Contesting the Analog/Digital Distinction

The analog/digital distinction is a recent conceptual development, only first explicitly emerging in the context of electrical engineering in the early 1940s. Although originally introduced to distinguish between two types of calculating machines, it was quickly projected into a variety of new contexts, from artificial intelligence and philosophy of mind to neuroscience, perceptual psychology and beyond. By the early 1960s, however, the extension of the distinction into these new contexts, alongside developments in computer science, led to debates about the very nature of the distinction itself. 

In this workshop, we propose to explore the debates surrounding the analog/digital distinction in a contemporary context. Our starting point will be an exchange between Nelson Goodman and David Lewis in the late 1960s and early 1970s. In 1968, Goodman proposed that the etymological roots of the terms in the distinction do not capture its true essence—that it has nothing to do with analogies and digits—and that it should be understood as a distinction between continuous and discrete systems. In 1971, David Lewis countered by pointing to examples in computer science of analog computers that are both continuous and discrete, and proposed that the distinction is better understood as a distinction between representation by covariation and representation by a series of digits. The debate between Goodman’s and Lewis’s views continued into the 1970s and 1980s, with Goodman’s view eventually becoming the received view endorsed by most computer science textbooks—even if the underlying philosophical debate remained unresolved. 

In the late 2000s and early 2010s, a series of articles by Matthew Katz, Corey Maley, and others revived the Goodman/Lewis debate and applied it to a variety of issues in philosophy of mind and cognitive science. Our workshop is intended to further explore these issues, updated to a contemporary context. The workshop will begin with a paper by Cory Maley on the physicality of analog systems. Next, Alistair Isaac will give a paper on the different sorts of noise that characterize analog and digital systems. Rachel Goodman will then give a paper on whether there are analog synactic types. The last paper will be by Jake Quilty-Dunn, on the relationship between iconic representation and analog systems. The workshop will end with a discussion of all four papers, moderated by Zed Adams.

Schedule:

SYMPOSIA SESSION ONE (11:00-13:00) 

First Talk (11:00-11:30) 

Corey Maley 

Associate Professor of Philosophy 

University of Kansas 

Title: “The Physicality of Analog Representation” 

Abstract: There are a couple of ways of characterizing the analog/digital distinction. On the traditional, received view, continuity is the essential feature of the analog; on the mirroring view, a specific kind of covariation between representation and what is represented is essential. While I think the mirroring view has some advantages over the received view, both ways of characterizing the analog illuminate an under-theorized feature of this kind of representation: the importance of the physicality of analog representation. I will discuss what I mean by this, particularly in contrast to other kinds of representation that abstract away from the physical details of their substrates, including what this means in the context of contemporary computational cognitive science. 

Second Talk (11:30-12:00) 

Alistair Isaac 

Senior Lecturer in Philosophy 

The University of Edinburgh 

Title: “Productive Noise” 

Abstract: I approach the analog/digital distinction through the distinctive form and potential for noise in each system type. Analog noise or distortion is typically thought of as graceful, sometimes even desirable, as it smooths a signal, while digital noise is harsh and brittle, often destroying the signal in manner that prevents it from playing its correct functional role. In the context of neural networks, analog noise may even equate to the kind of spontaneous activity associated with decision making and perceptual activity. Passing from music synthesis technologies, through computers proper, to models of cognition, I explore the implications of taking the type of noise found in a system as indicative of its status as analog or digital. 

Third Talk (12:00-12:30) 

Rachel Goodman 

Assistant Professor of Philosophy 

The University of Illinois at Chicago

Title: Are there analog types? 

Abstract: Traditional ways of making out the analog/digital distinction take the notion of a syntactic type to apply in both digital and analog systems. For example, Goodman (1976) and Haugeland (1981) understand syntactic types within an analog system, unlike those within a digital system, to be smooth and continuous. This paper voices skepticism that the notion of an analog syntactic type makes sense at all and defends this skepticism by examining a central role for the notion of a syntactic type within a representational system: providing a vehicular mechanism by which we can understand the presence of semantic coordination and frege cases within the system. Arguably, there are no phenomena of this kind to explain within analog systems so no role for syntactic types to play. To the extent that such phenomena do show up in analog systems, it’s arguable that they can do so only in virtue of a digital typing of the tokens involved. 

Fourth Talk (12:30-1:00) 

Jake Quilty-Dunn 

Research Fellow 

University of Oxford 

Title: “Icons as Complex Analog Representations” 

Abstract: Philosophers of perception often talk about perceptual representations as being “analog” or “iconic.” There is no clear statement of the relation between these apparently distinct representational formats. Iconic representations obey a correspondence principle: parts of the icon correspond to parts of what's represented. The correspondence principle need not be true of analog mental representations corresponding to numerosity or hue (Clarke forthcoming), which instead vary along vehicle-dimensions in ways that track variation in objective feature-dimensions (Maley 2011). But other perceptual representations seem to be iconic, including the elements of iconic memory (Quilty-Dunn forthcoming). This talk addresses two questions: (1) What is the relation between iconic and analog representations, and (2) How do analog representations compose? The answer to both questions is that icons are complex analog representations. Icons bind features not through composing discrete representational vehicles, but rather through coordinating values along multiple analog dimensions simultaneously. Primitive parts of icons can be modeled as sets of coordinates along analog dimensions. Iconic coordination is a sui generis form of compositionality that differs from predication and feature-placing. This “coordination model” of iconicity explains the holistic character of icons (i.e., why color and location are represented together in icons in a way they are not in linguistic representations). 

SYMPOSIA SESSION TWO (14:00-15:30) 

Extended discussion of the morning papers. Zed will moderate this discussion, asking focused questions of each of the presenters as well as directing audience questions towards the respective speakers.

Further conference information: