) take place in same way and rate in all vertebrate species and some invertebrates

Sophisticated forms of learning turn up in invertebrates, e.g. molluscs, arthropods

see notes

Initial response to weak stim strengthened by pairing CS and US

Aplysia initially responds weakly to gentle touch on siphon by withdrawing both gill and siphon, but when paired with aversive shock to tail this response more vigorous – increase in response over and above that seen in sensitisation

Siphon sensory neuron synapses on motor neurons for siphon and gill – connection strength from siphon  gill strengthens during course of conditioning because of fac by interneuron which occurs at same time as pathway activation

Question 14

conditional discrim (Colwill et al., 1988)

Answer 14

Conditional discrim in Aplysia

Found could learn to provide differential responses to same stim in diff contexts

Context 1 was smooth white round bowl w/ lemony seawater

Context 2 was dark grey rectangular container w/ ridges and turbulence (aerator)

Pair prod w/ shock

Other condition = unpaired

Learn context where paired and greater gill withdrawal

see notes

Question 15

learning in honeybees (Giurfa et al.)

Answer 15

Classical and instrumental conditioning

Contextual learning: C1: A+, B-; C2: A-, B+ - conditioning discrims

Categorisation:
o Bilaterally symmetrical v asymmetrical
o Diff types of abstract patterns
o Same v diff (matching and non-matching to sample)
o Neg patterning: A+, B+, AB-… w. olfactory stim – when paired = no reward – rat and humans take long time to learn but bees can learn quickly

Question 16

the comparability problem

Answer 16

How do we make sure everything except learning equiv between animals of diff species?
o Perceptual difficulty
o Mechanical difficulty of response
o Level of motiv
o Incentive value of reward

Bitterman’s suggestion: control by systematic variation of confounding factors

Problem: how can we ever explore all variations?

Another solution is to use comparable species, e.g. pigeons and corvids

Question 17

same v diff discrims

Answer 17

Matching to sample tasks and oddity from sample tasks (Zentall and Hogan, 1974, Wright et al., 1988, and others): responding based on identity/diff if enough exemplars used

Ev: transfer to novel stim on first trial – has been found in chimps, dolphins and corvids – eventually obtained (after some effort) with pigeons – Colombo, Cottle and Frost (2003)

Can we use this task to assess relative intelligence of pigeons and corvids?

Can they transfer same task to different stim?

Question 18

pigeons v corvids (Wilson et al., 1985)

Answer 18

Jackdaws show better transfer to novel matching problem after training on matching w/ diff stim – learn more slowly but show more transfer

Split into 2 groups – matching to sample of particular stim

Other group on conditional discrim – when shown triangle go for blue circle, and when shown square go for yellow circle

Transfer all to same task – e.g. duck and grapes exp

Each line is bird

Pre-trained on MTS show better perf – transfer

Pigeons = no ev of transfer

Pigeons outperf jackdaws – learn task faster

Some ev that corvids (crows, rooks, ravens, magpies and jays) perf better at learning sets than pigeons (Mackintosh, 1988)

see notes

Question 19

recency/familiarity explanations

Answer 19

Solution available to animal in terms of recency with which stim occurred

So – if learn to go for “more recently seen ‘stim – transfer across problems – diff strat that responds to configurations of stim won’t

Question 20

2 diff methods of solving the problem

Answer 20

Either, having studied sample which comes on first:

1. Pick bottom left comparison stim because feels more recent (familiar)
2. Pick bottom left because that’s the right response to configuration of stim

Could be that corvids are biased to use solution 1, whereas pigeons use 2, but does it make the corvids more intelligent?

Ev that pigeons can, after lots training, use solution 1, so might be that recency/familiarity info less salient for them

see notes

Question 21

another case of comparability

Answer 21

Language trained v non-language trained chimps, Premack (1983), Premack and Premack (1983)

Some tasks that language trained chimps succeed on that non-language trained cannot pass

Some tasks that both groups of chimps can do with equal facility

Question 22

same/diff tasks take 2

Answer 22

Language trained chimps could solve tasks, but others couldn’t

Both sets could solve successive same/diff problems, where one object shown and either same/diff object shown – can use recency when come to second object – other tasks = all objects feel recent

see notes

Question 23

analysis

Answer 23

Hard to explain result in terms of contextual variables

Harder to explain in terms of responding based on recency/familiarity

Extended training that language trained chimps received had effect – are they more intelligent as consequence?

Analogical reasoning (Gilan, Premack and Woodruff, 1981)

see notes