Goal neglect and inhibitory limitations: dissociable causes of interference effects in conflict situations

https://doi.org/10.1016/S0001-6918(99)00012-8Get rights and content

Abstract

Interference effects on task performance in conflict situations might reflect real limitations in inhibitory capabilities or failures to fully or consistently utilize such capabilities in executive control of task performance. We propose that useful clues regarding the actual cause of interference effects may be obtained from examination of their robustness within and between experimental conditions. We illustrate this approach for two major types of interference effects that have commonly been attributed to fundamental inhibitory limitations: Stroop-type interference and residual switch costs. We present results that indicate that these effects may not be unavoidable consequences of fundamental inhibitory limitations but may stem from goal neglect, i.e., failures to fully or effectively deploy inhibitory capabilities. These results indicate that, in addition to mean performance levels, variability of task performance may provide a valuable source of evidence regarding the actual cause of performance limitations or deficits in conflict situations.

Introduction

The human brain is capable of an enormous range of tasks. Yet, people are limited in the number of tasks they can perform concurrently, and usually they seem to be devoted to just one task at any moment. Two fundamental questions to be addressed by any theory of executive control of thought and action are why this serial constriction of external activities exists and how the processing system becomes temporarily committed to one activity rather than another. In this paper, we will be primarily concerned with the second of these questions, but the first one requires some consideration. An important insight, eloquently expressed by Simon (1994), is that a serial organisation of external activities should perhaps not be viewed as the result of resource scarcity prohibiting a presumably more efficient parallel organisation, but as an efficient solution to the problem of how to get a powerful parallel processing device, such as the human brain, to support coherent behavior in complex environments that offer multiple affordances for action. As stated by Simon:

Because of the dispersion of need-satisfying situations in the environment, most needs can be satisfied only after extensive activity involving collaboration of sense and motor organs in pursuit of a specific goal... An effective division of labor is not achieved by segmenting the organism into components that each work toward satisfying one of these goals. It is much more efficient to divide labor by time segments-the resources of the entire organism being devoted, in turn, to satisfying successive goals, the priorities being established by the signaling and attention control mechanisms (1994, p. 7–8).


Similar views have been expressed by other theorists Allport, 1987, Neumann, 1987. While acknowledging the high degree of parallel functioning observable in the sensory organs and in the control of internal processes, this view thus holds that the serial constriction of external activities is “… a response to the structure of the environment and of the organs that sense and act on it” (Simon, 1994, p. 8).

These considerations suggest a preliminary answer to the question of how a specific activity or task is selected: voluntary behavior is directed by goals and setting a goal temporarily commits the processing system to the task of attaining it. In this paper, we will focus on the case of well-trained speeded response tasks that can be performed in a second or so. Performance of such tasks may be viewed as being governed by a mental set or task set that, following Woodworth (1918), can be defined as an assembly of elementary processes, or processing modules, configured to deal with a specific task. Implementation of a task set involves the configuration of relevant processing modules for computing required input-output transformations and the selective enabling or disabling of inter-module connections in order to ensure a proper flow of information Allport, 1989, Monsell, 1996. Setting the task goal triggers retrieval of an abstract representation of the associated task set from procedural memory and the subsequent implementation of that set. When these control processes have finished, the processing system is committed to or prepared for that task, and subsequent task performance can be aptly described as a `prepared reflex' (Woodworth, 1918).

The schema control theory (Norman & Shallice, 1986) provides a convenient framework to expand upon these theoretical notions. Schemas are routine programs, one for each basic type of action or thought operation. For present purposes, a schema may be regarded as an abstract representation of a task set, and selection of a schema as corresponding to the retrieval and implementation of its associated task set. Schemas can be activated by sensory inputs, by other schemas, or by a general executive system, labeled the Supervisory Attentional System (SAS). Schemas compete for the control of thought and action by means of a contention-scheduling process that is probably mediated by lateral inhibition between activated schemas. Ideally, the most strongly activated schema would win this competition and suppress all other schemas. For routine actions in familiar contexts, contention scheduling will generally be sufficient to guarantee selection of the appropriate schema. In other circumstances, as discussed below, the SAS will be needed to bias or modulate the outcome of the contention scheduling process, by providing top-down activation or inhibition of schemas (Norman & Shallice, 1986).

These concepts can be made more concrete by considering the situation where subjects are instructed to prepare for a familiar reaction time (RT) task with an arbitrary stimulus-response mapping. Because the task is familiar, a well-integrated schema for the task has presumably already been acquired. The problem then is to select this schema and implement the associated task set. In order to do so, the appropriate task goal must be set. This goal then provides the drive for the SAS to activate the goal-relevant schema. In many cases, this top-down activation, followed by contention scheduling, will suffice to ensure selection of the relevant schema and adequate suppression of all other potentially applicable schemas. In other cases, further intervention of the SAS will be needed. Two such cases are of particular interest for present purposes. One is where the stimuli of the task are more strongly or naturally associated with other tasks; a familiar example is the Stroop task where naming the word is a more strongly automatic response than naming the color of the ink in which the word is printed. The other is where the stimuli of the task were recently encountered in a different task and there is residual activation of the schema for that task (Allport, Styles & Hsieh, 1994). Even if the competing schemas could be at least partially suppressed through contention scheduling, as only the relevant schema receives top-down activation from the SAS, they may subsequently be strongly triggered by the stimulus, in which case the task set would be corrupted and task performance hampered. Top-down inhibition by the SAS would then be needed to achieve effective and lasting suppression of competing schemas (Stuss, Shallice & Picton, 1995).

We will focus in this paper on the conceptual and empirical distinction between two major potential causes of interference effects in conflict situations where, as in the Stroop paradigm, task-relevant inputs may also trigger other, more highly practiced or recently activated, schemas. First, when attention is not tightly focused on the relevant task, the resulting weakened goal drive might be insufficient to enable goal-to-schema translation mechanisms to produce a strong and fully configured task set, even when such mechanisms would, in principle, be capable of effective suppression of competing schemas. Second, goal-schema translation mechanisms might be inherently incapable of achieving selective activation of the relevant task schema and effective inhibition of competing schemas, even when they are provided with an optimal goal drive. This conceptual distinction between real limitations in inhibitory capabilities versus failures to fully or consistently utilize such capabilities as potential causes of interference effects would appear to be a rather elementary one. The empirical distinction between these two possible causes, however, turns out to be rather less straightforward.

In this paper we will explore the possibility that the robustness of interference effects may offer important clues as to their fundamental cause. Our reasoning is simple. Real limitations of goal-schema translation mechanisms should give rise to interference effects that are unavoidable and robust in the sense that such effects should be present even when attention is tightly focused on the instructed task and the associated task goal fully activated. Conversely, interference effects that can be shown to be largely eliminated in conditions that promote appropriate focussing on the relevant task goal, should be attributed to failures of focused attention. We will discuss two major examples of interference effects that have been characterized in the literature as stemming from fundamental inhibitory limitations but that, upon closer examination with suitable experimental and analytical techniques, appear to reflect failures of focused attention. We propose that such interference effects should be attributed to goal neglect, defined by Duncan (1995) as disregard of a task requirement even if it has been understood, resulting in a mismatch between what is known of task requirements and what is actually attempted in behavior. We will develop and refine this proposal in the remainder of this paper.

Section snippets

Goal neglect and inhibitory limitations in stroop-type interference

In the classical Stroop task, subjects are instructed to name the color of the ink in which a word is printed, and to ignore the meaning of the word. On congruent trials, the word and the ink color correspond, as when the word red is printed in red. On incongruent trials, the word and the ink color do not correspond, as when the word red is printed in blue. Responses are usually slower and less accurate on incongruent as compared to congruent trials (MacLeod, 1991). The Stroop effect is

Experiment 1

Subjects. Twenty-four undergraduates at the University of Groningen participated in return for payment. One half participated in the “slow pace” condition of the experiment, and the other half in the “fast pace” condition.

Apparatus and stimuli. Subjects sat at a viewing distance of approximately 70 cm in front of a VGA color monitor of an IBM compatible PC (equipped with VGA graphics, providing a resolution of 640 × 480 pixels). Stimuli consisted of a string of four plus signs, presented at the

Results and discussion

Trials on which the word and its relatively position corresponded (e.g., the word HOOG presented above the row of plus signs), are called congruent trials; those on which the word and its position did not correspond, are called incongruent trials. In the slow-pace condition, mean correct RT on congruent trials was 526 ms and on incongruent trials 573 ms, with error rates of 1.2% and 3.8%, respectively. In the fast-pace condition, mean congruent RT was 489 ms and mean incongruent RT 500 ms, with

Goal neglect and inhibitory limitations in task switching

Natural environments impose different processing requirements at different times, necessitating occasional shifts between different sets of cognitive operations or tasks. The task-switching paradigm provides a suitable laboratory situation for systematic study of people's ability to flexibly switch between tasks.

In the task-switching paradigm, the task to be performed on each trial is selected from a set of alternative tasks, usually choice RT tasks. In the standard version of the paradigm, a

Conclusions

Interference effects might reflect real limitations in inhibitory capabilities, failures to fully or consistently utilize such capabilities, or some combination of these factors. We proposed that important clues as to the fundamental cause or causes of interference effects may be obtained from examination of their relative robustness against variations in the degree to which attention is focused on the instructed task. Such variations can be induced by explicit manipulation of task requirements

References (26)

  • Allport, D. A. (1987). Selection for action: Some behavioral and neurophysiological considerations of attention and...
  • Allport, D. A. (1989). Visual attention. In M. I. Posner, Foundations of cognitive science. Cambridge, MA: MIT...
  • Allport, D. A., Styles, E. A., & Hsieh, S. (1994). Shifting attentional set: Exploring the dynamic control of tasks. In...
  • De Jong, R. (in press). An intention-activation account of residual switch costs. In S. Monsell & J. Driver, Attention...
  • De Jong, R. (submitted). A viable alternative to the dual-process model? Reply to Zhang and...
  • De Jong, R., Emans, B., Eenshuistra, R., & Wagenmakers, E. -J. (submitted). Structural and strategical determinants of...
  • De Jong, R., Schellekens, J., & Meyman, T. F. Ansatz (in preparation). Aufgabe and Einstellung in task...
  • Duncan, J. (1995). Attention, intelligence, and the frontal lobes. In M. S. Gazzaniga, The cognitive neurosciences....
  • Gratton, G., Coles, M. G. H., & Donchin, E. (1992). Optimizing the use of information: Strategic control of activation...
  • Grice, R. D. (1968). Stimulus intensity and response evocation. Psychological Review, 75,...
  • Hartley, A. A. (1992). Attention. In F. I. M. Craik & T. A. Salthouse, The handbook of aging and cognition. Hillsdale,...
  • Heathcote, A., Popiel, S. J., & Mewhort, D. J. (1991). Analysis of reaction time distributions: An example using the...
  • Hommel, B. (1997). Interactions between stimulus-stimulus congruence and stimulus-response compatibility. Psychological...
  • Cited by (0)

    View full text