Perspectives on Cognitive Neuropsychology

G. Denes, C. Semenza, P. Bisiacchi

• Publisher: Routledge
• Year: 2018
• ISBN:
  9781317267980
• Edition: 1
• Topic: Psychology
• Subtopic: Cognitive Neuroscience & Neuropsychology
• Index: Psychology

Originally published in 1988, neuropsychology had evolved considerably in the decade prior to publication. Methods of research had changed, new topics of interests had emerged, and there had also been an ongoing debate as to the objectives of the field. This evolution had most radically affected the study of cognitive disorders – an area that had developed particularly rapidly in the years before and that ultimately became a field of inquiry in its own right, that of cognitive neuropsychology.

The contributors to this volume all share a fascination with the new perspectives for understanding how the mind works that have arisen from the study of impaired cognition. Yet, and this was very characteristic of the state of the art in cognitive neuropsychology at the time, they disagreed on many important issues, even those pertaining to the most basic assumptions of their discipline. Therefore, the first part of this book is devoted to an attempt to define and clarify these basic issues and to the confrontation of alternative views. The remaining parts present original studies on several topics of particular interest in cognitive neuropsychology.

PART I: BASIC ISSUES

1

A Function for Cognitive Neuropsychology

C. Semenza and P. Bisiacchi

Università di Padova, Italy

V. Rosenthal

INSERM, Hôpital Salpêtrière, Paris, France

INTRODUCTION

The current neuropsychological literature reveals the considerable influence of concepts and methods imported from cognitive science on the investigation into the effects of the neurological impairment of cognitive skills. Quite significantly, the commitment of an increasing number of neuropsychologists to the information-processing paradigm has taken place, together with a substantial change of attitude on the part of many cognitive psychologists who tend, at present, to view neuropsychological research as a valuable and potentially instructive source of data. This rapprochement manifests itself most obviously in the several attempts to develop working models of particular cognitive (sub) systems on the basis of both neuropsychological and “normal” findings (see, for instance, Cohen, 1983; Coltheart, Patterson, & Marshall, 1980; Ellis, 1982; Garrett, 1982; Marcel, 1983; Patterson, Marshall, & Coltheart, 1985). Given these recent developments, it is tempting to conclude that cognitive neuropsychology is in the process of becoming a respectable partner of the other disciplines that make up cognitive science.

This state of affairs appears to have several far-reaching consequences that we shall try to explore in the present chapter. It seems to us that the commitment, in the late 1970s, of many research workers in psychology, linguistics, computer science, philosophy, and sociology to an integrated cognitive science has been greatly inspired by the belief that none of these traditionally independent sciences can, by themselves, cope successfully with the complexity of cognition (cf. Norman, 1981, for further details). Linguistics, psychology, or artificial intelligence may ultimately pursue different goals; yet, many, or most of the questions that they are currently investigating can usefully be viewed as information-processing problems and, as such, an important preliminary task is that the structure of these problems be first understood. To be sure, it would be somewhat eccentric to compare man and computer in terms of their internal structure and circuitry. However, as information-processing systems engaged in tasks such as pattern recognition or natural language processing, man and computer systems face a number of common problems—the problems that are inherent in the realisation of such tasks (see Andreewsky & Rosenthal, 1985; Marr, 1977).

In this chapter we shall inquire into the rationale for the study of the effects of neuropsychological disturbances of cognitive skills in the research programme of cognitive science.

THE TARGET: INTELLECTUAL SKILLS

Intellectual skills are probably the most typical instances of activities that cognitive scientists are attempting to explain. The notion of skill is all the more important for neuropsychologists because they are engaged in observing individuals who fail to perform skilfully some typically cognitive tasks. The significance of this concept for the understanding of cognition has generally passed unnoticed, though it may be argued that the clarification of the structure of cognitive skills is crucial for any further advance in cognitive science. Perception, language comprehension and production, or everyday thinking are most obvious instances of cognitive skills, or, more appropriately put, of clusters of skills. Most significantly, however, it seems that it is a necessary condition for the effective realisation of such activities that they be carried out skilfully (i.e. smoothly, effortlessly, and outside of awareness). This may not necessarily be the case for all acts of cognition, but skilfulness seems to be a basic property of all complex mental activities that demand the deployment of considerable intellectual resources. It should be noted, on the other hand, that skills are never perfect—their execution always admits of a certain degree of failure—so one may wonder whether imperfection does not constitute another intrinsic characteristic of skilfulness.

In his seminal work, Personal Knowledge (1962), the philosopher Michael Polanyi devoted considerable attention to the analysis of the concept of skill. He noted that “…the aim of skilful performance is achieved by the observance of a set of rules which are not known as such to the person following them” (p. 49). It does not mean that we cannot tentatively define, at least, some rules, but—as Polanyi observed—such definitions are of little, if any, practical help. His well-known example concerned how to ride a bicycle. “A simple analysis shows that for a given angle of unbalance the curvature of each winding is inversely proportional to the square of the speed at which the cyclist is proceeding” (p. 50). But can this help one to ride a bicycle? “You obviously cannot adjust the curvature of your bicycle’s path in proportion to the ratio of your unbalance over the square of your speed; and if you could you would fall off the machine, for there are a number of other factors to be taken into account in practice which are left out in the formulation of this rule” (p. 50). The cyclist, of course, is not permitted to transgress this rule. But the implicit observance of a rule should not be confused with following it explicitly. Such rules are products of analyses carried out by external observers, yet the behaviour of an organism may not necessarily be governed by rules by which it is describable. These rules, in fact, may be only tacit.

The execution of skills is not improved by conscious knowledge of their particular components, but rather, as Polanyi noted, these particulars must remain beyond one’s focal awareness. In the present vocabulary, we would say that skilful performance is necessarily goal-directed and automatic, for if one pays attention to the particular components of a given activity the entire execution is liable to collapse. Thus, for instance, if a pianist shifts attention from the piece she is playing to what she is doing with her fingers, the execution of the musical work is impaired and she may even have to stop. Or, consider Polyani’s (1962, p. 56) description of stage-fright, “which seems to consist in the anxious riveting of one’s attention to the next word—or note or gesture—that one has to find or remember. This destroys one’s sense of the context, which alone can smoothly evoke the proper sequence of words, notes, or gestures.” The particulars of a skill are unspecifiable¹ for two reasons. First, because the performance is generally paralysed if we focus our attention on particulars. Second, and most important, skills are logically unspecifiable for it can be shown that the specification of particulars “composing” a skill would contradict what is implied in the whole performance. In fact, none of such particulars will account for the inherent purpose of the performance.

Skills are acquired through extensive practice. The performance of people who first try to speak a foreign language is clumsy, slow, effortful, and demands a good deal of attention; quite interestingly, these individuals may be aware of some particulars of their performance (e.g. looking for words, consciously combining them according to rules they have learned, etc….). Proficient speakers, on the other hand, are barely aware of what they are doing when they speak (even though they sometimes happen to commit word slips or mispronounce words). Whereas skilful performance is opaque and inaccessible to conscious inspection, unskilled performance appears to be at least partly transparent (though one may not put too much credence in such introspection). In this context, it is interesting to consider that some intellectual activities never attain the level of skill and remain apparently transparent. Formal reasoning, whereby, for instance, one has to find the appropriate conclusion from:

is always poorly performed unless concrete instances are substituted for x, y y and z.

Finally, as already mentioned, complex cognitive activities are best characterised as clusters of skills—such activities are never represented by a single task, but, instead, constitute a cluster of related tasks—and, as Kolers and Roedinger (1984) pointed out, it is probably in the very nature of skills that they are transferable between related tasks depending on the degree of overlap in the structure of such tasks.

The Means

The foregoing description of skills allows one to appreciate the difficulty faced by psychologists and computer scientists in their attempts to understand and model cognitive activities. Traditionally, psychologists proceeded by way of devising experiments so as to put constraints on (or simply manipulate) certain particular factors supposed to play a preeminent role in the execution of a given task. But if one considers seriously the above-mentioned characteristics of skills—especially the observation that the skilful performance of a task is not necessarily governed by the rules by which it may be described, that it is goal-directed, and it is not decomposable to specifiable particulars—it becomes clear that traditional methods of psychological research may not be the most appropriate for the investigation of intellectual skills.

The experience with machine translation in the 1960s will help to illustrate this point. This work was undertaken with a number of a priori linguistic and psychological concepts or, more precisely, with an a priori definition of the particulars of which a translation ought to consist. That this pioneer venture in artificial intelligence did not succeed was not due to the incorrect implementation of procedures nor to the unavailability of current high-level programming languages, nor could it be attributed to the weakness of the computer technology of the day. The failure of machine translation was due to the incorrect definition of the structure of translation tasks. The primary problem in effect does not consist in finding an appropriate theory and technology of implementation of the translation tasks but rather in conceiving an adequate theory of these tasks.

The emergence of cognitive science can be viewed as a consequence of the misadventures of the first generation of artificial intelligence models as well as of the growing interest on the part of psychologists and linguists in the digital computer as a tool for modelling intelligent behaviour. Obviously, however, for a domain of intellectual interest to become an autonomous scientific discipline it is necessary to define a conceptual framework within which to operate. It is generally assumed that cognitive science is defined by the information-processing paradigm—a paradigm that supposedly can be applied to the description of all cognitive activity. Yet, in actuality this may not constitute an adequate framework within which to define or conduct scientific activity. Indeed, the adequacy of this application has prompted a host of controversies regarding, for instance, the question of just how analogous to the human brain a computer may be considered to be (see Dreyfus, 1979). The problem is that information processing is merely a conceptual tool or method for describing behaviour. Although in physics and psychology, comparable statistical methods are often used, no one would claim that these sciences share a common conceptual framework. Neither information-processing analysis nor statistical methods constitute a proper level of scientific theory. And accordingly, information processing per se does not allow one to assume that, although studying language comprehension and natural language processing, psycholinguists and computer scientists are basically raising similar questions.

In order to address a proper level of cognitive theory it is thus necessary to define a framework that logically permits one to consider that whatever information-processing device is to be used it will not modify the structure of a given cognitive activity. Specifically, we submit that before considering how a given cognitive task is or ought to be implemented by a particular information-processing device, it is first necessary to develop a theory that specifies the problems that are inherent in the execution of that task. These inherent problems bear on the nature and the order of operations that a task necessarily implies, but not on how they may be carried out. In other words, a task theory defines the succession of operative sub-goals that have to be necessarily attained in the execution of a given cognitive activity. This theory does not, however, furnish a definition of processes that permit a system to achieve such goals nor does it describe the device by which these processes can be implemented. A task theory of a cognitive skill is not a psychological, artificial intelligence, linguistic, or biological theory proper—it is a cognitive theory that specifies of what a task consists and why. It is only at another level, at the level of implementation theory, that the specification of processes is necessarily device-oriented, and thus situated at a psychological or artificial intelligence level proper. It follows that one may conceive several alternative implementations for a task on the basis of the same theory of that task.

Implementation theories also depend on the characteristics of the device that must execute the task, because the processes that are conceived have to be “executable by the machine.” That is, an implementation theory should necessarily be compatible with a theory of the machine that physically “executes the program”—which is a theoretical level proper to biological or electrical engineering explanation (see Rosenthal, this volume, for more detailed discussion of this epistemic hierarchy).

These considerations have relatively clear implications with respect to the study of cognitive skills. Most obviously, if the currently frequent confusion between the above-mentioned three epistemic levels is to persist, it is unlikely that a clear picture of any cognitive skill whatsoever will emerge in the foreseeable future. Recently, this confusion has led to long debates between the proponents of the biological and the mentalistic bases of cognition or to speculations about the supposedly modular or interactive architecture of cognition. Most important, however, it let cognitive psychologists define prematurely their business as the study of cognitive processes. Although cognitive psychology may indeed propose to clarify the processes underlying cognitive tasks, such a goal is unlikely to be attained without a clarification of what has to be processed—that is, prior to the development of an adequate theory of the tasks being studied. In a way, one has to understand what actually has to be done in order to determine how it is or has to be carried out. This argument is so obvious that, to the best of our knowledge, no one has attempted to contest its logical validity in the field of cognitive psychology. However, the problem itself has often been eschewed by promptly adopting an intuitive pseudotheory of the activity under investigation. For instance, the standard pseudo-theory of the tasks of language comprehension postulates that, after the initial perceptual processing of a language stimulus (e.g. a sentence), it is necessary to access and activate in the mental lexicon the meanings of individual words of which the stimulus is composed, and then to combine these lexical meanings according to the structural arrangement of the corresponding words in the input-sentence. Accessing the meaning of a word represented in the mental lexicon amounts to comprehending this word, etc…. Of course, the present definition is rather schematic and does not take into account the alternative formulations of this theory, which are often enriched by intervening factors (such as context effects, etc….). The point is, however, that this standard pseudo-theory, which is based on common sense intuitions about language that can be traced back at least to Plato, and derives, in part, from Treisman’s (1960) conjecture about “dictionary store” or “mental lexicon,” has not been submitted to critical examination (either conceptual or empirical). O...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Original Title Page
• Original Copyright Page
• Contents
• List of Contributors
• Preface
• Part I: Basic Issues
• Part II: Language Comprehension
• Part III: Reading and Writing
• Part IV: Face Recognition
• Part V: Memory
• Indices