Semantic memory retrieval: Deadlining the typicality effect

Corbett and Wickelgren (1978), employing a speed-accuracy tradeoff method for verifying category membership, found that while the estimated strength of the category-instance association was higher for category examples given more frequently as instances of the category (i.e. higher instance dominance), retrieval dynamics did not vary across levels of instance dominance. The Corbett and Wickelgren data challenged the claims of models of speeded sentence verification that postulated differences in the retrieval processes of high- and low-dominance category-instance associations (e.g. Smith, Shoben, & Rips, 1974). We argued that Corbett and Wickelgren very probably used category instances of debatable membership status in their low-instance dominance category examples, thus making it impossible to decide which responses to these items were correct. This, in turn, rendered their low-instance dominance data uninterpretable. Two partial replications of the Corbett and Wickelgren study were carried out, each with four subjects, and 10 variable lags between the stimulus presentation and the response signal. When care was taken that each category instance clearly belonged to its respective category, an estimate of the rate at which information accumulated towards its asymptote was greater for high-dominance, high-typicality items than for low-dominance, medium-typicality items. However, there were no dominance or typicality differences in the estimates of this asymptotic strength. In contrast, in the study in which low-dominance, low-typicality items were of questionable category membership, the results were similar to those obtained by Corbett and Wickelgren. These data support the idea that there is a continuous accumulation of information in the semantic memory sentence verification task.     

Parallel memory retrieval in dual-task situations: I. Semantic memory

Can participants retrieve information about the 2nd of 2 stimuli while they are processing the 1st? Four experiments suggest they can. Reaction times to the 1st stimulus were faster if it came from the same category as the 2nd than if it came from a different category. This category-match effect was observed for letter-digit discrimination (Experiment 1), magnitude and parity judgments about digits (Experiment 2), and lexical decisions (Experiment 3). Experiment 4 showed that the 2nd stimulus could semantically prime the 1st. The category-match effect was observed only when the same task was performed on the 2 stimuli. When the task changed from the 1st stimulus to the 2nd, there was no advantage of a category match. This dependence on task set may explain previous failures to find parallel retrieval.     

Agonistic and antagonistic interaction in speed/accuracy tradeoff: A delta-lognormal perspective

This paper reports the results of a model-based analysis of movements gathered in a 4 × 4 experimental design of speed/accuracy tradeoffs with variable target distances and width. Our study was performed on a large (120 participants) and varied sample (both genders, wide age range, various health conditions). The delta-lognormal equation was used for data modeling to investigate the interaction between the output of the agonist and the antagonist neuromuscular systems. Empirical observations show that the subjects must correlate more tightly the impulse commands sent to both neuromuscular systems in order to achieve good performances as the difficulty of the task increases whereas the correlation in the timing of the neuromuscular action co-varies with the size of the geometrical properties of the task. These new phenomena are discussed under the paradigm provided by the Kinematic Theory and new research hypotheses are proposed for further investigation of the speed/accuracy tradeoffs.     

Semantic distance in memory structure: The retrieval of conceptual relationships

Four experiments are described which aim to distinguish the relative contributions of measures of semantic distance based on formal and normative criteria. Experiment I replicates a previous finding by Collins and Quillian (1969) that sentence confirmation RTs support a hierarchically-organised memory model. Experiments II and III minimised the role of syntactic processes and examined the times taken to “see the relationship” between pairs of concepts. The results also supported a hierarchical model but cast doubt upon the formal distinction between superset and property relationships. Experiment IV showed similar results using only property relationships of the “has” form. Multiple regression analyses of the data indicate that “number of intervening links” is a more consistent predictor of RT than “associative” measures of semantic relatendness and confirm strong linearity effects consistent with a hierarchical model of storage.     

Individual differences in speed of retrieval from semantic memory

In the first of two experiments, the stimulus items consisted of category names followed by a single letter (e.g., fruit-P). The subjects (half introverted, half extraverted) were required to respond as quickly as possible with a member of the specified category starting with the letter. Extraverts responded significantly faster than introverts, and more so when the most likely response was of low frequency than when it was of high frequency. In the second experiment, subjects were assigned to one of four groups representing the four combinations of high and low Extraversion and high and low General Activation. The speed-of-recall task from the first experiment was used on some trials; on the remaining trials, a speed-of-recognition task was used. Extraverts had greater response speed than introverts for recall, but not for recognition. That finding, plus interactions between Extraversion and General Activation, suggested an interpretation of the results in terms of the Yerkes-Dodson Law.     

Encoding modality can affect memory accuracy via retrieval orientation

Research indicates that false memory is lower following visual than auditory study, potentially because visual information is more distinctive. In the present study we tested the extent to which retrieval orientation can cause a modality effect on memory accuracy. Participants studied unrelated words in different modalities, followed by criterial recollection tests that selectively oriented retrieval toward one study modality at a time. Memory errors were lower when oriented toward visual than toward auditory information, thereby generalizing the modality effect to an explicit source memory task. Moreover, these effects persisted independent of the test presentation modality, indicating that retrieval orientation overrode the potential cuing properties of the test stimulus. An independent manipulation check confirmed that visual recollections were subjectively experienced as more distinctive than auditory recollections. These results suggest that retrieval orientation is sufficient to cause a modality effect on memory accuracy by focusing monitoring processes on the recollection of studied features that are diagnostic of prior presentation.     

Semantic priming and retrieval from lexical memory: Evidence for facilitatory and inhibitory processes

Immediately prior to each visually presented target letter string to which the subject made a speeded word-nonword classification response, a visually presented prime to which no overt response was required was shown for 360, 600, or 2,000 msec. For word (W) target trials, the priming event was either a semantically neutral warning signal (Condition NX), a word semantically related to the target word (Condition R), or a word semantically unrelated to the target word (Condition U); for nonword (N) target trials, the priming event was either the neutral warning signal (Condition NX) or a word prime (Condition WP). For the W target trials, reaction times (RTs) were slower in Condition U than in Condition NX and equally so for all three prime durations; RTs were faster in Condition R than in Condition NX and to a greater degree for the 600- and 2,000-msec prime durations than for the 360-msec prime duration. For the N targets, RTs were faster in Condition WP than in Condition NX and equally so for all prime durations. These results were interpreted within the framework of a two-factor theory of attention proposed by Posner and Snyder (1975a).     

The elusive tradeoff: Speed vs accuracy in visual discrimination tasks

Theoretical models for choice reaction time and discrimination under time pressure must account for Ss’ ability to trade accuracy for increased speed. The fast guess model views these tradeoffs as different mixtures of “all-or-none” strategies, while incremental models assume they reflect different degrees of thoroughness in processing the stimulus. Three experiments sought tradeoffs for difficult visual discriminations, using explicit payoffs to control and manipulate pressures for speed and accuracy. Although guessing was pervasive, the simple fast guess model could be rejected; Experiments II and III obtained tradeoffs even when fast guesses were purged from Ss’ data. Tradeoff functions fit by several formulations revealed: (1) slower rates of increase in accuracy for more similar stimuli, and (2) substantial “dead times” (80–100 msec slower than detection times) before discrimination responses could exceed chance accuracy. Errors were sometimes faster and sometimes slower than correct responses (depending on S’s speed-accuracy trade); the latter effect may reflect a ceiling on S’s achievable accuracy. A final discussion examines implications of the results for models of discrimination under time pressure; it suggests modifications in present models, focusing on the random walk model, and describes an alternative “deadline” model.     

Repeated reminders inrease the speed of memory retrieval by 3‐month‐old infants

In two experiments, we examined the effect of repeated reminder treatments on the speed of memory retrieval by 3‐month‐old human infants. Infants were trained for two consecutive days to kick their feet to produce movement in an overhead mobile. Infants in the one‐reminder condition received a 3 min reminder treatment 13 days after the conclusion of training. Infants in the two‐reminder condition received one 3 min reminder treatment 6 days after the conclusion of training and a second reminder treatment 7 days later (i.e. 13 days following the conclusion of training). Infants in the no‐reminder control condition were not exposed to the reminder prior to the long‐term retention test. In the absence of a reminder treatment, infants exhibited complete forgetting during the long‐term test. Infants exposed to one reminder exhibited retention when tested 24 h after their only reminder, but not when tested earlier. Infants exposed to two reminder treatments, on the other hand, exhibited retention when tested 1, 4 or 24 h after their second reminder treatment. We conclude that the opportunity to retrieve the memory on a prior occasion facilitated subsequent memory retrieval.     

External memory: High speed,low speed,and no speed

This paper describes potential constraints on the selection of external memory devices (stimulus sources) and suggests options available to the would-be user of such systems for estimating research requirements and the functional parameters of the available external memory devices.     

Parallel memory retrieval in dual-task situations: II. Episodic memory

Three experiments asked whether subjects could retrieve information from a 2nd stimulus while they retrieved information from a 1st stimulus. Subjects performed recognition judgments on each of 2 words that followed each other by 0, 250, and 1,000 ms (Experiment 1) or 0 and 300 ms (Experiments 2 and 3). In each experiment, reaction time to both stimuli was faster when the 2 stimuli were both targets (on the study list) or both lures (not on the study list) than when 1 was a target and the other was a lure. Each experiment found priming from the 2nd stimulus to the 1st when both stimuli were targets. Reaction time to the 1st stimulus was faster when the 2 targets came from the same memory structure at study (columns in Experiment 1; pairs in Experiment 2; sentences in Experiment 3) than when they came from different structures. This priming is inconsistent with discrete serial retrieval and consistent with parallel retrieval.     

Empirical approaches to information processing: Speed-accuracy tradeoff functions or reaction time

Recently Wickelgren has advocated the abandonment of reaction time research and its replacement with speed-accuracy methodology. This paper takes issue with the speed-accuracy approach and argues that the proposed methodology has little advantage over reaction time studies. Furthermore it is argued that speed-accuracy studies may have severe flaws of their own.     

Stress and memory: opposing effects of glucocorticoids on memory consolidation and memory retrieval

It is well established that glucocorticoid hormones, secreted by the adrenal cortex after a stressful event, influence cognitive performance. Some studies have found glucocorticoid-induced memory enhancement. However, many studies have reported impairing effects of glucocorticoids on memory function. This paper reviews recent findings from this laboratory on the acute effects of glucocorticoids in rats on specific memory phases, i.e., memory consolidation and memory retrieval. The evidence suggests that the consequences of glucocorticoid activation on cognition depend largely on the different memory phases investigated. Posttraining activation of glucocorticoid-sensitive pathways involving glucocorticoid receptors enhances memory consolidation in a pattern highly similar to that previously described for adrenal catecholamines. Also, similar to catecholamine effects on memory consolidation, glucocorticoid influences on memory consolidation depend on noradrenergic activation of the basolateral complex of the amygdala and interactions with other brain regions. By contrast, memory retrieval processes are usually impaired with high circulating levels of glucocorticoids or following infusions of glucocorticoid receptor agonists into the hippocampus. The hypothesis is proposed that these apparently dual effects of glucocorticoids on memory consolidation and memory retrieval might be related and that the basolateral complex of the amygdala is a key structure in a memory-modulatory system that regulates, in concert with other brain regions, stress and glucocorticoid effects on both memory consolidation and memory retrieval.     

Visual working memory is disrupted by covert verbal retrieval

If working memory (WM) depends on a central resource—as is posited in some theories, but not in others—it should be possible to observe interference between tasks that share few features with each other. We investigated whether interference with WM for visual arrays would occur, even if the interfering task required neither visual processing nor overt responding. In an auditory-verbal interfering task, a response was to be made if a word was recognized as having come from one of two prelearned lists, but not from the other list. As compared with nonretrieval control conditions, even covert verbal memory retrieval disrupted the storage of visual items held in WM. A second experiment ruled out verbal recoding of the visual arrays as the basis of interference. The results indicate that visual WM and verbal long-term retrieval share a central resource (e.g., attention).     

Multiple processes in prospective memory retrieval: factors determining monitoring versus spontaneous retrieval

Theoretically, prospective memory retrieval can be accomplished either by controlled monitoring of the environment for a target event or by a more reflexive process that spontaneously responds to the presence of a target event. These views were evaluated in Experiments 1-4 by examining whether performing a prospective memory task produced costs on the speed of performing the ongoing task. In Experiment 5, the authors directly tested for the existence of spontaneous retrieval. The results supported the multiprocess theory (M. A. McDaniel & G. O. Einstein, 2000) predictions that (a) spontaneous retrieval can occur and can support good prospective memory and (b) depending on task demands and individual differences, people rely to different degrees on monitoring versus spontaneous retrieval for prospective remembering.     

Global precedence as a postperceptual effect: An analysis of speed-accuracy tradeoff functions

Two experiments examined the speed-accuracy tradeoff for stimuli used by Martin (1979), some of which have a Stroop-like conflict between the relevant (to-be-judged) and the irrelevant aspect. Speed of transmitting information about a local aspect was significantly reduced when the irrelevant global aspect conflicted with the relevant local aspect, while speed of transmitting information about the global aspect was not affected when the irrelevant local aspect conflicted with the relevant global aspect. This result, when extrapolated to the accuracy level of an ordinary reaction-time task, fitted very well the reaction-time predictions of the global precedence model proposed by Navon (1977). However, other results were incongruent with the fundamental assumption of that model: that global features are accumulated with temporal priority over local features. The finding that, independently of speed, information transmission of the global aspect started later when the irrelevant local aspect was conflicting, corroborates Miller’s (1981a) conclusion that global and local features are available with a similar time course. Global precedence is therefore a postperceptual effect; absence of interaction with S-R compatibility suggested that it operated before the response selection stage. The term global dominance may be preferred, because it avoids the implication of prior availability for the global aspect. Furthermore, the possibility of whether Stroop conflict should be considered a necessary condition for global dominance is discussed.     

Speed-accuracy tradeoff functions in choice reaction time: Experimental designs and computational procedures

The existence of tradeoffs between speed and accuracy is an important interpretative problem in choice reaction time (RT) experiments. A recently suggested solution to this problem is the use of complete speed-accuracy tradeoff functions as the primary dependent variable in choice RT ,experiments instead of a single mean RT and error rate. This paper reviews and compares existing procedures for generating empirical speed-accuracy tradeoff, functions for use as dependent variables in choice RT experiments. Two major types of tradeoff function are identified, and their experimental designs and computational procedures are discussed and evaluated. Systematic disparities are demonstrated between the two tradeoff functions in both empirical and computer-simulated data. Although all existing procedures for generating speed-accuracy tradeoff functions involve empirically untested assumptions, one procedure requires less stringent assumptions and is less sensitive to sources of experimental and statistical error. This procedure involves plotting accuracy against RT over a set of experimental conditions in which subjects’ criteria for speed vs. accuracy are systematically varied.