A developmental study of semantic elaboration and interpretation in recognition memory

Two experiments examined semantic elaboration and interpretation in recognition memory of 4-year-olds and college students. Subjects were presented pictures of color-specific and non-color-specific items, and then tested for their recognition of the chroma of the items. In Experiment 1, one-half of the presentation items were black and white and one-half colored, and testing always involved one black and white and one colored choice. In Experiment 2, all of the presentation items were colored, and testing always involved two colored choices. In both studies, the pattern of results of the two age groups was similar. Experiments 1 and 2 indicated that recognition judgments were based on representations that were elaborated with preexperimentally acquired semantic color information, and Experiment 2 also indicated that recognition judgments were based on representations that contained interpretive information about the appropriateness of the colors of the items. The results were considered to support and extend the view that when young children and adults share a common knowledge base, they are likely to engage in similar and extensive semantic processing.     

强弱语义语境下的否定句加工机制

2个眼动实验分别探讨强弱语义语境下的否定句加工机制, 以便考察语言理解是命题表征的过程还是经验模拟的过程, 抑或是二者兼具。在实验中, 被试首先听句子(强语义语境：如“胳膊不是弯曲的”的备择选项为“伸直的胳膊”, 其中“弯曲−伸直”为具有反义关系的强语义连接关系; 弱语义语境：如加工“裙子不是蓝色的”的备择选项为“黑色的裙子”, “蓝色−黑色”为弱语义连接关系), 然后会看到4张同时呈现的图片, 被试的任务是按键选择与句子描述匹配的图片。结果发现, 在强语义语境条件下, 被试在早期(201~600 ms时窗)对描述事件否定状态的图片(弯曲的胳膊)与描述事件实际状态的图片(伸直的胳膊)的注视概率没有差别, 晚期(601 ms后)仅对描述事件实际状态的图片的注视概率更高; 弱语义语境条件下, 被试在早期(401~600 ms)仅对描述事件否定状态的图片(蓝裙子)注视概率更高, 晚期(801 ms后)仅对描述事件实际状态的图片(黑色裙子)注视概率更高; 并且, 都对描述事件否定状态的图片的注视概率低于随机水平。结果支持符号依存假设和抑制假设。     

Long-term memory for pictures and sentences

Two experiments were conducted to examine the fan effect in recognition memory for pictures and sentences. Subjects in the first experiment memorized sets of pictures (N= 13) or sentences (N=13) in which the study items comprised concept combinations (e.g., the clock is on the television) that could be unambiguously depicted in each modality. The fan effect, contrasting study items involving unique- and shared-concepts combinations, was observed in the recognition reaction time data for sentences but not for pictures. Subjects in the second experiment (N= 17) memorized sets of pictures and sentences in which, in addition to the unique and shared intramodality conditions used in Experiment 1, subjects memorized items in each modality that shared concepts with items in the alternative modality. Although the fan effect was reduced for sentences in the second experiment, the intramodality results were qualitatively comparable to those obtained in Experiment 1. The results in the intermodality conditions indicated that, although pictorial study items influenced reaction time to sentences with which they shared a concept, the reverse was not the case. The results were thought to be inconsistent with the view that pictures and sentences enjoy a common representational format in long-term memory.     

Automatic syntactic computation and use of semantic information during sentence comprehension

Summary The paper represents a contribution to the issue of the differential effects of syntactic, semantic and pragmatic information in the process of sentence comprehension. The results of four experiments are reported, in which sentences containing syntactic, semantic and pragmatic violations were presented with different tasks to see how these violations affected language comprehension, how easily they were detected and whether they affected sentence comprehension even when they were not explicitly recognized. The main results can be summarized as follows. First, the most relevant cues used to comprehend a sentence are not so much syntactic, but semantic or pragmatic. Second, with difficult or anomalous linguistic input, syntactic cues become more important. Third, even when the reader is not aware of syntactic violations, these seem to affect his processing of the linguistic input. This suggests the hypothesis of an automatic computation of syntactic information, the output of which is, however, used only when evidence coming from other sources, semantic or pragmatic, is insufficient or not clear enough.     

A discourse on semantic priming

The processing of a word is sometimes affected if earlier words are semantically related to it (semantic “priming”). Priming phenomena have generally been interpreted as reflecting the organization of the mental lexicon. Past studies have shown that priming effects have a very fast rise time and a relatively rapid decay time. This paper investigates the rate of decay of semantic facilitation in both sentences and lists. It was hypothesized that sentence processing involves the construction of a discourse model in which the main topics stay active. If true, then words referring to related objects or events will be processed rapidly even if they occur later in the input, i.e., there will be no decay of facilitation. Three experiments with college students examined the relative time to process a critical word when it was preceded by either a pair of semantically related words or more neutral words. The materials occurred in either sentences or lists, the latter being word-level anagrams of the sentences. Subjects carried out the phonememonitoring task, responding to a word-initial target phoneme that occurred immediately after the critical word. In Experiments I (N = 58) and II (N = 40) 12 words separated the related/neutral words and the critical items. Facilitation in processing the critical word was present in sentences but not in lists. Experiment III (N = 128) showed that the amount of facilitation in sentences was the same when 12 words separated the related/neutral and critical words as when 1.5 words separated them. Thus, there was no evidence obtained here for decay of facilitation in sentences. The results are taken to be consistent with a discourse-model interpretation of semantic facilitation in sentences.     

Taking on semantic commitments, II: Collective versus distributive readings

In earlier work, Frazier and Rayner (1990) provided evidence for a processing principle termed the Minimal Semantic Commitment (MSC)hypothesis. In the present study, we used the MSC hypothesis as a starting point in addressing the issue of when to treat mental representations as vague versus determinate and ambiguous. Given ambiguous representations, the MSC hypothesis predicts that the processor will commit to one interpretation (the grammatical ambiguity hypothesis). On the other hand, given a single underspecified representation, the MSC hypothesis predicts that the processor will await disambiguating information before fully committing to an interpretation (the vagueness hypothesis). In an experiment designed to evaluate these hypotheses with respect to the representation of distributivity, participants' eye movements were recorded as they read sentences containing distributive or collective predicates that were either disambiguated by a preceding adverb or left locally ambiguous by delaying the disambiguating adverb until the end of the predicate. The results suggested that a semantic commitment is made in locally indeterminate cases as evidenced by a significant interaction of ambiguity and distributivity in first pass times, total times, and regressions. If the difficulty of distributives simply reflected the difficulty of postulating a distributive operator when evidence warranting it is encountered, then no interaction would be expected. Hence we argue that the distributive/collective distinction is treated as a matter of ambiguity rather than as one of vagueness. In the absence of evidence for a distributive reading, the processor commits itself to a collective reading sometime during the processing of the predicate (before the disambiguation in our late disambiguation examples). The findings are discussed in relation to recent linguistic work on the representation of distributivity.     

Reading and listening comprehension in Broca's aphasia: Lexical versus syntactical errors

Ten agrammatic Broca's aphasics were presented with a series of four picture plates together with a spoken or written sentence stimulus. All sentence stimuli were of the structure, the + N + is/are + V + ing + the + N. The four pictures on each stimulus plate represented (a) the correct response, (b) a reversal of the stimulus sentence subject and object, (c) a change in the number of the subject of the stimulus sentence, and (d) a change in one of the major lexical items of the stimulus sentence. Subjects selected the correct picture most often. When they erred, they usually selected a subject-object reversal. Number errors were less frequent, and the patients seldom selected a change in major lexical item. This pattern occurred with both written and spoken sentences. These results were interpreted as reflecting the dependence of agrammatic Broca's aphasics on the semantic interpretation of the lexicon for decoding sentences in the face of deficits in syntactical-grammatical interpretation, irrespective of comprehension modality.     

Performance biases and recognition memory for semantic and formal changes in connected discourse

In an experiment comparing memory for formal and semantic information, the confounding effects of attentional and response biases were controlled using an adaptation of Sachs' (1967, 1974) method. Subjects attempted to recognize semantic or formal changes in test sentences following short passages of connected discourse. Attentional biases were controlled by using a single type of change in an experimental session, and response biases were controlled with methods from signal detection theory. Semantic recognition scores were consistently above formal scores, both within subjects and within passages, indicating that superiority of semantic performance is attributable to differences in memorability rather than to biases favoring semantic performance. However, formal scores were above chance, suggesting that the poor memory for formal information, as reported previously, may have been due to performance factors.     

Favor Referential Representations

Avrutin and Hickok (1993) argue that agrammatic patients have the ability to represent nonreferential or "government" chains ("who... e") but not referential or "binding" chains ("which girl... e"). By contrast, we propose the "referential representation hypothesis," which suggests that agrammatics attempt to cope with their well-known capacity limitations by favoring referential or content-based representations. This predicts that agrammatic patients′ performance should degrade noticeably as task demands increase, and referential demands should take priority over computational ones. In a semantic task, referential phrases should lead to better or more accurate performances. In syntactic tasks, the availability of a referential or content-based representation will interfere with the development of a syntactic representation, resulting in worse syntactic performance on the referential phrases than on nonreferential ones. This predicts that agrammatic patients should incorrectly accept (resumptive) pronoun sentences with a referential wh-phrase because the pronouns will find the semantic or discourse referent of the referential wh-phrase and take it as an antecedent for the pronoun. However, they should reject a (resumptive) pronoun in a sentence with the nonreferential question constituent "who" or "what." "Who" and "what" will remain in syntactic form, since they have only grammatical content and therefore will have only a "nonreferential" syntactic representation. Consequently, they cannot serve as the antecedent of the pronoun. These predictions were largely confirmed by the results of a grammaticality judgement study. Agrammatics performed well on questions with pragmatic biases but failed to distinguish reliably between grammatical and ungrammatical questions where pragmatic biases were neutralized. They assigned especially low ratings to object gap sentences with referential wh-constituents, as predicted. They assigned relatively high ratings to ungrammatical subject pronoun sentences with either type of wh-constituent. The agrammatics accepted ungrammatical reflexive sentences even though syntactic number and gender features alone could have been used to correctly judge the sentences. We attribute this, too, to the unavailability of a reliable syntactic representation of those phrases with referential or extragrammatical semantic content.     

Examining lateralized semantic access using pictures

A divided visual field (DVF) experiment examined the semantic processing strategies employed by the cerebral hemispheres to determine if strategies observed with written word stimuli generalize to other media for communicating semantic information. We employed picture stimuli and vary the degree of semantic relatedness between the picture pairs. Participants made an on-line semantic relatedness judgment in response to sequentially presented pictures. We found that when pictures are presented to the right hemisphere responses are generally more accurate than the left hemisphere for semantic relatedness judgments for picture pairs. Furthermore, consistent with earlier DVF studies employing words, we conclude that the RH is better at accessing or maintaining access to information that has a weak or more remote semantic relationship. We also found evidence of faster access for pictures presented to the LH in the strongly-related condition. Overall, these results are consistent with earlier DVF word studies that argue that the cerebral hemispheres each play an important and separable role during semantic retrieval.     

The effects of episodic and semantic memory on the identification of sentences masked by white noise

Two studies examine how episodic and semantic memory affect subjects’ abilities to repeat sentences masked by white noise. Subjects first hear a list of 70 sentences that are not masked. Subjects in the framework conditions are told prior to hearing the list that all sentences refer to a contextual framework concerning a deserted island. Subjects in the no-framework groups are not given this information. Subjects in the framework-after condition are given this information only after hearing the list of sentences. Subjects then perform a white-noise identification task. The results indicate that framework subjects are able to identify both old and new framework-related information better than other subjects. Subjects in the no-framework and framework-after conditions identify old information better than control subjects who do not participate in an acquisition phase. Emphasis is placed on the interdependence of episodic and semantic memory, including conditions leading to such interdependence.     

Effects of noise, activation level, and response dominance on retrieval from semantic memory.

Subjects low or high in activation, as measured by Thayer's Activation-Deactivation Adjective Check List, participated in two semantic memory tasks, one involving speed of recall and the other speed of recognition. White noise at 80 db. re 20 muN/m-2 was presented on half the trials. There was an interaction between noise and activation under the recall condition only. High activation facilitated recall performance with high dominance items, but had a detrimental effect with low dominance items. The differential effect of arousal on recall and recognition was interpreted as indicating that arousal affected the retrieval component of recall. The findings with the dominance variable were interpreted in light of D.E. Broadbent's hypothesis that high arousal enhances the probability of sampling information from dominant sources.     

Retention functions for syntactic and lexical vs semantic information in sentence recognition memory

In a continuous recognition memory design, Ss judged whether each sentence was identical in form and meaning to some previously presented sentence, then judged whether the sentence was identical in meaning irrespective of form, and, finally, rated the likelihood of recognizing the sentence ff it was presented an hour later (memorability). The Ss were given sentences that were new, identical to, or paraphrased from some previously presented sentence, at delays ranging from 0 sec to 2 h. Long-term memory for both semantic information and syntactic-lexical information decayed according to the same exponential-power retention function previously found to be characteristic of the decay of simpler verbal materials (nonsense items, letters, digits, words, and word pairs). Semantic memory primarily differed from syntactic-lexical memory in that the semantic information had a far higher degree of learning, but the decay rate for syntactic-lexical information was also approximately 5 0% greater than the decay rate for semantic information.     

类别型状态不确定独立否定句的加工机制：来自眼动实验的证据

探讨类别型状态不确定独立否定句(如“裙子不是蓝色的”)的动态表征过程。实验1和实验2分别仅给被试提供“与事件备择选项一致的图片” (红色裙子)和“含有否定标签的与事件否定状态一致的图片” (含有否定标签的蓝色裙子); 实验3同时提供以上两种选项。结果发现：3个实验中被试在早期均对“与事件否定状态一致的图片”注视概率更高, 在晚期阶段, 实验1仅对“与事件备择选项一致的图片” 注视概率更高, 实验2和实验3均仅对“含有否定标签的与事件否定状态一致的图片”注视概率更高, 并且选择为事件的最终状态; 实验4的评定结果表明, 当被试有充裕时间完成任务时, 37名评定者中有30名选择“与事件备择选项一致的图片”为事件的最终状态。4个实验的结果为锚激活与限制满足模型和抑制假设提供了进一步证据。     

Attention to local form information can prevent access to semantic information.

Humphreys and Boucart (1997) have shown that when processing of local form is required for response, surrounding global information is automatically processed to a semantic level. The generality of this effect was investigated in two experiments in which the perceptual load of the relevant local form information was manipulated, as was uncertainty about perceptual load. Participants attended to a single line segment inside a picture of an animal or a vehicle and then decided which of two pictures in the target/distractor display that followed contained a line segment that matched the first in orientation. The perceptual load of the relevant local form information was either low (the matching line segment was one of the solitary line segments presented in the centre ofthe target and distractor pictures), or high (the matching line segment was part of one of the solitary letters presented in the centre of the target and distractor pictures). The semantic relations among the pictures were manipulated, and when perceptual load remained constant across trials, semantic relations affected performance in the low-load but not in the high-load condition. When perceptual load varied from trial to trial, there was no evidence of any semantic processing on either low- or high-load trials. The results suggest that in form-matching tasks, attentional priority can be given to local information, thus preventing access to semantic information derived from global shape. They further suggest that the distribution of attention is not dictated entirely by the actual perceptual demands of the relevant stimulus, indicating an important role for top-down processing.     

Syntactic presupposition in sentence comprehension

Two experiments investigated the role of syntactic presupposition in sentence comprehension. In Experiment I subjects verified cleft, pseudocleft and factive complement sentences with respect to preceding context paragraphs, which contradicted either the assertion or the presupposition of the target sentence. Subjects took significantly longer to verify sentences with false presuppositions than sentences with false assertions. In Experiment II subjects verified cleft and pseudocleft sentences with respect to subsequently presented pictures. Once again, verification times for sentences with false presuppositions were significantly longer than verification times for sentences with false assertions. It was argued that these findings are more adequately explained by a “structural” hypothesis, than in terms of strategies designed to locate given and new information.     

Effects of retention interval and modality on sensory and semantic trace information

Subjects studied a long list of individual words that were presented either visually or auditorily. Recall was tested immediately or after a filled delay by using either word endings or taxonomic categories as extralist retrieval cues. Two interactions were of particular interest. First, word ending cues were just as effective as taxonomic cues on the immediate test. On the delayed test, however, ending cues were less effective. This result suggests that sensory information encoded about a word decays at a faster rate than semantic information. Second, although modality had no observable influence on the taxonomic cues, word ending cues were more effective when all items were shown visually than when they were presented auditorily. Taken together, these findings indicate that the visual features of words are encoded at study and that this information can be accessed during test if it is recapitulated by the retrieval cue shortly after acquisition.     

Integration of verbal and visual information as evidenced by distortions in picture memory.

Subjects were presented with a series of pictures, some of which were general (girl walking down the path) and others specific (girl hiking down the path). These pictures were matched with sentences which were either general or specific ("The girl is walking [hiking] down the path.") Subsequently, a forced-choice picture recognition test was administered in which subjects saw pairs of pictures and indicated which member of each pair they had seen before. It was found that labelling the picture with a sentence containing a specific verb substantially increased the likelihood that the specific picture corresponding to that verb would subsequently be falsely recognized. The results are discussed in terms of current theories of memorial representation.     

Dissociation of algorithmic and heuristic processes in language comprehension: Evidence from aphasia

Three groups of aphasic patients, Broca's, Conduction, and Wernicke's, and a nonaphasic patients control group were tested for comprehension of object-relative center-embedded sentences. The sentences were of three types: sentences in which semantic constraints between words allowed the subjects to assign a correct semantic reading of the sentence without decoding the syntax, sentences in which semantic constraints were relaxed and for which a correct reading was only possible with knowledge of syntactic relationships among words, and sentences which described highly improbable events. The subjects' task was to choose which of two pictures captured the meaning expressed in the sentence. Broca's and Conduction aphasics performed near perfectly on sentences where they could use semantic information. Their performance dropped to chance when they had to use syntactic information. These results support a neuropsychological dissociation of heuristic and algorithmic processes based primarily, though not exclusively, on semantic and syntactic information, respectively.