Danielle J. Williams

In The Brain Abstracted, Chirimuuta argues against the “literal interpretation” of computational models. The literal interpretation understands computational models as providing literally true descriptions of neural processes. According to Chirimuuta, the literal interpretation involves endorsing the computationalism thesis and adopting a theory of computational implementation. The connection between these views, Chirimuuta argues, is what forces the computationalist to reckon with two significant metaphysical problems, what I call the “problem of implementation” and the “problem of multiple realizability.” Here, I examine the connections between the various views that Chirimuuta considers, and I argue that one who endorses computationalism can avoid both metaphysical problems. I do this by distinguishing computationalism from the literal interpretation and the implementation of computational models in neuroscience. Although this paper engages with Chirimuuta’s project directly, it is helpful for those who wish to make sense of the intricacies of the computationalism thesis and computational functionalism.

In The Brain Abstracted, Chirimuuta argues against the “literal interpretation” of computational models. The literal interpretation understands computational models as providing true descriptions of neural processes. According to Chirimuuta, the literal interpretation involves endorsing the computationalism thesis and adopting a theory of computational implementation. The connection between these views, Chirimuuta argues, is what forces the computationalist to reckon with two significant metaphysical problems, what I call the "problem of implementation" and the "problem of multiple realizability." Here, I examine the connections between the various views that Chirimuuta considers, and I argue that one who endorses computationalism can avoid both metaphysical problems. I do this by distinguishing computationalism from the literal interpretation and the implementation of computational models in neuroscience. Although this engages with Chirimuuta's project directly, it is helpful for those who wish to make sense of the intricacies of the computationalism thesis and computational functionalism.

Recently, a version of realism has been offered to address the simplification strategies used in computational neuroscience. According to this view, computational models provide us with knowledge about the brain, but they should not be taken literally in _any_ sense, even rejecting the idea that the brain performs computations (computationalism). I acknowledge the need for considerations regarding simplification strategies in neuroscience and how they contribute to our interpretations of computational models; however, I argue that whether we should accept or reject computationalism about the brain is a separate issue that can be addressed independently by a philosophical theory of physical computation. This takes seriously the idea that the brain performs computations while also taking an analogical stance toward computational models in neuroscience. I call this version of realism “Analogical Computational Realism.” Analogical Computational Realism is a realist view in virtue of being committed to computationalism while taking certain computational models to pick out real patterns that provide a how-possibly explanation without also thinking that the model is literally implemented in the brain.

A roomsize computer equipped with a new type of circuitry, the Perceptron, was introduced to the world in 1958 in a brief news story buried deep in The New York Times. The story cited the U.S. Navy as saying that the Perceptron would lead to machines that “will be able to walk, talk, see, write, reproduce itself and be conscious of its existence.” More than six decades later, similar claims are being made about current artificial intelligence. So, what’s changed in the intervening years? In some ways, not much.

The term “medium independence” has different meanings. One sense maps onto “abstract-as-abstracta” descriptions while the other maps onto “abstract-as-omission” descriptions. Both senses have been deployed when it comes to understanding the nature of physical computation. However, because medium independence is a polysemic term, the sense being used should be clearly stated. If the sense is not clearly stated, then those who wish to engage in debates regarding medium independence and physical computation run the risk of conflating different but related issues which can have consequences when it comes to thinking about physical computation across contexts.

The Physical Signature of Computation is the most “robust” mapping view that’s ever hit the market. It is impressive in its detail and the careful attention paid to its characterization of both the physical system and the formal computational description—a true service to the philosophical literature. The book promises a “unified account of artifact and biological computation,” but here’s where things take a turn: after handling artifacts, the Robust Mapping Account fades from view and the Mechanistic Account of Physical Computation (in its neurocognitive form) emerges as the way to capture biological computation. To have a unified account, then, requires two distinct approaches to understanding the nature of physical computation. This raises questions about whether the Robust Mapping Account can be extended to neural computation, but most importantly, questions about the compatibly of these two views is questionable given the history of the Mechanistic Account and its underpinning framework.

There are two distinct approaches to Bayesian modelling in cognitive science. Black-box approaches use Bayesian theory to model the relationship between the inputs and outputs of a cognitive system without reference to the mediating causal processes; while mechanistic approaches make claims about the neural mechanisms which generate the outputs from the inputs. This paper concerns the relationship between these two approaches. We argue that the dominant trend in the philosophical literature, which characterizes the relationship between black-box and mechanistic approaches to Bayesian cognitive science in terms of the dichotomy between instrumentalism and realism, is misguided. We propose that the two distinctions are orthogonal: black-box and mechanistic approaches to Bayesian modelling can each be given either an instrumentalist or a realist interpretation. We argue that the current tendency to conflate black-box approaches with instrumentalism and mechanistic approaches with realism stems from unwarranted assumptions about the nature of scientific explanation, the ontological commitments of scientific theories, and the role of abstraction and idealization in scientific models. We challenge each of these assumptions to reframe the debates over Bayesian modelling in cognitive science.