interchangeable use of stimuli by monkeys
Our cognitive system has a unique ability to express ideas by different means. To refer to the animal “CAT’ for instance, we can pronounce the word “cat”, we can draw a cat or even write the word “CAT” using letter of the alphabet. All these expressions of the same concept can be used interchangeably. This ability to use different means to refer to the same objects, scenes or concepts implies that their different expressions are processed as functionally equivalent. The development of such equivalence relations is fundamental to form and use symbols, and therefore to our most critical human skill – language. Sidman & Taiby (1982) have proposed that this ability involves simple forms of associative learning that should be shared by nonhuman primates as well. However, a close look at the literature suggests that nonhuman primates have the most important difficulties to process stimuli as equivalent, and this difficulty must limit cognitive abilities which are central to human cognition, such as the ability for referential communication or abstract and symbolic thinking. In this context, Premilang2 aimed at testing if nonhuman primates (baboons) can process stimuli as functionally interchangeable, and to compare human and baboon performance in similar experimental contexts. The goal of this project was to provide information on the origin of human cognition and language evolution.
Behavioral experiments on baboons were conducted at the Primate Cognition and Behavior” (PCB) research platform which is installed within the CNRS primate station in Rousset-sur-Arc . This is a word-unique platform where a social group of baboons can freely interact with computer presenting computerized tasks on touchscreens. Correct responses to the tasks are food rewarded. This research platform allows highly efficient studies, because monkeys can be tested in parallel, with no need to capture them, on task involving many trials. The computerized tasks were also proposed to human subjects, for comparative purposes.
Monkeys were trained with different kinds of associative networks, and we discovered that they had the most important difficulties to process relations as symmetrical. Thus, from an initial training with the A->B relations between to objects, they were unable to infer the symmetrical B->A relation. Our set of experiments explained this phenomenon: contrary to humans, the baboons process in priority the information on the stimulus’ spatial location and their ordinal position, while humans tested in the same experimental context process the relation in priority.
This project opens avenues for a better understanding cognitives differences between humans and nonhuman primates.
Several publications were published in highly visible journals. This project was also at the origin of three Phds on the origin of language, and a large number of presentations in scientific meetings.
Our human cognitive system has a unique ability to express ideas by different means. To refer to the animal “CAT”, for instance, we can pronounce the word “cat”, we can draw a picture of a “cat” or even write the word “CAT” using the letters of the alphabet. All these expressions of the same concept can be used interchangeably. This ability to use different means to refer to the same objects, scene or concepts implies that their different expressions are processed as functionally equivalent. The development of such equivalence relations is fundamental to form and use symbols, and therefore to our most critical human skill - language (e.g., Devany, Hayes, and Nelson, 1986).
The current project deals with what may be considered as a fundamental evolutionary building block of our cognitive abilities. It concerns the ability to form equivalence relations between a variety of stimuli, such as (for humans) spoken words, written words, drawings and much more, processed as functionally equivalent and as interchangeable cognitive units. Sidman and his group (e.g., Sidman and Taiby, 1982) and their followers have suggested that equivalence relations of this type may emerge from relatively simple associative mechanisms, but evidence available so far in the comparative psychology literature suggests serious limitations of animals in that domain. It might be precisely this reduced ability to form equivalence classes that limits their mastering of complex operations that require interchangeability between cognitive units (e.g., high-level categorization, referential communication, symbolic and abstract thinking, etc…). Equivalence class formation therefore deserves a special attention in comparative psychology, because it is a core problem of major theoretical importance to account for the cognitive systems of humans and animals, and their differences.
In PREMILANG2, we will reconsider the issue of equivalence class formation in nonhuman primate using the innovative “Primate Cognition and Behavior” (PCB) platform installed in Rousset. This is a world-unique platform developed by the PI of the current project, in which monkeys living in their enclosure can participate freely to computerized experiments. The PCB platform allows collection of a very large number of trials (tenths of thousands per day), moreover on a large group of animals. It is therefore well suited to assess cognitive phenomena in monkeys that might be based on associative mechanisms, as the equivalence class formation that are usually emerging after repetitions of the associations.
The first line of research will measure the baboon’s ability to form genuine equivalence classes among triads of visual stimuli. Preliminary results presented in this project show the feasibility of this approach in the PCB platform. We will study the extent and the limits of this ability on a group of 20 baboons. The process of equivalence class formation is twined to the process of category formation. PREMILANG2 will test the baboon’s ability to form super-ordinate equivalence classes from the grouping of several independently learned equivalence classes under the influence of common auditory cues. Considering that equivalence classes are often bimodal in humans (as demonstrated by our use of verbal labels to refer to visual objects), cross-modal equivalence classes will also be investigated in PREMILANG2. For the first time in the animal literature, we will test if the presentation of common visual or auditory cues (analog to written or spoken words) for sets of distinct visual cues promotes equivalence class formation, as it does in children (Fulkerson and Haaf, 2003).
This project involves a team of animal and human cognition experts from two partners, the laboratory of cognitive psychology and the CNRS Station de Primatology (Rousset-sur-Arc). Premilang 2 is a resubmission of the Premilang project which has been retained in the short list (but not funded)
Monsieur joel FAGOT (Laboratoire public)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
Help of the ANR 139,958 euros
Beginning and duration of the scientific project: December 2013 - 36 Months