Numbers and Counting Studies

Emmerton et al (1997) - Relative numerosity discrimination

• trained pigeon to discriminate between ‘few’ and ‘many’ items

• the bird could still discriminate between the items when they were displayed so it wasn’t a light / dark discrimination

• there was evidence of 3,4,5 being treated on a continuum, suggesting knowledge transfer, and that the pigeons had some idea of what ‘many’ and ‘few’ mean

Koehler (1913) - concept of absolute number

Jakob the raven

• could choose a pot with 5 spots out of an array even when the size of the pot changed 50x

Matsuzawa (1985) - concept of absolute number

• Ai the chimp achieved above 90% accuracy selecting keys 1-6 when she was shown arrays of red pencils

• Researchers argued she could transfer the ability to different arrays of the item, but the number was often confounded with other factors like time and space

Pepperberg (1994) - concept of absolute number

African grey parrot could count different components out of the same array

Meck and Church (1983) - concept of absolute number

Rats

• 2 pulses of white noise - rewarded for left lever response

• 8 pulses of white noise - rewarded for right lever response

• confounded by noise?

• when noise stimuli altered so both presentations lasted 4 seconds, rats could still discriminate

• evidence against perceptual matching

Biro and Matsuzawa (2000) - ability to count

• Ai was trained to touch Arabic numerals in an ascending order

• This was just rote learning of a particular S-R sequence, no actual understanding of the quantitative relation between numbers was required

Dacje and Stinivasan (2008) - ability to count

Bees

• Trained - sucrose rewards at 1st-5th landmark

• Distance between landmarks varied on each trial in training

• Then tested without reward

  • 1 - same landmarks (yellow stripes)

  • 2 - different landmarks (yellow disks)

  • 3 - could only see one landmark at a time

• Evidence showed bees can count to 4, and transfer number to novel items

• They also don’t integrate the whole image

Brannon and Terrace (2000) - ability to count

Chimps

• Order arrays of 1-4 items in ascending, descending, or random order

• They could learn the ascending and descending but not the random order

• Tested with novel displays of 5-9 items

• Chimps taught an ascending order could generalise immediately to higher number

• The ones taught in descending order needed further training

• Implies limited understanding of quantities

Boysen and Berntson (1989) - ability to do arithmetic

Chimp

• Trained to label arrays with counters, then Arabic numerals

• Performed well when the items were swapped for everyday objects

• Did extensive training with the numbers 0-4

• In the final test, oranges were hidden in the lab in any of 3 hiding places. The chimps had to find them all, then pick the Arabic numeral representng the sum of all the hidden oranges

• After 12 training sessions of about 20 trials per session, performing was at about 85% correct

• Chimps could perform accurately when the experimenters hid cards with numbers written on them rather than oranges - they performed above chance right away, implying an understanding of interval scale

Boysen and Bertson (1995) - contextual variables

• Chimp A had to choose between 2 amounts of gumdrops or chocolate covered peanuts

• Chimp A’s choice was then given to B, and A got the unchosen one

• A should choose the smaller quantity to get the larger quantity, but was unable to solve the task

• Animals showed considerable behavioural distress over the inequitable distribution

• If the candy was substituded by numerals, then they were fine

• Is it evidence they can count, or can they just not resist the treat?