Marr's Attacks: On Reductionism and Vagueness

It has been suggested that Marr took the three levels he famously identifies to be independent. In this paper, we argue that Marr's view is more nuanced. Specifically, we show that the view explicitly articulated in his work attempts to integrate the levels, and in doing so results in Marr attacking both reductionism and vagueness. The result is a perspective in which both high-level information-processing constraints and low-level implementational constraints play mutually reinforcing and constraining roles. We discuss our recent work on Spaun—currently the world's largest functional brain model—that demonstrates the positive impact of this kind of unifying integration of Marr's levels. We argue that this kind of integration avoids his concerns with both reductionism and vagueness. In short, we suggest that the methods behind Spaun can be used to satisfy Marr's explicit interest in combining high-level functional and detailed mechanistic explanations.     

Grounding Cognitive-Level Processes in Behavior: The View From Dynamic Systems Theory

Marr's seminal work laid out a program of research by specifying key questions for cognitive science at different levels of analysis. Because dynamic systems theory (DST) focuses on time and interdependence of components, DST research programs come to very different conclusions regarding the nature of cognitive change. We review a specific DST approach to cognitive-level processes: dynamic field theory (DFT). We review research applying DFT to several cognitive-level processes: object permanence, naming hierarchical categories, and inferring intent, that demonstrate the difference in understanding of behavior and cognition that results from a DST perspective. These point to a central challenge for cognitive science research as defined by Marr—emergence. We argue that appreciating emergence raises questions about the utility of computational-level analyses and opens the door to insights concerning the origin of novel forms of behavior and thought (e.g., a new chess strategy). We contend this is one of the most fundamental questions about cognition and behavior.     

Thirty Years After Marr's Vision: Levels of Analysis in Cognitive Science

Thirty years after the publication of Marr's seminal book Vision (Marr, 1982) the papers in this topic consider the contemporary status of his influential conception of three distinct levels of analysis for information-processing systems, and in particular the role of the algorithmic and representational level with its cognitive-level concepts. This level has (either implicitly or explicitly) been downplayed or eliminated both by reductionist neuroscience approaches from below that seek to account for behavior from the implementation level and by Bayesian approaches from above that seek to account for behavior in purely computational-level terms.     

Can Computational Goals Inform Theories of Vision?

One of the most lasting contributions of Marr's posthumous book is his articulation of the different “levels of analysis” that are needed to understand vision. Although a variety of work has examined how these different levels are related, there is comparatively little examination of the assumptions on which his proposed levels rest, or the plausibility of the approach Marr articulated given those assumptions. Marr placed particular significance on computational level theory, which specifies the “goal” of a computation, its appropriateness for solving a particular problem, and the logic by which it can be carried out. The structure of computational level theory is inherently teleological: What the brain does is described in terms of its purpose. I argue that computational level theory, and the reverse-engineering approach it inspires, requires understanding the historical trajectory that gave rise to functional capacities that can be meaningfully attributed with some sense of purpose or goal, that is, a reconstruction of the fitness function on which natural selection acted in shaping our visual abilities. I argue that this reconstruction is required to distinguish abilities shaped by natural selection—“natural tasks” —from evolutionary “by-products” (spandrels, co-optations, and exaptations), rather than merely demonstrating that computational goals can be embedded in a Bayesian model that renders a particular behavior or process rational.     

Logic as Marr's Computational Level: Four Case Studies

We sketch four applications of Marr's levels-of-analysis methodology to the relations between logic and experimental data in the cognitive neuroscience of language and reasoning. The first part of the paper illustrates the explanatory power of computational level theories based on logic. We show that a Bayesian treatment of the suppression task in reasoning with conditionals is ruled out by EEG data, supporting instead an analysis based on defeasible logic. Further, we describe how results from an EEG study on temporal prepositions can be reanalyzed using formal semantics, addressing a potential confound. The second part of the article demonstrates the predictive power of logical theories drawing on EEG data on processing progressive constructions and on behavioral data on conditional reasoning in people with autism. Logical theories can constrain processing hypotheses all the way down to neurophysiology, and conversely neuroscience data can guide the selection of alternative computational level models of cognition.     

Rational Use of Cognitive Resources: Levels of Analysis Between the Computational and the Algorithmic

Marr's levels of analysis—computational, algorithmic, and implementation—have served cognitive science well over the last 30 years. But the recent increase in the popularity of the computational level raises a new challenge: How do we begin to relate models at different levels of analysis? We propose that it is possible to define levels of analysis that lie between the computational and the algorithmic, providing a way to build a bridge between computational- and algorithmic-level models. The key idea is to push the notion of rationality, often used in defining computational-level models, deeper toward the algorithmic level. We offer a simple recipe for reverse-engineering the mind's cognitive strategies by deriving optimal algorithms for a series of increasingly more realistic abstract computational architectures, which we call “resource-rational analysis.”     

Grounded Cognition: Past,Present, and Future

Thirty years ago, grounded cognition had roots in philosophy, perception, cognitive linguistics, psycholinguistics, cognitive psychology, and cognitive neuropsychology. During the next 20 years, grounded cognition continued developing in these areas, and it also took new forms in robotics, cognitive ecology, cognitive neuroscience, and developmental psychology. In the past 10 years, research on grounded cognition has grown rapidly, especially in cognitive neuroscience, social neuroscience, cognitive psychology, social psychology, and developmental psychology. Currently, grounded cognition appears to be achieving increased acceptance throughout cognitive science, shifting from relatively minor status to increasing importance. Nevertheless, researchers wonder whether grounded mechanisms lie at the heart of the cognitive system or are peripheral to classic symbolic mechanisms. Although grounded cognition is currently dominated by demonstration experiments in the absence of well‐developed theories, the area is likely to become increasingly theory driven over the next 30 years. Another likely development is the increased incorporation of grounding mechanisms into cognitive architectures and into accounts of classic cognitive phenomena. As this incorporation occurs, much functionality of these architectures and phenomena is likely to remain, along with many original mechanisms. Future theories of grounded cognition are likely to be heavily influenced by both cognitive neuroscience and social neuroscience, and also by developmental science and robotics. Aspects from the three major perspectives in cognitive science—classic symbolic architectures, statistical/dynamical systems, and grounded cognition—will probably be integrated increasingly in future theories, each capturing indispensable aspects of intelligence.     

Beyond Single-Level Accounts: The Role of Cognitive Architectures in Cognitive Scientific Explanation

We consider approaches to explanation within the cognitive sciences that begin with Marr's computational level (e.g., purely Bayesian accounts of cognitive phenomena) or Marr's implementational level (e.g., reductionist accounts of cognitive phenomena based only on neural-level evidence) and argue that each is subject to fundamental limitations which impair their ability to provide adequate explanations of cognitive phenomena. For this reason, it is argued, explanation cannot proceed at either level without tight coupling to the algorithmic and representation level. Even at this level, however, we argue that additional constraints relating to the decomposition of the cognitive system into a set of interacting subfunctions (i.e., a cognitive architecture) are required. Integrated cognitive architectures that permit abstract specification of the functions of components and that make contact with the neural level provide a powerful bridge for linking the algorithmic and representational level to both the computational level and the implementational level.     

Cognitive Neuroscience: The Troubled Marriage of Cognitive Science and Neuroscience

We discuss the development of cognitive neuroscience in terms of the tension between the greater sophistication in cognitive concepts and methods of the cognitive sciences and the increasing power of more standard biological approaches to understanding brain structure and function. There have been major technological developments in brain imaging and advances in simulation, but there have also been shifts in emphasis, with topics such as thinking, consciousness, and social cognition becoming fashionable within the brain sciences. The discipline has great promise in terms of applications to mental health and education, provided it does not abandon the cognitive perspective and succumb to reductionism.     

Wayfinding in familiar and unfamiliar environments in a case of progressive topographical agnosia

A 71-year-old right-handed man (F.G.) presents with prosopagnosia and with an inability to recognize famous and familiar buildings. Despite his deficit, F.G. obtained normal scores on neuropsychological tests of executive functions, language, praxis and primary visuoperceptual skills. Brain MRI showed atrophy predominantly in the right temporal lobe, particularly in the fusiform gyrus and the parahippocampal cortex. The present study investigated F.G.’s ability to orient himself in familiar and new environments. His wayfinding abilities in a familiar environment (i.e., his hometown) were preserved despite an inability to recognize familiar and famous buildings, monuments and landmarks in this environment. Wayfinding was achieved through a heavy reliance on written indications (e.g., names of restaurants and streets), preservation of a pre-existing cognitive map of this familiar environment, and normal executive functions necessary to plan the execution of a given trajectory. In an unfamiliar environment, F.G.’s topographical agnosia resulted in severe wayfinding difficulties and in the inability to build an adequate cognitive spatial representation. F.G.’s topographical agnosia results from a high-level visuoperceptual deficit, characterized by an inability to access a global configuration of complex visual stimuli such as familiar and famous monuments, and an over-reliance on the processing of local features.     

What multiple sclerosis could bring to cognitive neuroscience?

The relevance of multiple sclerosis for cognitive neuroscience has evolved significantly during the last decades. After a relative and enduring disinterest, the 1980's has been marked by a first wave of studies aiming at characterizing the cognitive dysfunctions associated with this disease. Once identified, and grouped under the relatively vague and nonspecific concept of “subcorticofrontal syndrome”, these cognitive symptoms had to wait until the end of the 1990's to give rise to a new and vigorous resurgence of attention. Interestingly, this genuine contemporary revival of interest originates in the promotion of the very same arguments that served until there to explain the weak investment of multiple sclerosis by neuropsychology and cognitive neuroscience. The early disseminated nature of brain lesions, their dynamic and unstable nature, the prevalence of white-matter lesions, and the alteration of non-modular aspects of cognition: all these arguments have discouraged neuropsychologists for a long time. Today, these very same specific properties of multiple sclerosis offer an extremely relevant model to explore cognitive dimensions of brain plasticity, to revivify the concept of disconnection in neuropsychology, and to evaluate some neuroscientific models of consciousness.     

The quandary of reductionism: relevance to Alzheimer disease research

Modern science has embraced reductionism, seeking ever-smaller parts to explain the whole. Although reductionistic approaches are successful in very simple biological modelling, they are not necessarily appropriate for systems of increasing complexity. Drawing on famous historical examples of how non-reductionist thinking has benefited mankind, and of how reductionism has sometimes led to erroneous conclusions, we call attention to the need to move away from purely linear reasoning in order to succeed in addressing many of the problems we face with the predicted demographic increase in seniors, and the increase in numbers of those afflicted with Alzheimer disease. The time has come to reconsider and seriously question our most basic assumptions and beliefs surrounding what we believe Alzheimer disease to be, without which we run the risk of missed opportunities and failure.     

Phrenology and methodology,or “playing tennis with the net down”

Background: In 1835, the British Association for the Advancement of Science exhumed the skull of Jonathan Swift, author of Gulliver's Travels, to submit it to phrenological scrutiny and ascertain the cause of his final illness. The behaviour Swift exhibited during the final 3 years of his life—including memory impairment, personality alterations, language disorder, and facial paralysis—was the cause of much speculation among his contemporaries. Aims: This paper will review the debate between Phrenologists and Alienists, which was focused on the significance of the physical evidence presented by Swift's skull, and its implications for explaining behavioural patterns during his lifetime. His skull was the subject of research and rebuttal over a 12-year period, played out in the major medical publications of the day. Main Contribution: The focus of the arguments hinged on two issues that resonate today in research on cortical localisation of function: the correlation between anatomy and physiology, and the implications of pathology for both. Conclusion: Examination of the 19th-century debate over the evidence represented by Jonathan Swift's skull for brain/behaviour correlations illuminates methodological and theoretical assumptions.     

On the Relation Between Marr's Levels: A Response to Blokpoel (2017)

Blokpoel reminds us of the importance of consistency of function across Marr's levels, but we argue that the approach to ensuring consistency that he advocates—a strict relation through exact implementation of the higher level function at the lower level—is unnecessarily restrictive. We show that it forces overcomplication of the computational level (by requiring it to incorporate concerns from lower levels) and results in the sacrifice of the distinct responsibilities associated with each level. We propose an alternative, no less rigorous, potential characterization of the relation between levels.     

Wayfinding in familiar and unfamiliar environments in a case of progressive topographical agnosia

A 71-year-old right-handed man (F.G.) presents with prosopagnosia and with an inability to recognize famous and familiar buildings. Despite his deficit, F.G. obtained normal scores on neuropsychological tests of executive functions, language, praxis and primary visuoperceptual skills. Brain MRI showed atrophy predominantly in the right temporal lobe, particularly in the fusiform gyrus and the parahippocampal cortex. The present study investigated F.G.'s ability to orient himself in familiar and new environments. His wayfinding abilities in a familiar environment (i.e., his hometown) were preserved despite an inability to recognize familiar and famous buildings, monuments and landmarks in this environment. Wayfinding was achieved through a heavy reliance on written indications (e.g., names of restaurants and streets), preservation of a pre-existing cognitive map of this familiar environment, and normal executive functions necessary to plan the execution of a given trajectory. In an unfamiliar environment, F.G.'s topographical agnosia resulted in severe wayfinding difficulties and in the inability to build an adequate cognitive spatial representation. F.G.'s topographical agnosia results from a high-level visuoperceptual deficit, characterized by an inability to access a global configuration of complex visual stimuli such as familiar and famous monuments, and an over-reliance on the processing of local features.     

Personality and complex brain networks: The role of openness to experience in default network efficiency

The brain's default network (DN) has been a topic of considerable empirical interest. In fMRI research, DN activity is associated with spontaneous and self‐generated cognition, such as mind‐wandering, episodic memory retrieval, future thinking, mental simulation, theory of mind reasoning, and creative cognition. Despite large literatures on developmental and disease‐related influences on the DN, surprisingly little is known about the factors that impact normal variation in DN functioning. Using structural equation modeling and graph theoretical analysis of resting‐state fMRI data, we provide evidence that Openness to Experience—a normally distributed personality trait reflecting a tendency to engage in imaginative, creative, and abstract cognitive processes—underlies efficiency of information processing within the DN. Across two studies, Openness predicted the global efficiency of a functional network comprised of DN nodes and corresponding edges. In Study 2, Openness remained a robust predictor—even after controlling for intelligence, age, gender, and other personality variables—explaining 18% of the variance in DN functioning. These findings point to a biological basis of Openness to Experience, and suggest that normally distributed personality traits affect the intrinsic architecture of large‐scale brain systems. Hum Brain Mapp 37:773–779, 2016. © 2015 Wiley Periodicals, Inc .