Chapter 15

The degree to which an organism discriminates between stimuli is inversely related to the degree of generalization:
- The more an organism discriminates between stimuli, the less it generalizes between them.
- Conversely, the less an organism discriminates between stimuli, the more it generalizes between them.
Suggestion: It may not be necessary to propose the existence of two separate processes to explain behavioral outcomes.

One argument for maintaining the concepts of both discrimination and generalization is that together they can provide insight into the process of conceptualization.

Example illustrating conceptual behavior:
- When a child learns to say “shoe” for various members of a shoe category and “book” for different examples of books, we observe conceptual behavior.
Keller and Schoenfeld posited that conceptual behavior is not necessarily verbal nor restricted to humans.

Evidence supporting the notion that conceptual behavior is neither purely verbal nor unique to humans includes:
- Pigeons can discriminate the presence versus absence of objects concerning natural concepts.
- Additionally, pigeons can discriminate objects that belong to multiple natural concepts.

Research by Herrnstein and colleagues involved:
- Pigeons presented with 80 color slides, consisting of:
- 40 slides that included examples of a basic or natural concept (designated S+).
- 40 slides without examples of a basic or natural concept (designated S-).
- Pecking at an S+ slide results in food rewards, while pecking at an S- slide does not.

The concepts that were classified for the pigeons included:
- trees, water, fish, photos of specific persons.
The nature of these concepts lacked clearly defined features guiding the pigeons' pecking responses.
Pigeons demonstrated rapid mastery of discrimination tasks and successfully applied learned behaviors to brand new slides.

An experiment by Wasserman and colleagues involved training pigeons using various color slides based on multiple natural concepts.
- The pigeons were required to categorize which of four natural concepts they perceived by pecking one of four circular keys surrounding a square viewing screen.

The performance of pigeons in learning tasks was evaluated:
- Percent Correct response over sessions indicated learning progress.
- As examples per category increase, human performance reflects a slowed learning process while enhancing transfer to new exemplars.

The correlation between the number of examples per category and the percent correct response was documented:
- A detailed graph showcased the learning outcomes across varying examples (e.g., 25 versus 100 examples).

Questions explored regarding repetition include:
- What is the effect of repetition on the categorization process?
- Can pigeons learn without any repetition?
- Is facilitation observed in learning due to repetition?
Results indicated:
- Pigeons can indeed learn without repetition, yet repetition does facilitate the learning process.

The distinction between true categories versus pseudo categories was analyzed:
- An investigation into whether categories given to pigeons were arbitrary.
- True category learning should manifest faster than pseudo category learning if these categories are not arbitrary.
Data from the experiments showed:
- True Category Discrimination: consistently resulted in higher correct performance than Pseudocategory Discrimination, suggesting non-arbitrariness in categorization.

Conducted experiments featured arrangements with multiple natural concepts:
- Involved schedules with various instances of reinforcement (S+ and S-).
- 43% of all errors made were S- from the same category as the S+.

These findings confirmed a tendency toward greater within-category generalization than between-category generalization among pigeons.
This indicates that pigeons are capable of understanding natural or basic concepts similar to human concept comprehension.
The primary stimulus generalization theory is put forward as the best explanation for basic-level categorization phenomena.

Different stimuli linked to the same context or response can be categorized together, despite perceptual dissimilarity (e.g., tables and rugs categorized as “furniture”).
Such collections based on perceived differences yet shared contexts yield higher-order (superordinate) categories.

Experiment framework allowing assessment of higher-order category learning was implemented:
- Utilizing a three-step design for comparisons between pigeons and preschool-aged children to evaluate the learning capabilities across species.

The matching-to-sample methodology(TMS) was employed to determine if pigeons can learn abstract concepts:
- Performance results indicated that pigeons were capable of same-different learning tasks.

Transfer success from Training Set 1 to subsequent sets was assessed:
- While chimpanzees exhibited excellent transfer, pigeons showed poorer transfers unless higher sample response requirements were implemented.

The evidence suggests that pigeons can acquire abstract concepts through same-different discrimination tasks.
Robust findings across both simultaneous arrays and successive lists affirm that animals can form abstract concepts.
Overall evidence shows that conceptual behavior may be consistent across species, challenging the divide between human and non-human behavior regarding cognition.