analogy and relational reasoning

Definition: Role-based relational reasoning involves using relational terms such as:
- „bigger than”
- „less than”
- „is equivalent to”
- „is similar to”
- „is identical with”
- „part of”
- „lives in”
- „is friend of”
- „is sister of”
  -Analogy as Example: Analogy serves as a key example of role-based relational reasoning. For instance, there is an analogy formed between A and B.
Key Processes in Analogical Transfer:
- A target situation triggers a retrieval cue for a relevant source analog.
- The establishment of a mapping occurs, involving systematic correspondences to align elements of the source and target.
- Analogical inferences are derived from this mapping.

Four-term Proportional Analogies:
- These take the form A:B::C:D.
- Example: HAND:FINGER::FOOT:?
- The missing term D can be inferred (TOE) based on the relationship defined between A and B for C.
- Here, A:B represents the source analog, while C:D is the target.
Nature of Analogy: Analogy is fundamentally a relationship.
Metaphors as Special Analogies:
- Metaphors represent a unique form of analogy where source and target domains are semantically distant.
- Example: „the evening of life”, where life (target domain) is mapped from the time of day (source domain).
- Understanding Time: Time is often conceptualized through spatial metaphors:
- Examples:
  - „My birthday is fast approaching”
  - „The time for action has arrived.”
- A recurring metaphor is that life is viewed as a journey.

Questions of Interest:
- What functions does human relational reasoning serve?
- What algorithms enable it?
- How is it neurologically implemented? (p. 239)
Computational Goals: Include:
- Forming and evaluating hypotheses
- Problem-solving
- Comprehending new concepts
- Persuasive arguing
Inductive Uncertainty: Reasoners may confront inductive uncertainty; for example, the assumption that all swans are white is an inductive generalization that can be invalidated.

Symbolic Connectionism:
- Represents structured relations within complex neural networks.
- Operates under constraints of limited human-like working memory (p. 240).
Neural Substrate:
- The prefrontal cortex plays a crucial role in facilitating relational reasoning.

Study Overview:
- Glick and Holyoak (1980) utilized an analogy involving a general attempting to capture a fortress from a dictator.
- The general deploys his army in groups across various roads, timing their arrival to capture the fortress effectively.
Analogous Problem: A tumor problem developed by Duncker (1945) is explored:
- A doctor must use rays to destroy a tumor while avoiding healthy tissue damage, as high intensity harms both, and low intensity harms neither.
- The solution parallels the military strategy: instead of one high-intensity ray, multiple low-intensity rays from various directions are applied, collectively achieving a harmful effect at the tumor site while preserving surrounding tissue.

Correct Solutions:
- Without a source analog: 10% of college students solved the tumor problem correctly.
- With the General story used without hints: 20% solved correctly.
- With instruction hints: 75% of college students found the correct solution.
Mapping Importance: Essential for analogical inference; involves identifying corresponding elements between source and target analogs.

Early Presence of Analogical Reasoning: According to Goswami, such reasoning is evident from early infancy.
Understanding Other Minds: The role of analogy includes:
- Analogical Solution to the Problem of Other Minds (J.S. Mill, Bertrand Russell)
- Analogical Theory of Mind (AToM) computational model (Rabkina et al., 2017, 2018)
- „Like-me” Hypothesis (A. Meltzoff): Infants employ themselves as analogies to comprehend others.