- “Instinct and learning are mutually exclusive factors in the control of animal behaviour.” Discuss, using examples from birds and primates
This is most closely related to week 1 and week 2. The theme in Week 2, as in Week 1, was the interaction of instinct and experience. The standard conclusion is that instinct and experience, even in very young animals, are not mutually exclusive, but may work “hand in hand” (Bateson, 1990, 2000 and Shettleworth, 1993)
Week 2, on early social development in birds and primates, is more relevant than any other week, but the conclusion referred also to the week 1 question.
Further material on imprinting was on page 9 of the week 2 handout and some further material on rhesus monkeys was available on the intranet.
[Week 1 handout | pdf version]
[Week 2 handout | pdf version]
- To what extent is temporal contiguity the critical determinant of association formation? Discuss in the context of evidence from classical conditioning experiments.
This is similar to the essay for the second lecture in week 3. There was a section covering the question at the bottom of page 17 of the handout. The contrast is between continguity (the pairing together of a conditioned stimulus with the UCS, and contingency, which broadly corresponds to the meaningfullness of the pairings. The three key experimental areas which suggest that contiguity is NOT the only factor in determining the strength of typical Pavlovian conditioned responses are conditional probability experiments, blocking experiments and findings relating to the relevance of the conditioned stimuli used to the nature of the unconditioned stimulus, as in the phenomena of taste-aversion learning.
[Week 3 handout | pdf version]
- How far was Thorndike correct in assuming that instrumental learning can be fully accounted for by the strengthening of connections between preceeding stimuli and successful responses?
The most direct answer is based on page 8 of the handout for week 6, in which it is emphasised that experimental evidence for 'expectancies' and 'cognitive maps', as proposed by Tolman (1932, 1948), and reward re-evaluation experiments, point to additional more cogntive factors being involved in instrumental learning. It is possible to mention that the Thordikean rule may apply to simple nervous systems (e.g. Aplysia) and that connectionists (Week 12) and some neurobiologists (Kandel, 2001; Kandel & Pittenger 1999) still suggest that even human cognition and memory need to be explained by the strengthening of connections at the neural level. Lieberman (2000) has suggested the basic assumptions of modern connectionism are virtually identical to those made by Pavlov, and exactly the same thing could be said about Thorndike’s belief in the importance of the “frequency of connections”.
[Week 6 handout | pdf version]
- Compare and contrast classical and instrumental conditioning.
This was the topic for the second lecture of week 6. There were tables on page 18 of the handout listing “Contrasting features” and “Shared features” of the two kinds of conditioning. The omission procedure was identified as one of the key experimental procedures for making a contrast. There is still an argument about whether these kinds of learning reflect “One process or two?” (Lieberman, 2000; page 442-447)
[Week 6 handout | pdf version]
- How is conditioned fear relevant to the understanding of the development and treatment of anxiety disorders?
This was considered in some detail in week 7 and to some extent in week 6. The recent references of Bouton et al, (2001) and Ohman and Mineka (2001) are included again in the week 13 handout.
There are several articles by Brewin which are relevant to this question, most recently Brewin (2001) and also Brewin (2006)
[Week 7 handout | pdf version]
- Consider the evidence for visual pattern recognition and categorization in animal learning experiments.
This was the essay for both lectures in week 8. There is sufficient material in either one of the lectures to provide an
examination answer. E.g. the first lecture was mainly spent in covering 3 papers about the ability of human participants and rhesus monkeys to correctly categorise large numbers of visual images.
The second lecture concentrated on similar experiments using pigeons, in which the stimuli were letters of the alphabet (as discussed in pp 291-3 of Walker, 1987), line drawings or paintings. Although there may be too much material in both lectures to fit into one examinanation answer, it is better to briefly allude to the material which is not covered.
This essay could also be answered by drawing on the material in pages 502-519 of Lieberman (2000) which ends up with the “Neural Network” solution or by using the material in Chapter 11, "Concept Learning", in the book by Roberts (1998).
[Week 8 handout | pdf version]
- Consider experimental evidence and the theoretical background for the distinction between memory for recent events and response learning in animal behaviour?
Much of Week 9 was devoted to evidence that the hippocampus is used, even in rats, for storage of “episodic-like” material which could be used in answering “where, what and when?” questions i.e. which is different from response learning (There was a quote from Morris (2001) at the top of p. 17 of the handout. The diagram from Squire and Zola (1996) emphazised the distinction between procedural and episodic memory. The radial maze, and memory of hoarded food, are two of the experimental prodedures used to test episodic memory. Plus DTMS. There are many papers on food recovery and the title of the review by Griffiths, Clayton and Dickinson (1999) is obviously relevant.
However Kandel (2001; Kandel and Pittenger, 1999) would say that the molecular biology of memory storage may be similar for response learning and other kinds of memory process. Similarly connectionists in week 12
[Week 9 handout | pdf version]
- ‘Apes cannot be taught language, but there is evidence that they
have special abilities in the areas of social learning,
imitation, and self-recognition.’ Discuss
This covers both week 10 and week 11, but the emphasis is very much more on week 11, with evidence that chimpanzees show self-recognition, social learning of tool using in the wild, etc. The Tomasello references are especially relevant.
Tomasello and Rakoczy (2003) have a short section on chimpanzees, and this article is a useful recent example of the third position on chimpanzee cognition, that chimpanzees have some above average knowledge of psychological states in conspecifics, although clearly this paper emphasises that human social cognition is unique to humans.
Tomasello, Call and Hare (2003) is about evidence that subordinate chimpanzees remember something about what a competing dominant individual has or has not seen.
Two substantial papers from Povinelli's laboratory provide evidence for self-recognition in chimpanzees: Povinelli et al., (1993)
and Povinelli et al., (1997) A review by Boysen and Himes (1999) surveyed the evidence in relevant areas. The textbook reference is Chapter 12 in Roberts (1998).
[Week 10 handout | pdf version]
[Week 11 handout | pdf version]
- Thorndike called himself a connectionist — is this just a co-
incidence, or can comparisons be made between modern accounts of
neural networks and previous theories of animal learning?
This was the theme for the Week 12 lectures. There was something in roughly the same area in 2001 and 2003 (e.g. “Compare and contrast the types of learning observed in animals and the types of learning made use of in connectionist modelling”).
The material for this week (week 13 see conclusion) can be used to criticise modern connectionist claims.
Similarities and differences betwen modern connectionism and associate theories of animals learning were listed in some detail in the Easter handout (these lists were not repeated in the week 12 handout, but the week 12 handout covers the basic points). One of the differences referred to in the Easter handout was the relative absence of constructs corresponding to motivation and emotion modern connectionist theories and in particular the relative lack of use of reinforcement learning mechanisms, but this has changed somewhat in the last 6 or 7 years as there is greater interest in the use of reward mechanisms both in theoretical models of human behaviour and in brain imaging studies.
[Week 12 handout | pdf version]
I will attempt to mark and comment on essays
within 2 weeks of receiving them.
Essays can either be “practice exam” versions, hand-written
in about an hour, or longer versions, possibly word-processed.
March 15th 2007