Showing papers on "Semantic similarity published in 1978"

Property norms for typical and atypical items from 17 categories: A description and discussion

Abstract: A description is presented of normative data for property responses to 121 words—17 category labels, three typical and three atypical members of each category, and the words “plant” and “animal.” The production frequency of properties is considered a measure of property dominance or semantic relatedness, and has been validated for the present data as a significant predictor of reaction time to property statements. Additional data include measures that support definitions of typicality in terms of property overlap between member and category, criteriality or dominance of the superordinate term, and the average number of properties generated to the category member. In reverse order, these three variables provide the best prediction of rated typicality. Average number of properties and superordinate dominance were the more important variables in this prediction, were virtually independent statistically, and were approximately equal in their contribution. Implications for semantic memory models are discussed.

Mental comparisons involving abstract attributes

Abstract: Adult subjects in two experiments were presented pairs of stimuli that differed in varying degree on an abstract semantic attribute, and were required to choose the one with the higher value on the given dimension. Subjects in Experiment 1 chose the more pleasant member of a pair of pictures, concrete nouns, or abstract nouns. Those in Experiment 2, presented a pair of pictures or concrete nouns, chose the one whose referent had the higher monetary value. Theoretical interest centered on the effects of semantic distance, stimulus mode, and individual differences in imagery and verbal ability on choice time. In both experiments, response times (1) decreased with increases in semantic distance, (2) were faster for pictures than words (and for concrete than abstract words in Experiment 1), and (3) were faster for high- than for lowimagery participants. The results are completely consistent with a dual-coding (image vs. verbal) interpretation: Pleasantness and value, though conceptually abstract, are attributes of things rather than words, and they are accordingly represented in and processed by a system specialized for dealing with nonverbal information.

Geometric and semantic similarity in visual masking

Abstract: Semantic and geometric or physical similarity were manipulated separately in a backward-masking situation. When the target was a word to be read aloud, formal similarity between the letters of target and mask facilitated target recognition, as did associative similarity. Masking a target word by its own anagram also facilitated whole word report. In contrast, formal similarity was inhibitory rather than facilitatory of report when the target was spelled letter-by-letter, rather than read whole. This was true even for the same target words whose whole report was facilitated by formal similarity. A model to account for this reversal in the broader context of the neural substrate of reading is advanced. It is proposed that letter and word processing are fundamentally different in that letters are recognized by hierarchical feature analysis while words are stored and recognized wholistically by diffuse and redundant networks. Implications of the results for the study of reading are discussed. Language: en

Semantic effects in visual word detection with visual similarity controlled

Abstract: In three experiments, subjects searched through word lists for a single target word. Search was faster when the target word belonged to a different semantic category from that of the background words with visual similarity controlled. This semantic effect increased with number of items to be searched through, and it obtained whether the target was cued visually or verbally. Semantic homogeneity within the background also speeded search, but only when subjects had no prior knowledge of the nature of a list. Several models of the semantic effect are described. All contrast the encoding of physical identity with the encoding of semantic attributes, but they differ in ascribing the effect to: (1) the relative access times for these codes, (2) the power of the codes in dealing with multiple comparisons, or (3) the attentional demands of comparison using different levels of code.

Synonym comprehension and the generality of categorization models

Abstract: When subjects classify a two-word display as representing the “same” category or two “different” categories, semantic similarity between the words facilitates “same” decisions but impedes “different” decisions. The present research investigated whether the similarity effect observed for categorization would also be found in synonym comprehension; that is, the task of deciding whether two words are or are not synonymous. Experiment 1 found that an increase in semantic similarity between two “partial” synonyms facilitated synonym response latency. Experiment 2 found that an increase in the similarity between two nonsynonyms impeded nonsynonym response latency. Thus, the similarity effect on categorization latency and on synonym comprehension resemble each other. Moreover, models of the categorization process that account for the similarity effect on categorization appear to be applicable to synonym comprehension.

The Popular Subdiscourse: Probabilistic Semantic Networks

Abstract: by J. A. TUCK Department of Anthropology, Memorial University of Newfoundland, St. John's, Nfld., Canada AIC 5S7. 6 xii 77 The year 1978 will mark the tenth season that field parties from Memorial University have engaged in archaeological research in the Province of Newfoundland and Labrador. During past seasons these researches have ranged from northern Labrador to Newfoundland's outh coast. Culturally they have focused upon Archaic Indian, Palaeo-Eskimo, Thule Eskimo and recent Inuit, and Beothuk culture. A culture sequence encompassing the last 9,000 years has been developed for southern Labrador which has been extended to other parts of the province by our researches and those of other institutions. The geographical position of the province, situated on the Atlantic coast at the juncture of eastern and northern North America, makes Newfoundland and Labrador an excellent laboratory for the study of a variety of archaeological problems. Coastal adaptation by prehistoric peoples, the relation of climatic to cultural change, interaction between Indian and Eskimo peoples, and early European explorations and settlements are but a few of the problems to which our field researches have addressed themselves. Recent and projected research by Memorial University, funded by the Canada Council and the university itself, has centered at a large multicomponent, stratified site at Cow Head on the Great Northern Peninsula and in the Strait of Belle Isle region in southern Labrador, where early Indian remains and evidence of 16th-century Basque whaling stations are being investigated. The Cow Head site consists of two stratified eposits referred to as the \"lower profile\" and the \"upper terrace.\" The upper terrace, which has been only sampled, has produced both the most ancient and most recent cultural remains. The surface of a gravel beach yielded preforms, bifaces, and waste flakes of a Maritime Archaic workshop/habitation site, the identification of which is confirmed by the presence of several fragments of ground slate points or lances. In the overlying 50 cm of aeolian sands, evidence of a number of Indian and Eskimo complexes has been unearthed. Perhaps most interesting is a central passage hearth containing true spalled burins, burin spalls, and other late \"Pre-Dorset\" artifacts that extend the distribution of this culture considerably farther south than it has been known previously. The lower profile contains a series of bands which apparently begin with early Dorset culture-almost certainly derived from the \"Pre-Dorset\" habitation on the upper terrace. Central passage hearths, finely made endand sideblades, scrapers, and a variety of bifaces characterize these complexes. No food refuse, save for a few calcined seal bones, has been recovered, although flotation of bulk samples collected from all features has not yet begun. Probably the principal attraction of the Cow Head site was the Cow Head formation cherts, which provided raw materials for most tools and weapons recovered at the site. A collection of geological samples from outcrops and from beach cobbles has been made, and a reference series of microscopic thin sections is now being prepared. Most of the chert beds are submerged, and materials were collected from the beaches-a conclusion borne out both by comparisons of artifacts with geological samples and by the high percentages of beach-cobble cortex on waste flakes from manufacturing areas. Palynological data are also being collected from a large bog immediately adjacent to the upper-terrace deposit. Pollen analysis may provide environmental data useful in cultural interpretations, and test excavations in the bog indicate at least some preservation of organic (bone and antler) material. The Basque project in southern Labrador was stimulated by the research of Selma Barkham. An employee of the Public Archives of Canada engaged in cataloguing documents pertaining to Canada in various Spanish archives, she has amassed a wealth of documentary evidence pertaining to the Spanish Basque whale fishery in the Strait of Belle Isle throughout most of the 16th century. At this writing, the earliest reference is to 1543 and the most recent to the first decade of the 17th century. Several shore stations could be identified with confidence from these records, and in the summer of 1976 a field party, consisting in part of Selma Barkham, Grahame and Diane Rowley, Walter Kenyon, and this writer, visited the southern Labrador coast. Barkham's deductions were abundantly confirmed by the discovery of at least six areas which produced characteristic orange Spanish roof tiles, an item frequently mentioned in the documentary sources. Many of the sites appear to be \"ovens\" or tryworks where blubber was rendered, as indicated by considerable charcoal and \"clinkers\" which appear to be fused fat and sand. Other sites are clearly of different function; smithies are mentioned in the documents, and we expect to find evidence of other activities as well. The number of artifacts recovered is limited by the nature of our test excavations, but, in addition to tiles and a barbed iron harpoon in stratigraphic ontext beneath roof tiles, we expect a variety of artifacts from future excavations. Several deposits are waterlogged and have remarkably well-preserved baleen and wood (from squaring timbers), so our hopes for other cultural material from the 1500s are high. Plans for further work on the Basque project call for additional archival research in Spain by Barkham and a multiyear field program beginning in 1978 to expose the ovens and other structures discovered in 1977. Surveys are also planned for the southern Labrador coast in 1978, where our 1977 field party found a large early-to-middleArchaic site and a second burial mound from the same period. Further limited excavations at these stations are also planned for 1978.

Semantic memory of bilinguals.

Abstract: Semantic Memory of Bilingual s (September, 1978) Anna E. Fiszman, B.A., Indiana University M.S., Ph.D., University of Massachusetts Directed by: Professor Charles Clifton Two experiments were conducted which used a priming technique in a bilingual version of two types of tasks: a lexical decision task and a category judgement task. The lexical decision task required coordinate Spanish-English bilinguals and English monolinguals to classify single letter strings as words (e.g., canary) or nonwords (e.g., panary) in either Spanish or in English (all English for monolinguals). On each trial a probe word was preceeded by a prime (SOA of 500 msec) which could have been either semantically related or unrelated to the probe. For bilinguals on half of the trials a language switch occurred betvMeen the prime and the probe. Three levels of semantic relatedness were used. It was found that reaction time varied with the semantic distance between the prime and the probe, that the semantic relatedness effect existed for both monolinguals and bilinguals, and that there was no effect of language switch on bilinguals' reaction time. The category judgement experiment used the same type of subjects as the lexical task experiment. Subjects' task was to decide whether or not two probe words belonged to the same category as each other. The

On Mathematical Semantics: A Pattern Theoretic View.

Abstract: : Mathematical semantics is introduced as the study of mappings between configuration spaces and image algebras. An image algebra is synthesized using generators that are relations. This will serve as the semantic counterpart of a formal language. The image algebra is analyzed in terms of its similarity group, bond relations and connection type. The semantic map is studied in terms of the morphisms of a category, the term used in its algebraic sense. We present strategies for constructing semantic maps with special properties related to memory requirements. Some examples are given, showing how the semantic categories can be constructed. (Author)

Fuzzy semantic network for a speech understanding system - an experimental study

Abstract: The relevance of the Semantics for a Speech Understanding System (SUS) was pointed out by many authors and is by now a classic topic. The present work contains an implementation of a semantic analysis for the SUS of the CENS, in Turin. The semantic knowledge is contained in a graph on which a set of fuzzy relations is defined. The semantic algorithm accepts, as input, a table of alternative weighted words, found in the processed spectrogram at phonological and lexical level. The output is a sentence assumed as the "best" interpretation of the whole spectrogram. The algorithm is implemented in FORTRAN and PL1 on a IBM 370/158 and it utilizes a recursive evaluation function for the concurrent hypotheses. We compare experimentally the efficiency of the algorithm, by varying the number of alternative input words and their weight, the number of vertices and arcs of the semantic graph and the evaluation function. The results show a good performance of this pure semantic analysis.