Following the seminal structure mapping theory by Dedre Gentner, the process of mapping the corresponding structures of relations defining two analogs has been understood as a key component of analogy making. However, not without a merit, in recent years some semantic, pragmatic, and perceptual aspects of analogy mapping attracted primary attention of analogy researchers. For almost a decade, our team have been re-focusing on relational structure mapping, investigating its potential mechanisms across various analogy tasks, both abstract (semantically-lean) and more concrete (semantically-rich), using diverse methods (behavioral, correlational, eye-tracking, EEG). I will present the overview of our main findings. They suggest that structure mapping (1) consists of an incremental construction of the ultimate mental representation, (2) which strongly depends on working memory resources and reasoning ability, (3) even if as little as a single trivial relation needs to be represented mentally. The effective mapping (4) is related to the slowest brain rhythm – the delta band (around 2-3 Hz) – suggesting its highly integrative nature. Finally, we have developed a new task – Graph Mapping – which involves pure mapping of two explicit relational structures. This task allows for precise investigation and manipulation of the mapping process in experiments, as well as is one of the best proxies of individual differences in reasoning ability. Structure mapping is as crucial to analogy as Gentner advocated, and perhaps it is crucial to cognition in general.

Building conceptual abstractions from sensory information and then reasoning about them is central to human intelligence. Abstract reasoning both relies on, and is facilitated by, our ability to make analogies about concepts from known domains to novel domains. Structure Mapping Theory of human analogical reasoning posits that analogical mappings rely on (higher-order) relations and not on the sensory content of the domain. This enables humans to reason systematically about novel domains, a problem with which machine learning (ML) models tend to struggle. We introduce a two-stage neural net framework, which we label Neural Structure Mapping (NSM), to learn visual analogies from Raven's Progressive Matrices, an abstract visual reasoning test of fluid intelligence. Our framework uses (1) a multi-task visual relationship encoder to extract constituent concepts from raw visual input in the source domain, and (2) a neural module net analogy inference engine to reason compositionally about the inferred relation in the target domain. Our NSM approach (a) isolates the relational structure from the source domain with high accuracy, and (b) successfully utilizes this structure for analogical reasoning in the target domain.

Human cognitive abilities seem qualitatively different from the cognitive abilities of other primates, a difference Penn, Holyoak, and Povinelli (2008) attribute to role-based relational reasoning—inferences and generalizations based on the relational roles to which objects (and other relations) are bound, rather than just the features of the objects themselves. Role-based relational reasoning depends on the ability to dynamically bind arguments to relational roles. But dynamic binding cannot be sufficient for relational thinking: Some non-human animals solve the dynamic binding problem, at least in some domains; and many non-human species generalize affordances to completely novel objects and scenes, a kind of universal generalization that likely depends on dynamic binding. If they can solve the dynamic binding problem, then why can they not reason about relations? What are they missing? I will present simulations with the LISA model of analogical reasoning (Hummel & Holyoak, 1997, 2003) suggesting that the missing pieces are multi-role integration (the capacity to combine multiple role bindings into complete relations) and structure mapping (the capacity to map different systems of role bindings onto one another). When LISA is deprived of either of these capacities, it can still generalize affordances universally, but it cannot reason symbolically; granted both abilities, LISA enjoys the full power of relational (symbolic) thought. I speculate that one reason it may have taken relational reasoning so long to evolve is that it required evolution to solve both problems simultaneously, since neither multi-role integration nor structure mapping appears to confer any adaptive advantage over simple role binding on its own.