Savage-Rumbaugh

Spontaneous symbol acquisition and communicative use by Pygmy Chimpanzees.

Brief Background:

Language is made of symbols (sounds and words) which we learn can have meanings and can be combined in infinite ways to produce an infinite number of sentences.

Various research studies have tried to teach primates to use human language - Gardner & Gardner 1969 in order to examine whether language is learner or innate - part of the nature-nurture debate.

Learning Theory - B.F. Skinner: when children produce sounds that are like language they are rewarded - sounds are then shaped by the rewards as the parents repeat the word correctly e.g. repeating "Dadada" as "Dad" or "Daddy"

Language Acquisition Device - Chomsky: children have an innate ability to learn language as we are born with an understanding of the rules of language.

Aim: To study the language acquisition of two pygmy chimpanzees compared with two common chimpanzees.

Research question: Can Kanzi learn symbolic language without training in the same way children do?

Pp's

• 2 Pygmy chimpanzees: Kanzi (male) and Mulika (female)
  
• 2 Common chimpanzees: Austin and Sherman

Method:

• Design - Quasi-experiment, longitudinal case study report

Setting:

• Language Research Centre, Georgia State University
• Outdoor environment - 55 acres of wood

IV - species of chimpanzee - Pygmy or Common
DV - language acquisition

The visual symbol system:

Indoors - battery powered keyboard with geometric symbols that brighten when touched, then speech synthesiser 'speaks' the word.

Outside - chimps pointed to symbols on laminated keyboards. They were also exposed to spoken words and approximately 100 ASL (American Sign Language) gestures.

Each symbol was called a lexigram.

Data recording: Kanzi's (from the age of 2 and a half) and Mulika's utterances for 17 months were recorded completely using an automatic computerised record from the keyboard. Two observers (improves inter-rater reliability) made notes and 4.5 hours were recorded on a videotape.

When assesing Kanzi's symbol use utterances were categorised :

• Correct/incorrect
• Spontaneous/imitated
• Structured (responds to question)
• Behavioural concordance (agreement)

Kanzi produced 2540 correct combinations of symbols:

• More than 80% of Kanzi's single-word and combination utterances were spontaneous (learnt with no help).
• Only 11% (265 combinations) were prompted/imitated

The pygmy chimps learnt more easily than the common chimps and made spontaneous associations between lexigrams and objects whereas the common chimps did not.

The pygmy chimp's understanding of spoken English was not context dependent whereas the common chimps needed contextual.

- Pygmy chimpanzees began to use gestures spontaneously at a younger age (Mulika at 12 months) than the common chimpanzees; and their gestures were more explicit.

- Pygmy chimpanzees did better on formal tests; understood the lexigrams more easily; understood more spoken english.

- Kanzi was able to refer to requests involving others. In the 'blind' test in 55 acre forst, Kanzi was able to lead the experimentor out of the forest by correctly using photos and lexigrams.

.........Matching symbol - English......Matching photo - English ....Matching photo - symbol

Kanzi ...................65/66..................... 56/59 .........................................55/59
Mulika ..................41/42 .....................36/41 .........................................41/42
Austin ..................not tested............... 3/30 ..........................................30/30
Sherman...............not tested ...............2/30 ...........................................30/30

Kanzi learned to use the symbols spontaneously.

Bonbos (pygmy) chimps can understand spoken english whereas common chimps can't.

Compared to common chimps, pygmy chimpanzees appear to be able to learn and use language more like a human child.

Kanzi and the language universals:

• Semanticity ( the use of symbols to mean objects or actions) Yes
• Creativity - Yes
• Structure dependence - Yes
• Generalisation - Yes
• Displacement (refers to things that are distant in time and space) Yes
  

Strengths:

• Researchers implemented many controls
• The data gathered from formal tests is more likely to be valid and reliable
• Longitudinal study - gathers in-depth data
• High ecological validity
• Inter-rater reliability - Observers agreed 100% on 37 utterances

Weaknesses:

• Ethical issues involved - studying animals, the right to withdraw, informed consent, de-briefing
• Uncontrolled environments
• Small sample - cannot be generalised across whole/wider population
• Not all features of language were displayed e.g. Arbitrariness of the symbols
• Researcher can be biased