Evolution of social cognition and the cognitive bases of transitive inference in monkeys Open Access

Paxton, Regina Lynn (2012)

Permanent URL: https://etd.library.emory.edu/concern/etds/w95051313?locale=en
Published

Abstract

The social intelligence hypothesis posits that life in complex social groups created
selection pressures that resulted in the evolution of sophisticated cognitive adaptations
specialized to each species' life history. These cognitive abilities may be most readily
engaged and measured under conditions that approximate a species' natural environment.
I validated a novel method for cognitive testing in animals housed in a semi-natural social
group by showing that these subjects learn and perform similarly to laboratory animals on
a battery of cognitive tasks including perceptual discrimination, classification, memory,
and transitive inference (TI). Transitive inference is a cognitive ability theorized to have
been under selection pressures related to learning dominance relations. For example, if
animal A is dominant to B (A>B) and B is dominant to C (B>C), the relation between
animals A and C can be logically inferred ( A>C). TI can also serve non-social purposes;
many species solve nonsocial laboratory TI tasks in which the relations between items are
defined by reinforcement contingencies (if A+B-, and B+C-, then A+C-). However,
because the relations are trained by reinforcement, it is unclear whether non-social TI
performance is controlled by inference or by associative values accrued to individual
stimuli. TI based on associative values would not support learning dominance
relationships, because observation of social interactions is not followed by explicit
reinforcement. Rhesus monkeys were presented with a series of tests to determine the
contributions of associative values and logical inference to TI performance. We found
that associative value can influence TI choice in extreme circumstances, but the inferred
order of stimuli better explains TI choice under normal testing conditions. Measurement
of the associative values of TI stimuli showed that these values did not correspond to
performance on TI tests. Monkeys linked two previously learned 7-item TI lists into one
14 item list after training on just one linking pair, performance that cannot be explained
by associative value alone. The viability of socially housed animals as cognitive test
subjects will allow for future studies that examine the extent to which monkeys use TI to
learn natural dominance hierarchies.

Table of Contents

Table of Contents

Chapter 1.

Introduction

1

The social intelligence hypothesis

Social hierarchies

Transitive inference

Use of transitive inference to learn social hierarchies

Chapter 2.

Comparison of automated cognitive test performance by monkeys in laboratory and large semi-natural social groups

8

Introduction

8

Experiment 1- Perceptual discrimination

15

Experiment 2- Perceptual classification

20

Experiment 3- Transitive inference

25

Experiment 4- Delayed matching to sample

30

General Discussion

35

Chapter 3.

Mechanisms underlying transitive inference performance in rhesus macaques

37

Introduction

37

Experiment 1- Transitive inference

40

Experiment 2- Measurement of associative values

47

Experiment 3. Manipulation of associative values

55

Experiment 4- List linking

61

General discussion

68

Chapter 4.

Summary, synthesis, and future directions

71

Overall discussion of findings

Contributions of findings towards the social intelligence hypothesis

Future directions

References

78

Figures and Tables

Chapter 1.

Introduction

1

Figure 1.

Matrilines in rhesus monkey groups

4

Chapter 1.

Comparison of automated cognitive test performance by monkeys in laboratory and large semi-natural social groups

8

Figure 1.

The testing system used by the Laboratory monkeys

11

Figure 2.

The four testing stations used by the Field station subjects

13

Table 1.

Demographics of subjects working at the Field station

15

Figure 3.

Trial progression during size and brightness discriminations

17

Figure 4.

Stimuli used in size and brightness discriminations

18

Figure 5.

Mean proportion correct on the two criterion sessions

19

Figure 6.

Example trial from perceptual classification

23

Figure 7.

Mean proportion correct during the one transfer session

24

Figure 8.

Training and test pairs used in transitive inference

28

Figure 9.

Symbolic distance effect

30

Figure 10.

Example trial from the memory experiment

33

Figure 11.

Performance on delayed matching to sample

34

Chapter 2.

Mechanisms underlying transitive inference performance in rhesus macaques

37

Table 1.

Training and test phases in Experiment 1

43

Figure 1.

Symbolic distance effect

45

Figure 2.

Symbolic distance effect by first item

46

Table 2.

Results of RM ANOVA for symbolic distance effect by first item

47

Figure 3.

Correlation between object discrimination performance and associative value difference score.

53

Figure 4.

Correlation between TI choice and associative value difference score and symbolic distance

54

Figure 5.

Errors to criterion in the congruent and incongruent conditions

59

Figure 6.

Premise pair performance in the first intermixed session

65

Figure 7.

Symbolic distance effect across between list pairs

67

About this Dissertation

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
School
Department
Degree
Submission
Language
  • English
Research Field
Keyword
Committee Chair / Thesis Advisor
Committee Members
Last modified

Primary PDF

Supplemental Files