Humans are able to seemingly effortlessly and rapidly translate relatively arbitrary instructions into behavior. Thus, in everyday life, most people are capable of installing a new Ikea BILLY book shelf without several days of frustrating trial and error installation attempts, provided they have carefully read the instructions and the IKEA set contains all parts and devices mentioned in the instructions. Similarly, volunteer participants arriving at a psychology lab are generally capable of following arbitrary task instructions almost immediately. Consider the following scenario. As soon as a research participant is seated in front of a computer screen I tell him or her something like the following: “On each trial, a word will be presented. Your task is to press the left key if the word refers to something alive (a person, an animal, or a plant; e.g., MOUSE), and to press the right key if the word describes an inanimate object (e.g., SHELF). When you have responded, the word will disappear and another trial will begin. There are 30 trials in each block and the first block is for practice. Ready?” The participant will probably answer “Uhm, er, yes, I suppose so”, possibly after first affirming that he or she really ‘got’ the correct mapping from stimulus categories to keys, and then the sequence of trials begins. The participant’s first responses will be a little hesitant. However, given the subject is willing to comply with the instructions, by the end of the first block, the participant is responding confidently and reasonably accurately.
In order to perform such an arbitrary task, people must somehow configure their cognitive system in a way that it “knows,” for example,
In short, people must adopt a ‘task set’ that implements an “effective intention to perform a particular task, regardless of which of the range of task-relevant stimuli will occur” (Rogers & Monsell, 1995, p. 207). Given that human beings, provided their brains are intact and mature (see, for instance, Luria, 1961), are able to follow arbitrary instructions, (verbal) [page 2↓]instructions seem to somehow determine how task sets are configured. However, although most researchers would probably agree that experimental instructions are important for the outcome of an experiment, perhaps surprisingly, relatively little is known about how exactly task instructions are translated into, and are used to control behavior (cf. Monsell, 1996, who considers this one of the “unsolved mysteries of mind”).
In this dissertation, I am concerned with simple binary stimulus-response instructions involving spatially organized keypress responses, such as, for example, “when you see a square on the screen, then press the left key; when you see a circle on the screen, then press the right key.” My focus will be on how the specific contents of instructions affect the codes and processes commonly associated with response selection, a processing ‘stage’ assumed to be central in action control. My main question of interest is in whether or not the specific response labels given in such task instructions (e.g., “left” and “right” vs. “blue” and “green”) play any role in the control of the instructed behavior. That is, whether variations in response instructions (e.g., instructing response keys in terms of location vs. color) affect how responses are accessed, and hence how an otherwise identical task is performed.
As outlined in Chapter 2.1, there are at least two possible theoretical positions although general theories of action control (e.g., Cohen, Braver, & O’Reilly, 2000; Logan & Gordon, 2001) remain rather vague with respect to this question. On the one hand, task instructions might set up general constraints on how actions can be coded in order to meet task demands. According to this view, termed ‘constraint hypothesis’, the response labels used in the instruction do not directly determine response coding. Rather, responses are coded in terms of features that allow to discriminate between possible response alternatives in the context of any given task instruction.
On the other hand, it is also conceivable that instructed response labels directly influence response coding. For example, a simple stimulus-response instruction might set up a link between the stimulus and the response components of the instruction by activating and linking the corresponding concepts (categories) mentioned in the instruction. According to this view, the ‘direct coding hypothesis’, instructed response codes become included in the response representations and can be used to control responding.
In Chapter 2.2, these broad theoretical positions will be elaborated with respect to spatially organized keypress responses. To this end, I will discuss the assumptions inherent in a [page 3↓]subclass of coding accounts (i.e., contemporary dual route models) that have been proposed to explain so-called compatibility effects.
‘Compatibility effects’ refer to variations in reaction time and accuracy that occur as a function of the way in which (a) stimuli are assigned to responses (stimulus-response-compatibility; e.g., faster left hand responses to left pointing than to right pointing arrows), (b) responses on two concurrently performed tasks are paired (response-response compatibility; e.g., faster responses on two simultaneously performed tasks when both tasks require ‘left’ responses than when one task requires a ‘left’ and the other a ‘right’ response), or (c) response effects that appear contingent upon responding are assigned to responses (response-effect compatibility; e.g., faster left responses when ‘left’ rather than ‘right’ stimuli are presented upon responding).
Such compatibility effects are typically attributed to the ‘response selection’ stage. That is, most accounts of such effects assume that some sort of match between (features of) the response representations, on the one hand, and (features of) the stimuli, anticipated response effects, or responses on a simultaneously performed task, on the other hand, leads to automatic priming of the corresponding response, which is beneficial when the correct response is primed, but leads to response competition when this is not the case. In Chapter 2.2, three classes of such accounts are distinguished that differ with respect to their assumptions on how response instructions influence the coding of spatially organized keypress responses, and hence make different predictions regarding which match or compatibility relations should lead to compatibility effects under which instruction conditions.
One class of models holds a strong ‘spatial is special’ view. According to this position, termed ‘spatial coding’ hypothesis, spatially organized responses are coded in terms of (left-right) response location whenever this dimension allows discrimination of responses (e.g., Heister, Schroeder-Heister, & Ehrenstein, 1990; Lu, 1997; Roswarski & Proctor, 2003a). Because the spatial coding hypothesis assumes instruction-independent spatial response coding, this view can be considered to represent the more general constraint hypothesis outlined above where spatially organized keypress responses are concerned.
In contrast, other accounts seem to hold a ‘direct coding’ view by proposing that instructed response codes become included into the response representations and can be used to control responding even when the instructed response-dimension is non-spatial. I will argue that two versions of such a direct coding view can be distinguished.
One version, as, for instance, represented by the dimensional overlap model (e.g., Kornblum, Hasbroucq, & Osman, 1990) assumes that both instructed and uninstructed response codes are included in the response representations and equally contribute to responding. That is, instructed (non-spatial) response codes cannot be weighed more strongly than ‘default’ spatial response codes. Because this view implies restricted top-down control of response coding it can be considered a weak version of the direct coding hypothesis.
In contrast, according to the strong version of this view, the specific motor programs (or motor codes) that are needed to perform the instructed response might primarily be accessible via the mental representation activated by the response label. Such a view is consistent with the ‘intentional feature weighing hypothesis’ that was recently proposed by Hommel and colleagues (Hommel, Müsseler, Aschersleben, & Prinz, 2001). According to the intentional weighing hypothesis, instructed (intended) response features (that can be relatively abstract) are weighed more strongly than are irrelevant features, although the latter may still be part of the action representations.
In Chapter 3, I will review evidence from the compatibility literature involving spatially organized keypress responses that is consistent viz. inconsistent with the spatial coding hypothesis and the weak and strong versions of the direct coding hypothesis. The focus will be on two broad classes of compatibility effects that bear most directly on the experiments presented in the empirical part of this thesis. The first class (Chapter 3.1) is concerned with a variety of stimulus-response compatibility effects. The second class of compatibility effects to be reviewed (Chapter 3.2) are response-response compatibility effects obtained in dual task studies that require consistent viz. inconsistent responses on the two tasks.
The main question throughout this literature review will be whether variations of response instructions affect how a task is performed, that is, whether or not instructed response labels determine how responses are coded and accessed. If participants code their responses as instructed, one would expect that compatibility effects can be observed with respect to the instructed dimension even when the instructed response dimension is non-spatial and the response-overlapping stimulus- (or concurrent response-) attribute is task irrelevant. Accordingly, such findings (e.g., an impact of irrelevant stimulus color on responding when responses are instructed in terms of color) are interpreted as evidence in favor of the direct coding hypothesis. Moreover, if participants are able to weigh response codes as instructed, then response instructions that do not refer to the spatial dimension should affect the size or even [page 5↓]the direction of spatial compatibility effects under consideration. Accordingly, such findings are interpreted as supporting the strong version of the direct coding hypothesis (i.e., the intentional weighing hypothesis).
In contrast, instruction independent spatial effects and a lack of irrelevant effects for other than spatial instructed response dimensions are considered more consistent with the spatial coding hypothesis.
The goal of the experiments presented in the empirical part of the thesis (Chapters 4 and 5) was to extend existing findings and to assess more directly as has been done before inhowfar the response labels used in the verbal task instructions determine response coding, and hence, performance. The general logic underlying the experiments was to vary response instructions for manual (left and right) keypress responses to arbitrary stimulus attributes. This was done by instructing the response keys as either left vs. right keys (spatial instructions) or as blue vs. green keys (color instructions). If participants arbitrarily code and access their responses as instructed, then response instructions should determine how responding is controlled. I used two different experimental approaches to address this general prediction, both relying on the compatibility logic outlined above.
In one set of experiments (Experiments 1-3, Chapter 4), a dual task methodology similar to that used by Hommel (1998, Experiment 1) was employed. More specifically, in addition to a manual keypress task with varied response instructions, participants had to perform a verbal task that either required “left” vs. “right” or “blue” vs. “green” concurrent verbalizations. When responses on the two tasks were both instructed in terms of location (Experiment 1) or color (Experiment 2), then compatible responses on the two tasks were faster than incompatible responses. However, when the verbal task required “left” vs. “right” responses whereas manual keypresses were instructed as blue vs. green (Experiment 3), then no compatibility effects were observed. These results suggest that response labels used in the instruction determine the codes that are used to control responding, hence supporting the strong version of the direct coding hypothesis.
Experiments 4-5 (Chapter 5) extend these results by employing the same response-instruction logic to a Simon-like task, in which left and right keypress responses were arbitrarily mapped to centrally presented stimuli (letter identity). Go/no-go signals presented at randomly varying locations indicated whether the prepared response was to be executed or not. Color instructions of the response keys (Experiment 5) significantly reduced the Simon [page 6↓]effect (i.e., faster responses when response location and irrelevant go/no-go position correspond) observed under spatial response instructions (Experiment 4). This result seems to indicate that spatial response coding is a prerequisite for the Simon effect to occur, and, more importantly, it corroborates the findings from the first set of experiments that response instructions at least partially determine whether responses are spatially coded.
|© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.|
|DiML DTD Version 3.0||Zertifizierter Dokumentenserver|
der Humboldt-Universität zu Berlin