This preview shows page 1. Sign up to view the full content.
Unformatted text preview: PSYC 103
Communication Monday, March 14 Warren Lecture Hall Room 2005 8-9:20pm
Final: Friday March 18th 8AM Different calls elicit different behaviors in receivers Do vervet calls have meaning?
Measure habituation/“dishabituation” between different calls that are used
in the same context (i.e. different sound - same meaning) Transfer across call types Individual recognition;
No transfer Do vervet calls have meaning?
Habituation transfers across calls from different species! Duration of response Call listed first is “habituation”
Second is “control” & “test” Vervets treat starling alarm calls similarly to vervet alarms
Show audience effect (like chickens)
Suggests an internal representation of the referent Human Language and Animal Communication
Limited vs unbounded signal sets: Most animals communicate about
a limited range of items (food, sex, predators, etc) using a fixed and
usually small set signals.
Reference and situational freedom: Animal communication often
reflects internal states. Other signals are referential, i.e.they are reliably
given in the presence of an object and not under other conditions (bees
chickens, monkeys). Humans can discuss events or items that have
occurred in the past or may occur in the future. This is called situational
freedom or displacement.
Intentionality: Humans generally use language with the intent of
informing (e.g. clarification), by suiting our communication to the
audience and modifying it based on whether or not it is having its
intended effects. There is no solid evidence that other animals do this. Development of communication
Communication and intention
Is animal communication made with the intention of passing on knowledge to the
- If communication is intentional, animals should signal more when near conspecifics:
Marler et al (1990) jungle fowl
Cheney & Seyfarth (1990)
vervet monkeys Show an audience effect - The audience effect may, however, reflect nothing more than an innate disposition to
make calls in the presence of conspecifics.
Hostetter, Cantero & Hopkins (2001):
- Trainers held out a banana to chimps that was just out of reach
- If trainer facing chimp: chimp held out its hand
- If trainer had back to chimp: chimp vocalized
- A consequence of trial and error learning? 6 (iii) A definition of language
Criteria that communication must meet for it to be considered
(1) Arbitrariness of units
- The units of communication (e.g. words) must be arbitrarily related to the event they
- Displays of aggression or submission in dogs do not count. Different alarm calls by vervet
monkeys do count (2) Semanticity
- The units of communication must have a meaning, or evoke a representation of an
- Alarm calls of the vervet monkey fulfil this criterion (3) Displacement
The units of communication must be able to relate to an object (or event) that is spatially or
temporally remote. Dance of the honey-bee fulfils this criterion (4) Productivity
A large number of meaningful utterances can be created from a limited vocabulary (iv) Language and cooperation
- Two dolphins situated in adjacent pools (arranged so that they could not see each other)
- Steady and flashing lights presented to female dolphin
During steady light: male dolphin had to press panel A for both dolphins to be fed
During flashing light: male dolphin had to press panel B for both dolphins to be fed
- Male dolphin mastered this discrimination. Success was accompanied by female dolphin
making different sounds during the different lights
- Is this cooperation?
- Maybe, but Boakes & Gaertner (1977) say results can be explained in terms of
- Random sounds made by the female during the lights could be accompanied by a
chance correct response by the male
- This would serve to reward the female for making that sound during that light
- And reward the male for making a particular response during a particular sound Can an ape create a sentence?
(1) Training to speak
- Furness (1916):
Spent many hours attempting to teach an orang-utan to speak English.
Orang-utan only managed: “Papa”, “cup” and “th”
- Hayes (1952, 1961):
Raised a chimp, Vicki, as if it were a child
Chimp only managed “Mama”, “Papa”, “cup” and “up”
- Duchin (1990): These limited successes are not a surprise
Chimp is constrained by its inability to move its
tongue to make [a], [i] and [u] sounds The adult vocal tract of a human (left-hand side) and a chimpanzee (right-hand side)
(adapted from Lieberman, 1975). Can an ape create a sentence?
(A) Training methods
(2) Plastic tokens
- Premack (1971, 1976)
- Created an artificial language in which plastic objects represented words
- Chimp called Sarah learned about 130 words
- Sarah praised for putting objects in correct order Can an ape create a sentence?
Training methods American Sign Language
Gardner & Gardner (1969)
- Trained a chimp, Washoe, to communicate with her hands
Successfully learned about 132 different words:
Nouns (“Sweet”, “Key”)
Pronouns (“Me”, “you”)
Verbs (“tickle”, “open”) -
- Washoe could combine
these words to form
e.g. “you tickle me” Another strategy for teaching a chimp a language was to to place an infant in the care of Washoe.
The infant died and the trainer signed “Baby gone, baby finished”. Washoe…
“…dropped her arms to her lap and she broke eye contact and slowly moved
away to a corner of the cage. She was demonstrating all the clinical signs of
depression. She continued for the next several days to isolate herself from any
interactions with the humans and her signing dropped off to almost nothing.
Her eyes appeared to be vacant or distant” 11 Can an ape create a sentence?
Lexigrams Can an ape create a sentence?
- Symbols (lexigrams) displayed on a keyboard that was connected to a computer
Pressing a key projected the symbol onto a screen. Sentences created by pressing keys in
Sentence had to follow grammatical rules
Lana (a common chimpanzee) learned to press lexigrams for “Please machine give juice” Rumbaugh & Savage- Rumbaugh (1994)
- Tried to teach a bonobo chimp, Matata, to use lexigrams
Matata’s adopted son, Kanzi, observed this training and spontaneously began using
lexigrams Can an ape create a sentence?
- Kanzi given intensive training with lexigrams which were now associated with human
speech - Could respond to commands such as “Get object X from location Y” (even if X was right in
front of him) - Concluded that Kanzi had comprehension skills of a 2-year-old child - Early onset of training seems to be important - But, bonobo had a better vocabulary than the common chimpanzee Can an ape create a sentence?
Assessment of language training with apes
- Discreteness and displacement – good evidence for these two criteria.
Savage-Rumbaugh et al. (1983): Tested 2 chimps, Austin and Sherman, who were
trained with methods used by Rumbaugh (1977), i.e. discreteness.
Both could request food that was currently out of sight, some distance away: spatial and
temporal displacement - Semanticity
Savage-Rumbaugh et al. (1988): Austin and Sherman shown a lexigram that
represented an object
Correctly retrieved this object from a box of many objects
“Meaning” (and hence semantics) of lexigram appreciated
But not always the case. A chimp called Nim capable of signing “apple” and “banana” but
could not respond correctly when given both fruit and asked by the trainer to hand over the
apple (Savage-Rumbaugh, 1984) (v) Can an ape create a sentence?
(B) Assessment of language training with apes
- Sentence comprehension: Can chimps understand sentences on the basis of their
Savage Rumbaugh et al. (1993): Showed that Kanzi could respond appropriately to
the following instructions:
“Pour the lemonade in the Coke” vs “Pour the Coke in the lemonade” - Sentence production
Fouts (1975): Reported that a trainer placed a doll in Washoe’s cup
Washoe produced the novel sentence “baby in my cup”
But, Terrace (1979) discovered that the trainer pointed first to the doll and then the
cup prior to asking Washoe to sign – Do not have to assume a knowledge of grammar Can an ape create a sentence?
Assessment of language training with apes
Sentence production (continued)
- Terrace (1979): Analyzed the utterances of language-trained apes
- Nim signed almost 20,000 multi-word utterances in 18 months (of 5235 different types)
Many were of limited type though:
Many began with “More”
Many were copied from trainers
No evidence for syntax or grammar Mean length of utterance stayed
at about 1.5 throughout training - Rivas (2005): Replicated many of Terrace’s findings Requirements for learning a language
(1) Language acquisition device
- Chomsky (1972): Humans posses a language acquisition device for generating universal
- Macphail (1982): Language is uniquely human
- But, little known about properties of a “language acquisition device” or if it exists (2) Motivation
- A barrier that may restrict animals is that they be unwilling to communicate
- Spontaneous acts of communication are rare in apes (Savage-Rumbaugh, 1984)
- But, Nim would, in informal settings, produce spontaneous utterances
- Gardner & Gardner (1974): Washoe would sign “quiet” to herself when moving to a
forbidden part of the yard Requirements for learning a language
(3) Sentence production
(a) Need to be able to order of words
- Monkeys can appreciate serial order
(b) Need to be able to use words to represent a category
- Many demonstrations of categorization in pigeons (4) The representation of knowledge
- Sentences convey relational information: e.g. “Kanzi poured the Coke”
- Premack (1983): Language training will only be successful if the animal possesses an
abstract code to represent relationships
- Overall, evidence that animals can encode relations is poor Principles of Behavior
The end ...
View Full Document
This note was uploaded on 03/24/2012 for the course PSYC 103 taught by Professor Pearlberg during the Spring '07 term at UCSD.
- Spring '07