Dillon’s work, recently published in the esteemed journal Trends in Cognitive Sciences, introduces what she calls the “Wanderers Hypothesis for Geometry.” This framework postulates that the foundations of geometric understanding stem from navigation-like cognitive mechanisms that evolved to facilitate spatial orientation and route planning in both humans and non-human animals. Unlike the traditional “language-of-thought” hypothesis, which posits that complex mental languages underpin uniquely human faculties like mathematics and music, Dillon argues that the competencies necessary for geometric reasoning are evolutionary products of spatial awareness systems. These systems allow creatures as diverse as rats, chickens, and fish to navigate their environments effectively, demonstrating an awareness of fundamental geometric concepts such as distance, direction, and shape, albeit without full adherence to formal Euclidean rules.

Exploring decades of experimental research, Dillon elucidates how animals engage in mental simulations of navigation, planning routes through spatial landmarks without explicit training in geometry. Such behaviors indicate that an approximate, evolved form of geometric cognition exists beyond the human species. This finding disrupts the conventional assumption that geometry is an exclusive product of human symbolic thought. Humans, she explains, are not unique because they possess a special geometric module; rather, they are distinguished by their capacity to incorporate language into their cognitive toolkit. Language, in Dillon’s view, transforms these spatial navigation processes, enabling humans to manipulate geometric concepts abstractly and solve geometry problems internally, without relying on direct physical exploration.

The hypothesis further gains credence when considering infant cognition. Dillon’s recent studies reveal that even preverbal babies demonstrate an intuitive grasp of spatial parameters, suggesting that fundamental geometric understanding precedes formal education and linguistic development. Intriguingly, these abilities often outstrip current AI systems in tasks requiring commonsense psychological reasoning, highlighting the sophistication and evolutionary depth of these navigation-based cognitive faculties. Thus, the mental operations underlying geometric reasoning appear to be rooted in inherited spatial navigation mechanisms that are both ancient and widely shared among vertebrates.

Dillon’s analysis gains additional power by integrating comparative studies across diverse cultures and developmental stages. Evidence shows that children and adults worldwide engage with geometric concepts in ways that align with navigation-based mental processing, rather than relying solely on abstract mathematical instruction. The cross-cultural consistency of geometric perception and learning undermines the notion that formal Euclidean geometry arises exclusively from culturally transmitted symbolic thought. Instead, it supports the idea that a more primitive, approximate geometric competence grounded in space-finding and route-planning instincts serves as a universal cognitive substrate.

Additionally, Dillon addresses cutting-edge developments in artificial intelligence, particularly efforts like Google DeepMind’s AlphaGeometry project, which attempts to replicate human geometric reasoning. These AI systems, while proficient in many respects, still struggle to emulate the flexible, navigation-based cognitive strategies that animals and humans naturally exhibit. This limitation underscores the significance of evolutionary inherited processes as the foundation rather than formal symbolic geometry alone. Dillon’s synthesis suggests that AI might benefit from models that better mimic biological spatial cognition, rather than relying solely on classical mathematical formalism.

The implications of Dillon’s Wanderers Hypothesis extend beyond cognitive psychology and neuroscience; they resonate within evolutionary biology and philosophy of mind. By framing geometric understanding as an outgrowth of ancient navigation faculties, her work illuminates how complex abstract thinking might arise from more fundamental, embodied cognitive processes. It also invites a reexamination of what truly distinguishes human intelligence—shifting the emphasis from specialized mental modules to the transformative power of language. Language, she argues, uniquely grants humans the ability to mentally traverse geometric space without physical movement, enabling problem solving, abstract thought, and cultural transmission of mathematical knowledge.

Moreover, Dillon’s theory offers a cohesive explanation for how geometry could evolve across species with differing ecological demands and navigational needs. Animals inhabiting environments requiring precise spatial mapping would benefit from refined geometric cognition bolstered by navigation mechanisms. This underlying commonality might explain why diverse vertebrates show remarkable spatial reasoning abilities despite wide variations in sensory modalities and neural architectures. The Wanderers Hypothesis thus bridges gaps between behavioral ecology, cognitive psychology, and linguistics, synthesizing multidisciplinary findings into a unified theoretical framework.

Throughout her article, Dillon carefully situates her perspective within the broader intellectual lineage, acknowledging the insights and limitations of earlier thinkers while pushing the scientific dialogue forward. By focusing on empirical evidence—from animal navigation experiments to infant cognitive assessments and AI benchmarking—she constructs a persuasive case that challenges deeply entrenched assumptions about human exceptionalism in geometric reasoning. Her work underscores the importance of viewing cognition as an adaptive, evolutionary phenomenon that can be profoundly influenced by cross-species shared mechanisms.

In conclusion, Moira Dillon’s research reorients our understanding of the cognitive origins of geometry. Rather than arising from a discrete, innate mental geometry module present only in humans, geometric knowledge appears to be scaffolded upon ancient, evolutionarily conserved navigation processes that permit animals to perceive and mentally simulate spatial relationships. The uniquely human leap emerges from linguistic capabilities, which allow abstract manipulation of these spatial representations independent of immediate environmental interaction. This paradigm not only reshapes theories about geometric cognition but also holds potential to inspire innovations in education, AI development, and understanding of cognitive evolution.

As ongoing research delves deeper into the neural and behavioral bases of spatial reasoning, Dillon’s Wanderers Hypothesis for Geometry stands as a bold and illuminating model. It challenges us to rethink the cognitive architecture from which geometry arises, emphasizing the intertwined roles of biology, environment, and language in shaping one of humanity’s most treasured intellectual achievements. The convergence of psychology, neuroscience, ethology, and AI research in this endeavor exemplifies the dynamic, integrative future of cognitive science.

Subject of Research: Cognitive origins of geometric understanding and its evolutionary and cross-species foundations
Article Title: The cognitive origins of geometry
News Publication Date: 6-Apr-2026
Web References: https://doi.org/10.1016/j.tics.2026.01.005
Image Credits: Jonathan King/NYU
Keywords: Mathematics, Learning processes, Cognitive psychology, Cognitive development

The essence of the study lies in the concept of “label extension,” a critical aspect of early language acquisition in human infants, where a single word becomes generalized to encompass a category of objects sharing functional similarities. For example, young children learn that the term “cup” may refer to a range of drinking vessels, regardless of their varying appearances. Until now, demonstrations of this complex ability in non-human species have been scarce and mostly limited to highly trained captive animals subjected to lengthy conditioning protocols.

What sets this research apart is the observation of Gifted Word Learner dogs naturally expanding their verbal vocabulary within their everyday environments. Unlike traditional captive studies, where supervision and formal training dominate, these dogs acquire toy names through spontaneous interactions with their owners, highlighting the significance of social context and play in cognitive development. This naturalistic methodology allowed the research team to probe whether dogs can categorize objects functionally, beyond superficial appearance.

The research design encompassed a play-based experimental framework involving seven dogs—six Border Collies and one Blue Heeler—each known for their extraordinary aptitude for retrieving and recognizing the names of an extensive array of toys. The experiment was divided into four distinct phases, meticulously crafted to simulate an authentic learning process: first, the dogs were introduced to two novel verbal labels linked not to individual items but to groups of toys unified by their use during play, such as “Pull” and “Fetch.”

Following this initial learning phase, an assessment confirmed that the canines had correctly associated these labels with the appropriate toy groups. This verification was essential to ascertain that the dogs had effectively learned the meaning of the labels before undergoing further testing. The pivotal moment arrived in the generalization phase when the dogs encountered entirely new toys with diverse shapes and appearances, devoid of any verbal labels during play—yet involving the same functional interactions previously taught.

Remarkably, in the final test, each dog was prompted to fetch either a “Pull” or a “Fetch” toy without verbal cues during the gameplay. Despite the absence of direct labeling, the dogs consistently selected the correct unlabelled toys significantly above chance levels, demonstrating their capability to extend learned verbal labels to entirely novel objects based on their function. This outcome challenges the presumption that such sophisticated label generalization is uniquely human or restricted to extensively trained captive animals.

This study does not merely add a new chapter to canine cognition; it opens a vast realm for understanding the cognitive mechanisms underlying language-related behaviors across species. The dogs’ ability to generalize labels based on functional attributes instead of simple visual categorization suggests cognitive processing akin to concepts of meaning and category formation, foundational elements in language comprehension. Such findings propel forward our understanding of non-verbal communication and semantic reasoning in animals.

Moreover, the fact that this ability emerges naturally within the intimate setting of a human family amplifies the potential ecological validity of the findings. The social dynamic between dogs and their owners, characterized by playful interactions and spontaneous learning, mirrors aspects of human infant language acquisition, emphasizing the interplay between environment, social context, and cognitive development. These insights prompt reevaluation of how communicative skills may evolve and function beyond linguistic species.

From a neuroethological perspective, the results provoke intriguing questions about the neural substrates involved in such functional label generalization. Does this reflect a specialized system for semantic processing in dogs? Or is it an extension of more generalized associative learning mechanisms fine-tuned by social bonds and environmental influences? Future research leveraging neuroimaging and comparative cognition techniques could illuminate these pathways.

Additionally, the discovery that different toy categories like “Pull” and “Fetch” can serve as abstract functional classes for dogs challenges earlier assumptions that animal classification behavior is predominantly perceptual. Instead, it underlines an impressive representational capacity—dogs appear to form mental categories defined by action-related properties, a sophisticated trait aligning with foundational elements of symbolic thought.

The implications extend beyond theoretical interest; understanding how dogs comprehend verbal labels functionally has practical applications in enhancing communication between humans and working dogs, potentially improving training methods for service, therapy, and assistance roles. The ability to generalize commands based on function might foster more adaptive and resilient behaviors in dynamic environments.

In sum, this pioneering study reshapes the conversation regarding animal cognition and language evolution. Gifted Word Learner dogs not only memorize toy names but demonstrate a nuanced comprehension of their use, enabling them to generalize these labels to new objects within the functional categories they represent. This phenomenon, discovered through natural social interactions and play, underscores the depth and subtlety of canine intelligence.

As Dr. Claudia Fugazza, lead author of the study, notes, these findings reveal that dogs’ label learning reflects more than rote memorization; it encompasses an understanding of meaning that transcends mere appearance. Dr. Adam Miklosi, coauthor of the study, highlights that such capacities in a non-human, non-linguistic species challenge traditional views and usher in exciting possibilities for studying the evolutionary pathways of language-related skills.

This research encourages a shift in scientific paradigms, inviting us to explore language not solely as a human faculty but as a complex cognitive ability with roots and parallels in other species, particularly those integrated into human social frameworks. Indeed, the remarkable linguistic aptitude of Gifted Word Learner dogs provides a compelling window into the ancestral capabilities that may have set the stage for human language, while simultaneously deepening our appreciation for the sophisticated minds of our closest animal companions.

Subject of Research: Gifted Word Learner dogs’ ability to extend verbal labels based on functional object classification.

Article Title: Dogs extend verbal labels for functional classification of objects

News Publication Date: 18-Sep-2025

Web References: http://dx.doi.org/10.1016/j.cub.2025.08.013

Image Credits: Department of Ethology / Eötvös Loránd University

Keywords: Dogs, Ethology, Animal communication, Behavioral ecology, Life sciences