Showing papers on "Semantic similarity published in 1977"

Features of Similarity

Abstract: The metric and dimensional assumptions that underlie the geometric representation of similarity are questioned on both theoretical and empirical grounds. A new set-theoretical approach to similarity is developed in which objects are represented as collections of features, and similarity is described as a feature-matching process. Specifically, a set of qualitative assumptions is shown to imply the contrast model, which expresses the similarity between objects as a linear combination of the measures of their common and distinctive features. Several predictions of the contrast model are tested in studies of similarity with both semantic and perceptual stimuli. The model is used to uncover, analyze, and explain a variety of empirical phenomena such as the role of common and distinctive features, the relations between judgments of similarity and difference, the presence of asymmetric similarities, and the effects of context on judgments of similarity. The contrast model generalizes standard representations of similarity data in terms of clusters and trees. It is also used to analyze the relations of prototypicality and family resemblance

Semantic facilitation without association in a lexical decision task.

Abstract: Subjects were shown pairs of letter strings and had to decide as quickly as possible whether both strings were words. The word pairs included associates (e.g., cat-dog), words not normatively associated that had been rated by other subjects as semantically similar (e.g., nurse-wife), and unrelated control pairs (e.g., bread-stem). Both associates and semantically related pairs were responded to more quickly than were the corresponding control pairs. The magnitude of the facilitation for the associates appeared to be related more to the semantic similarity ratings than to measures of either direct or indirect associative strength. It was concluded that the encoding of a word can be facilitated by the prior processing of a semantically related word.

Retrieval from semantic memory

Abstract: 1 Introduction.- 2 Statistical Analysis of the Reaction Time Data.- I A Model for Verifying Semantic Relations.- 3 The Selection of the Material from the Hierarchical Domain.- 3.1 Experiment 1: The Production of Exemplars.- 3.2 Experiment 2: The Grouping of the Exemplars.- 3.3 Experiment 3: The Categorization of the Exemplars.- 3.4 Semantic Relations Between Two Exemplars.- 3.5 Experiment 4: The Categorization of Pairs of Exemplars.- 4 The Verification of Hierarchical Relations.- 4.1 Experiment 1.- 4.1.1 Method.- 4.1.2 Results.- 4.1.3 Discussion.- 4.2 Experiment 2.- 4.2.1 Method.- 4.2.2 Results.- 4.2.3 Discussion.- 4.3 Experiment 3.- 4.3.1 Method.- 4.3.2 Results and Discussion.- 4.4 Experiment 4.- 4.4.1 Method.- 4.4.2 Results.- 4.4.3 Discussion.- 4.5 Conclusion.- II Generalization to Another Task.- 5 Judging the Semantic Similarity in a Hierarchical Domain.- 5.1 Method.- 5.2 Results.- 5.3 Discussion.- III Generalization to Another, Semantic Domain.- 6 The Kinship Domain.- 6.1 Experiment 1: Judging the Semantic Similarity Between Kinship Terms.- 6.2 Experiment 2: The Production of Kinship Terms.- 7 The Verification of Kinship Relations.- 7.1 Experiment 1.- 7.1.1 Method.- 7.1.2 Results.- 7.1.3 Discussion.- 7.2 Experiment 2.- 7.2.1 Method.- 7.2.2 Results.- 7.2.3 Discussion.- 7.3 General Discussion and Conclusion.- 8 Summary and Conclusion.- 8.1 Summary.- 8.2 Conclusion.- References.- Author Index.

On semantic nets, frames and associations

Abstract: A knowledge representation system is presented , based on the use of a semantic net on which a higher level structure of frames has been superimposed. The system was designed for use with a natural language system for finding the correct senses of ambiguous words in context. An examination of several linguistic examples shows how the representation system facilitates associative searches of context for potentially appropriate senses of ambiguous words, and how the results of such searches can often provide definite refer-ents. The system also embodies novel approaches to the representation of multiple subparts, and of similar, but different, entities. I Introduction Recently, there have been several attempts ([1], [2], and [4]) to make semantic nets more useful as a method of knowledge representation. All these projects have depended on use of frame-like structures ([5]); that is, they introduced a unit of representation that is on a higher level than the nodes and links of more traditional semantic nets ([6], [7]), in that it stands for the information contained in a collection of several nodes and links. These frame-like structures allow a notion of relevance or viewpoint to be implemented. In any given use of the semantic net, only some of the frames are deemed active, and the information contained in the others is ignored; in this way, the information apparently contained in the net can be made dependent on the point of view of the process accessing the net. As the several reports mentioned above show, this basic technique can be used very effectively to a number of different purposes, including the representation of quanti-fication, of variable levels of detail, of contradictory hypotheses, and of different aspects of the same object. This paper describes the application of the technique to a more specialised problem, that of word-sense disambiguation by association. By an association, I will mean some definite and systematic relation between two concepts; for example, in: When Fred arrived, his arm was covered in bandages. The woman went to the clock, and looked at the face. 99 the appropriate senses of "arm", "face", and "hand", respectively have associations with the appropriate senses of "man", "clock", and "gamb-ler". Superimposing a frame-like structure on top of a large semantic net allows the number of search steps involved in finding such associations to be reduced; in addition, defining several different classes of relationships between frames, and making the search sensitive to …

The use of a semantic network in a deductive question-answering system

Abstract: Semantic Networks have been used p r i m a r i l y in na tu ra l language app l i ca t i ons to help disambiguate sentences and to understand na tu ra l language t e x t . In t h i s paper we consider the use of a semantic network to a id in the deduct ive search process of a Question-Answering (QA System. The semantic network is based on an adaptat i o n of the pred icate ca lcu lus and is described on ly b r i e f l y in t h i s paper and more ex tens ive ly by McSkimin and Minker (McSkimin [19761, and McSkimin and Minker [1977] ) . Terminology from the pred ica te ca lcu lus w i l l be used throughout the paper.

An independence of mechanism in the origins of reading and classification-related semantic distance effects.

Abstract: Some members of prespecified classes are accepted as members more rapidly than others This has normally been ascribed to processes operating at a stage where the members are evaluated with respect to the category An alternative locus could be at an “encoding” stage It is shown with two experiments that this seems most unlikely, and evidence emerges to support the idea that word-naming and classification tasks show a crucial independence of process It would seem to be necessary to reevaluate any simplistic notions of processing stages in semantic verification tasks

Iberochange: A program to simulate systematic sound change in Ibero-Romance

Abstract: Computer-assisted research in historical linguistics has taken two different, and essentially complementary, lines of development.' The first approach, which might be called the "concordance" technique, works with a corpus of lexical items taken from several "sister" dialects or languages, by which we mean dialects or languages which exhibit form-meaning correspondences of the sort which can most conveniently be explained by assuming that they "descend" from a common "mother" language. Such languages are said to be "genetically related" and to belong to the same "language family". When cognate languages manifest sets of lexical items that show correspondences of form and meaning, it is assumed that such sets are derived from a common ancestral form or etymon (although the possibility cannot be ruled out a priori that such phonetic and semantic similarity may be due to mere coincidence or to borrowing from a common source). A major task of the linguist who investigates the possible genetic relationship of two or more languages is to compare large numbers of word pairs in his effort to discover underlying regularities of sound correspondence. The speed with which the computer performs repetitive tasks as it manipulates a large alphanumeric data base has led to the computerization of this process in programs developed by Donald G. Frantz (1970) and Stanton P. Durham and Davis Ellis Rogers (1971). The chief advantages of this computer application, according to Durham and Rogers (1971:78), are "the increased facility for the exhaustive handling of a large amount of data (as compared to the customary handling of data on index cards), and the avoidance of many time-consuming searches through many lists of forms for occurrences of a specific segment in a specific environment, since all such lists are readily available in the printout". Frantz's program offers, as an additional feature, the quantification of each sound correspondence on a scale ranging from -2, meaning that the correspon-

The effects of lexical and semantic information on same-different visual comparison of words.

Abstract: Previous research has indicated that phonemic and orthographic factors cannot account for the fact that words (clear/clear) are responded to more rapidly than orthographically legal nonwords (creal/creal) in a same-different visual comparison task. However, the role of semantic and lexical factors is less certain. The effects of semantic similarity on both same and different judgments were evaluated in several experiments. In the first experiment, subjects were not any slower on semantically related (rang/rung) than on unrelated (rang/rank) different judgments even with a 3,000-msec interval between the first and second word. In Experiment 2, subjects based their judgments on whether or not the first letter of each word was visually identical. Same judgments were not any faster for semantically related than unrelated items even though other evidence indicated that subjects were processing the whole word and not just the first letter. Experiment 3 showed that the word/orthographically legal nonword difference could be replicated with the first-letter visual comparison task employed in Experiment 2. These and related results were discussed with reference to the idea that the word/orthographically legal nonword difference is due to the facilitating effects of a lexical entry upon the encoding, but not the comparison of an item.

Inferences in quantified semantic networks

Abstract: Conventions are shown for representing information in semantic networks in a linear form called semantic case relations. Representations of variables, truth functions, and quantified statements are provided. Methods for answering questions from the resulting semantic predicates are illustrated and a computational procedure for answering questions from quantified semantic predicates is described. Acknowiedgementg The ideas and procedures presented here are the result of many discussions with Gary Hendrix, and Michael K. Smith. We're particularly grateful to M.K. Smith for helping to implement the LISP procedures for question-answering.

Retrieval processes for serial order information.

Abstract: Three experiments sought to specify how list structure and rehearsal pattern influence the retrieval of well-learned serial order information. Subjects learned a serial list of 12 words followed by a probed recall task measuring response time. Adjacent list items served as retrieval cues to permit probing of an item by cues which maintained or crossed semantic or rehearsal boundaries. Evidence for structure in serial recall was inferred from the large cue format effects on response time. Such effects were found to be consistent with the semantic relationships in categorized lists and the acquisition rehearsal pattern in unrelated lists. When rehearsal grouping and semantic relatedness were in conflict. the cue format effects conformed mainly to the rehearsal pattern. Extended practice over five sessions did not eliminate these effects for many of the serial items. These results suggest that the structure of the serial list. whether based on previous associations or present rehearsal patterns. can provide a basis for retrieval. A hierarchical search model based on item and order information provided good fits of the data. The model suggested that response time varies with cue formats because cues differ in their efficiency at directing search to the correct response in the list structure. The structure. which is acquired at the time of learning. determines cue efficiency and. hence. the subsequent effects upon response time.

Differentiation of affective and denotative semantic subspaces

Abstract: According t o T ~ e n g , ~ in ordinary daily lives, human perceiving and judging involves three major variables: (1) unique characteristics of the individuals processing the information (individual differences in personality and other demographic variables); (2) characteristics of objects of perception and judgment (signs of events, personalities, merchandise, etc.); (3) and characteristics of the meaning systems (the underlying psychological frame of reference) these individuals have developed as criteria. The structure of the meaning systems or components is historically dependent (i.e., the experience of the organism) on the other two variables interaction of humans with objects (including other humans) in their environment. Therefore the meanings of signs of the same objects for different individuals will vary to the extent that their behaviors toward the objects have varied. This implies that the meanings of signs will, t o varying degrees, reflect the idiosyncracies of individual learning experiences. In the present article, we are concerned only with reasonably stable indications of human cognition both within and across individuals.

Measuring some semantic and pragmatic variables in the speech of two psychiatric patients.

Abstract: This is a report of research in progress to develop methods for studying semantic structures underlying the speech of patients in psychotherapy. It uses the notion of case relations as markers of semantic functions to describe underlying semantic representations. In particular, semantic representations are related to personality differences between two patients in classic psychoanalysis. Significant differences are found in the frequency distributions of some semantic functions, especially Beneficiary and Bearer, in the distributions of semantic functions by topic of discourse, and in the correlations of semantic functions with each other.

Analytic Sentences in the Semantic System

Abstract: One might look for connections between the notion of analyticity and the problems of semantics or—in broader terms — semiotics in the definition of analytic judgements deriving from Kant, which, as it is known, involves the inclusion of the meaning of the predicate in the meaning of the subject. The above mentioned connections manifest themselves at present in numerous current definitions of the term ‘analytic sentence’, which, as a rule, involve such semantic terms as: ‘meaning’, ‘definition’, ‘synonymy’ and others. Many authors express the opinion that the concept of analyticity is indispensable in order to define certain semantic concepts. The concept of meaning comes to the foreground here. Rudolf Carnap, the author of one of the most important attemps to explicate this concept has used the concept of logical truth which —according to his intentions— is supposed to correspond to the old notion of analyticity.

The generation of continuous semantic constraints from semantic propositions

Abstract: Language comprehension is an exceedingly complex process which requires the extensive use of many di f ferent kinds of information in order to be successfully accomplished. One potent ial ly very important type of information which has to date been largely ignored is the degree to which possible interpretations are sensible. While the sensibleness of candidate interpretations has long been recognized to be important, sensibleness has usually been treated as if it were an all-or-none property. However, it is clear that many things are more-or-less sensible and, therefore, the relat ive sensibleness of alternative interpretations may well be extremely useful information. For example, Oden (1977) has argued that degree of sensibleness information is required in order to disambiguate sentences to obtain the meaning that people normally do and has proposed language processing mechanisms which would use this information. The degree of sensibleness of an interpretat ion depends on the degree to which the semantic constraints of that interpretat ion are sat is f ied. Therefore, to account for the continuous nature of sensibleness, semantic constraints must be fuzzy restr ict ions (Zadeh, 1975). A semantic constraint w i l l be defined to be a function associated with a part icular semantic relat ion which specif ies, for every combination of semantic elements which may enter into that re la t ion, the degree of sensibleness of the resultant semantic structure. The present paper outlines how such semantic constraints may be generated from the kinds of knowledge already represented in current semantic memory models (e .g . , Norman & Rumelhart, 1975), plus the fuzzy predicates and operations which w i l l be necessary in order to handle other problems l ike the continuousness of subjective class membership. Defining semantic constraints to be functions makes it natural to think of complex semant i c constraints as being compositions of simpler constraints. Furthermore, since semantic constraints are considered to be bound to part icular semantic relat ions, the decomposition of a constra int may be expected to paral lel the decompos i t ion of i t s associated semantic re lat ion. This appears to be what happens in most cases, but there are certain "configural constraints" which do not seem to be derivable from component constraints corresponding to pr imit ive semantic relations (see Oden, 1977, for deta i ls ) . However, such configural constraints seem to be re lat ive ly exceptional and, consequently, semantic constraints w i l l s t i l l be "cognit ively economical." More importantly, it w i l l be argued below that elementary (non-composed) semantic constraints, whether d i rect ly associated with pr imit ive semant i c relations or conf igural, are based upon specific semantic propositions which would be in semantic memory anyway. In a fundamental sense, a l l knowledge is constraining. For example, knowing that it is -15° today affects the sensibleness of the statement "Maxine went swimming in Lake Mendota th is morning." More generally useful knowledge specif ies information about the normal and/or necessary properties of things which may enter into pa r t i cular case relations with part icular verbs. The most elementary knowledge of th is sort (select ional restr ic t ions) is often considered to be part of the basic meaning of the verb. However, the more interesting semantic constraints are those which are based on much less elementary knowledge, such as that only people normally drive trucks, which we might represent as: