Archive for the ‘Explanation’ Category


Can there be spatially distributed parts in a mechanistic explanation?

June 9, 2014

Morgan Thompson

Abstract: Some mechanists maintain that mechanistic explanation is the ideal type of explanation in neuroscience and that mechanistic explanations can contain spatially distributed parts. There is a tension between these two claims. Carl Craver (2007) lists the following criteria for parts that are appropriate for featuring in mechanistic explanations: being detectable by multiple theoretically independent techniques (a la Wimsatt), having a stable cluster of properties (a la Boyd), being manipulatable, and being physiologically plausible. However, spatially distributed parts are not clearly appropriate on those criteria. I appeal to examples from neuroscience where researchers use multivariate pattern analysis to analyze spatially distributed networks. I conclude that mechanists should either adjust their criteria for parts appropriate for mechanisms so that they include spatially distributed parts or they should give up mechanistic explanation as the ideal explanatory scheme in neuroscience.



The Cognitive Neuroscience Revolution

February 24, 2014

Trey Boone

Abstract: Once upon a time, there was cognitive science—the interdisciplinary study of cognition. It included (aspects of) six disciplines: psychology, computer science, linguistics, anthropology, neuroscience, and philosophy. The six disciplines were supposed to work together towards understanding cognition, but there was also a clear division of labor between them. On one side stood psychology, with the help of computer science, linguistics, anthropology, and philosophy; on the other side stood neuroscience. Psychology etc. studied the functional, cognitive, or—in Marr’s terminology—the computational and algorithmic levels; neuroscience investigated the neural, mechanistic, or implementation level. These two approaches were considered to be autonomous from one another. This division of labor leaves no room for cognitive neuroscience. Indeed, from this perspective, the very term “cognitive neuroscience” is almost an oxymoron, because neuroscience is supposed to deal with the mechanisms that implement cognitive processes, not with cognition proper. Yet cognitive neuroscience has emerged as the new mainstream in cognitive science. What gives?

We argue that cognitive science as traditionally conceived is on its way out and is being replaced by cognitive neuroscience, broadly construed. Cognitive neuroscience is still an interdisciplinary investigation of cognition. It still includes (aspects of) the same six disciplines (psychology, computer science, linguistics, anthropology, neuroscience, and philosophy). But the old division of labor is gone.

The old two-level view (functional/cognitive/computational vs. neural/mechanistic/implementation) is being replaced by a view on which there are many levels of mechanistic organization. No one level has a monopoly on cognition proper. Instead, different levels are more or less cognitive depending on their specific properties. Old psychological theories pitched at the “functional level” are simply sketches of mechanistic explanations at one of many levels of mechanistic organization (Piccinini and Craver 2011). The disciplines contributing to cognitive science are not autonomous from one another. Instead, these different disciplines contribute to the common enterprise of constructing multilevel mechanistic explanations of cognitive phenomena.


When is Harmonics a Science?

November 22, 2013

Marina Baldissera Pacchetti

Abstract: Aristotle bases the principle of harmony on the diesis being the indivisible unit that can be represented by number. This, in a passages of De Anima and De Sensu, sets aesthetic standards for ratios of numbers representing melodious concords (`symphonia’), which are demonstrated by arithmetic. This allows him to define harmonics to be one of the `more physical of the mathematical sciences’ because it shares the principle of indivisible unit number with arithmetic, and the demonstrations that provide knowledge about concords are arithmetical.

Aristoxenus says that this is not an appropriate treatment of harmonics and a proper explanation of what justifies aesthetic perception of consonant sounds. The principles of this science cannot rely on mathematics, but on movement with respect to pitch space. Harmonious sound arises from intervallic movement (to be defined), which is recognized only in terms of its phenomenology. He does not discard the use of arithmetic, but he rather sees it as a useful tool to calculate modes – as Barker says (note 50 GMW II p.135), Aristoxenus does not have a problem in including intervals smaller than a quarter tone (the diesis) in his `science of harmonics’. The inclusion of tones smaller than a quarter tone was problematic in explanations of concords for his predecessors (esp. Pythagoreans) and those who will come afterwards (Ptolemy 2C AD, Boethius 6C AD, Gafurio 15C AD, Zarlino 16C AD). Aristoxenus does not deny that the use of mathematics is not useful in harmonics, but as his `first principles’ are not mathematical, he seems to be able to accommodate standard calculations within his framework without incurring in inconsistencies in his philosophical framework.


Day-O-WIPs Beta

June 17, 2013

The second installment of the “Day-O-WIPs” series:

“Toward a Philosophy of Synthetic Science” Julia Bursten

“Can Genes be Darwinian Individuals?” Haixin Dang

“Group Theory or No Group Theory: Understanding Atomic Spectra” Joshua Hunt

“Dynamical Models: A Type of Mathematical Explanation in Neuroscience and Medicine” Lauren Ross

“The Wax & the Mechanical Mind: Reexamining Hobbes’s Objections to Descartes’ Meditations” Marcus Adams