Adele Abrahamsen

Why Do Biologists Use So Many Diagrams

Diagrams have distinctive characteristics that make them an effective medium for communicating research findings, but they are even more impressive as tools for scientific reasoning. To explore this role, we examine diagrammatic formats that have been devised by biologists to (a) identify and illuminate phenomena involving circadian rhythms and (b) develop and modify mechanistic explanations of these phenomena.

Beyond the exclusively propositional era

Contemporary epistemology has assumed that knowledge is represented in sentences or propositions. However, a variety of extensions and alternatives to this view have been proposed in other areas of investigation. We review some of these proposals, focusing on (1) Ryles notion of knowing how and Hansons and Kuhns accounts of theory-laden perception in science; (2) extensions of simple propositional representations in cognitive models and artificial intelligence; (3) the debate concerning imagistic versus propositional representations in cognitive psychology; (4) recent treatments of concepts and categorization which reject the notion of necessary and sufficient conditions; and (5) parallel distributed processing (connectionist) models of cognition. This last development is especially promising in providing a flexible, powerful means of representing information nonpropositionally, and carrying out at least simple forms of inference without rules. Central to several of the proposals is the notion that much of human cognition might consist in pattern recognition rather than manipulation of rules and propositions.

From Reactive to Endogenously Active Dynamical Conceptions of the Brain

We contrast reactive and endogenously active perspectives on brain activity. Both have been pursued continuously in neurophysiology laboratories since the early 20th-century, but the endogenous perspective has received relatively little attention until recently. One of the many successes of the reactive perspective was the identification, in the second half of the 20th century, of the distinctive contributions of different brain regions involved in visual processing. The recent prominence of the endogenous perspective is due to new findings of ongoing oscillatory activity in the brain at a wide range of time scales, exploiting such techniques as single-cell recording, EEG, and fMRI. We recount some of the evidence pointing to ways in which this endogenous activity is relevant to cognition and behavior. Our major objective is to consider certain implications of the contrast between the reactive and endogenous perspectives. In particular, we relate these perspectives to two different characterizations of explanation in the new mechanistic philosophy of science. In a basic mechanistic explanation, the operations of a mechanism are characterized qualitatively and as functioning sequentially until a terminating condition is realized. In contrast, a dynamic mechanistic explanation allows for non-sequential organization and emphasizes quantitative modeling of the mechanisms’s behavior. For example, with appropriate parameter values a set of differential equations can be used to demonstrate ongoing oscillations in a system organized with feedback loops. We conclude that the basic conception of mechanistic explanation is adequate for reactive accounts of brain activity, but dynamical accounts are required to explain sustained, endogenous activity.

Cognitive Science: History

This article describes the development of cognitive science as a collaborative endeavor of psychology, computer science, neuroscience, linguistics, and related fields, emphasizing how the collaboration began in the 1950s and the subsequent interactions between the collaborators. Key early developments, which were featured at a conference in 1956, were Newell and Simons program for proving logic theorems, Chomskys generative grammar, and Millers behavior research on cognitive interference and limits of short-term memory. Major developments in subsequent decades include the development of list-processing program languages and production systems as well as larger-scale knowledge structures such as frames and scripts in AI, attempts to employ linguistic grammars in psycholinguistic research, and development of information processing models of many cognitive activities. In the mid-1970s the Sloan Foundation and the System Development Foundation both began supporting cognitive science research. In 1977 the journal Cognitive Science was established, and in 1979 the Cognitive Science Society began, providing institutional identities for cognitive science. The identity of cognitive science, though, has been changing in the 1980s and 1990s, especially with the development of cognitive neuroscience and with the emergence of a situated, embodied focus in some cognitive science research.

Bridging Interdisciplinary Boundaries: The Case of Kin Terms

A great debate over kin term analysis was waged for two decades beginning in the mid-1950s. By the end of this period, the principal dispute was between two different approaches to fractionating the meanings of the terms. In componential analysis, each term took a unique set of values on a set of abstract dimensions such as sex, generation and lineality; the results were displayed in a tabular format (componential paradigm). In relational analysis, each term was analyzed as a composition of one or more relational components; for example, a grandparent is the parent of a parent. This dispute petered out by the mid-1970s despite the fact that it remained unresolved. In this paper I take up where the debate left off, for two reasons: (1) an integrated account can now be offered, within which the two approaches that appeared to be at odds fit together as pieces of a solution; and (2) the history of the debate and its resolution in an integrated account provide a particularly illuminating case study of how disciplines such as linguistics and psychology relate to one another. In what follows, first, the debate over kin terms is placed in the context of an account of disciplinary structure and interrelationships which I recently proposed (Abrahamsen, 1987, from which some of the following exposition is drawn); second, the high points of the debate are introduced, and the integrated account is discussed in some detail.

Interfield Connections and Psychology

Woodward and Devonis propose a program for analyzing the history of psychology based on the notion of interfield theory. Both of us have long been interested in relations between fields or disciplines of science, and one of us (WB) has analyzed a number of examples of interfield theories. The notion of an interfield theory was developed for quite specific purposes, however, and it is not clear that it can bear the load Woodward and Devonis seek to place upon it. We will first indicate some of the challenges that will confront any attempt to apply the notion of interfield theory to analyzing the development of psychology as a discipline and then offer an alternative perspective on psychology as a discipline, from which we might then apply the notion of interfield theory fruitfully in a more restricted way. In addition to advancing a framework for analyzing the development of psychology, Woodward and Devonis also advance interpretations of specific aspects of psychology’s history. Some of these interpretations are open to challenge or alternative formulations; however, we have chosen to restrict our focus to questions raised by their application of interfield theory.

Using Diagrams to Reason About Biological Mechanisms

In developing mechanistic explanations for biological phenomena, researchers have their choice of several different types of diagrams. First, a mechanism diagram spatially represents a proposed mechanism, typically using simple shapes for its parts and arrows for their operations. Beyond this representational role, such diagrams can provide a platform for further reasoning. Published diagrams in circadian biology show how question marks support reasoning about the proposed molecular mechanisms by flagging where there are knowledge gaps or uncertainties. Second, an annotated mechanism diagram can support computational modeling of the dynamics of a proposed mechanism. Each variable and parameter needed for the model is added to the diagram adjacent to the appropriate part or operation. Anchoring the model in this way helps with its construction, revision, and interpretation. Third, a network diagram fosters a different approach to mechanistic reasoning. Layout algorithms are applied to data generated by high-throughput experiments to reveal modules that correspond to mechanisms. We present examples in which network diagrams enable viewers to advance hypotheses about previously unknown mechanisms or unknown parts and operations of known mechanisms as well as to develop new understanding about how a given mechanism is situated in a larger environment.