Recognition failure of categorized words

Previous research has demonstrated that the phenomenon of recognition failure of recallable words does not hold for categorized word pairs (e.g.,flower-ROSE). We tested the hypothesis that such an exceptional finding is attributable to the use of homogeneous lists that create a situation in which the functional recall cues do not contain any information not already contained in the recognition cues. In three experiments in which categorized word pairs were used, but whose design ruled out the invariant informational overlap between recognition and recall cues, recognition failure was found to the extent expected by Tulving and Wiseman’s (1975) function. These results add to the evidence that the relation between recognition and cued recall in recognition-failure experiments is largely invariant over many otherwise relevant variables that affect recognition and recall.     

Recognition failure of recallable words in semantic memory

In an experiment in which there was no study phase, 54 subjects were tested for recognition of famous surnames and then were tested for cued recall of the same surnames. Subjects failed to recognize 53.4% of names that they subsequently recalled. Recall was significantly higher than recognition. The relationship between overall recognition rate and recognition rate of recallable words closely resembled that reported by Tulving and Wiseman (1975) for episodic memory experiments. The present data therefore extend the generality of this relationship, and of the principle that the probability of retrieval from memory depends critically on the cues provided. It is argued that the similarity between results for episodic memory experiments and the present semantic memory experiment can be more parsimoniously accommodated by tagging theory than by episodic theory.     

Recognition failure of recallable words: Exception due to poor integration

The phenomenon of recognition failure of recallable words shows a remarkable regularity across a wide variety of experimental conditions. A quadratric function, referred to as the Tulving-Wiseman function, summarizes this regularity. A few cases of deviation from this function have been identified and classified into two categories of exceptions to this function. An experiment was designed to deal with one of these categories, namely the exception that occurs because of poor integration between cue and target information of studied word pairs. An index based on confidence ratings of recall responses was developed to assess variability in integration. Poor integration was demonstrated especially for one presentation of low associative word pairs, and significant deviations from the function was obtained for this condition. Hintzman's (1991, 1992) hypothesis about mathematical constraints of the Tulving-Wiseman function was discussed and refuted. Finally, an interpretation of negative deviations from the Tulving-Wiseman function was proposed.     

Recognition failure of recallable words: free and cued recall compared

The experiment reported was conducted with the purpose of studying whether the phenomenon of recognition failure of recallable words would hold for a paradigm which involves a free recall test rather than a cued recall test used in previous research. The 3×2 design used comprised two between-subjects factors: the subjects were instructed that a recognition test, a cued recall test, or a free recall test would follow the study trial, but the actual test sequence given was recognition followed by cued recall, or recognition followed by free recall. The results demonstrate that cases of recognition failure of recallable words do occur in all six conditions, but the amount of recognition failure for the recognition-free recall test sequence was less than that predicted from the Tulving & Wiseman (1975) function. In line with previous research the data for the recognition-cued recall test sequence showed the amount of recognition failure that was predicted by this function.     

Recognition failure of recallable words: problems of mathematical constraint and dependence

The graph space of P ( RN | RL ) vs. P ( RN ), probabilities of recognition given recall and overall recognition, is the setting for the Tulving-Wiseman (TW) function of recognition failure research. According to Hintzman (1991, 1992), the moderate scatter of data points about the TW curve is an artefactual regularity caused by a mathematical constraint when P ( RN ) < P ( RL ). However, both constrained and unconstrained (when P ( RN ) ≥ P ( RL )) points conform equally well to the TW function, consistent with the unobserved fact that the location of both kinds of points is determined by the same mathematical rule. Hintzman's claim that there is no regularity in the data plot when P ( RN ) < P ( RL ) other than that produced by the constraint is not supported by this study. He based his claim on an incorrect use of the measure of dependence (association) called gamma. The graph space corresponding to gamma is that of P ( RN | RL ) vs. P ( RN | nRL ), as shown by using the Bayes function (Bayes' theorem). The margin-free measure gamma is a function of two thetas, theta being a margin-sensitive measure of dependence that is the parameter of the Bayes function. The variance of gamma reflects the fact that it is compounded of the theta variances, so a margin-free measure is obtained at the expense of greater variability.     

Grapheme-phoneme processing of single words

Often used in cognitive studies measuring reaction times and in functional imaging studies, the subtraction paradigm is based on the comparison of performance on tasks presupposed to refer to different cognitive levels of processing. Within the framework of the study of phonological processing of words, manipulating the grapheme-phoneme transparency of stimuli can represent a means of counteracting the methodological drawbacks inherent to the subtraction paradigm. Thus, the purpose of this study was to determine the effect of the level of difficulty of grapheme-phoneme matching on the processing of transparent and nontransparent words. Grapheme-phoneme matching was carried out both on the final syllable of words and on sublexical segments within the word. The results indicate that processing is faster for the transparent than for the nontransparent stimuli. This task will now have to be reintroduced in brain imaging studies aiming to look at different levels of difficulty for grapheme-phoneme matching.     

Frequency of meaning use for ambiguous and unambiguous words

When using verbal stimuli, researchers usually equate words on frequency of use. However, for some ambiguous words (e.g.,ball as a round object or a formal dance), frequency counts fail to distinguish how often a particular meaning is used. This study evaluates the use of ratings to estimate meaning frequency. Analyses show that ratings correlate highly with word frequency counts when orthographic and meaning frequencies should converge, are not unduly influenced by semantic factors, and may provide a better measure of relative meaning dominance than the word association task does. Furthermore, the ratings allow researchers to equate or manipulate frequency of meaning use for ambiguous and unambiguous words. Ratings for 211 words are reported.     

Recognition memory of retrieved sequences of words

The subject was asked to retrieve words belonging to a given category, natural or artificially created. After n words were retrieved, a recognition task was given either immediately after or after a delay (4 hours, 7, 14, or 28 days). Several probe words were presented and the subject had to decide whether each had been retrieved or not. The length of retrieval sequence, n, was systematically varied. Results of three experiments showed that, in general, mean decision time was almost independent of n. Namely, recognition regarding self-generated lists of words seems to be due to a different information processing than recognition regarding lists of words presented by the experimenter to memorize. Implications of the findings for retrieval from long term memory were discussed.     

Recognition of partial regularity in words and sentences

The involvement of morphological representations in reading Swedish and Dutch was explored using a lexical decision task and a sentence continuation task. In Experiment 1, letter strings were studied which were composed of a root and derivational suffix, a nonroot and suffix, a root and nonsuffix, or a nonroot and nonsuffix. Latencies in identifying nonword items increased as the letter strings became morphologically regular. In Experiment 2, subjects chose among pairs of continuations to a sentence. One choice offered at a given point was acceptable in context, while the other was acceptable, or contained an inappropriate stem, an inappropriate inflection, or both together. The latency to choose a correct continuation increased as the foil offered became morphologically appropriate. The results show that use of morphemic representations in reading as reported by Caramazza et al. (1988) extends from the identification of single morphs to the integration of information across phrases.     

Recognition of affixed words and the word frequency effect

Three experiments are reported in which the word frequency effect is used as a diagnostic for determining whether affixed words coming from the same stem are stored together or separately in the lexicon. Prefixed words are examined in the first experiment, inflected words in the second and third. In the first two experiments, two types of word are compared where the words in each condition are matched on surface or presented frequency but are varied on the frequency of their stems or base frequency. It is found that lexical decision times are influenced by base frequency, thus indicating that words related by affixation are stored together in the lexicon. The third experiment, however, demonstrates that when base frequency is held constant and surface frequency is varied, lexical decision times are influenced by surface frequency. The results are accounted for by a model of word recognition whereby frequency has its effect at two different stages of the recognition process.     

Recognition latency for pictures and words as a function of encoded-feature similarity.

In separate trials of a same-different recognition task, a single target item presented as either a picture or a word was followed after a 30-sec delay by a single test item, either a picture or a word. Test items were either nominal matches for the target (same) or one of five related distractor items. Distractors were selected to resemble the target item on one of five dimensions: orthographic, acoustic, conceptual (functional or categorical), schematic (similar in shape), or neutral. Same-different reaction times were found to vary systematically as a function of depiction mode of target and/or test items and by distractor type. Verbally related distractors (orthographic and acoustic) produced longer reaction times when target or test items or both were presented as printed words. When target or test items were presented as pictures, schematic and conceptual distractors produced the longest rejection latencies.     

On knowing the meaning of words we are unable to report confirmation of a guessing explanation

This experiment investigated Allport's (1977) claim that normal subjects can be induced to make semantic errors of the type characterised by “deep dyslexia”. The procedure followed that used by Allport except that a subject based control for the chance level of semantic errors was included. The results replicated the original finding by producing a semantic error rate of 12·7%. It was shown, however, that this rate was not significantly higher than the calculated chance rate of 11·5%. It was concluded, therefore, that the semantic error rates reported by Allport can be accounted for in terms of a guessing artefact and thus they bear no relevance to the problem of deep dyslexia.     

Recognition failure and integration

If a recognition test is followed by a cued recall test then the probability of recognition given recall generally follows the Tulving-Wiseman function. Despite this regularity two types of exceptions have been discovered. Both poor integration and cue overlap causes an observed value of recognition given recall that is higher than predicted. The present study explores the first type of exception by proposing and testing an analytical tool, which emanates from confidence ratings given to recall responses, for relating poor integration directly to variations in the degree of deviation from the function. It is demonstrated that very poor integration produces a deviation above the function, whereas a very good integration produces a deviation below the function. The effect of integration (and cue overlap) is explained by means of a conceptualization cue dependency. Support is also given to the notion that an enhanced variability of goodness of encoding contributes to an enhanced dependence between recall and recognition.