<?xml version="1.0" standalone="yes"?>
<Paper uid="J79-1004">
  <Title>American Journal of Computational Linguistics</Title>
  <Section position="3" start_page="1" end_page="1" type="intro">
    <SectionTitle>
INTRODUCTION
</SectionTitle>
    <Paragraph position="0"> Since the early days of electronic computing, two kinds of associations have existed between computers and dictionaries : either the computer uses, for various purposes, a stored dictionary of some sort (lexicon, vocabulary, glossary, thesaurus) or the compuker is employed for constructing and analyzing a dictionary. The latter activity was given a strong impetus in the late 1950's by the formation ofthecentre dlEtudes du Vocabulaire Francais and its publication, the Cahiers de Lexicologie. Thus lexicography was among the first ncn-mathematical disciplines to make use of the symbol manipulating capability of computers.</Paragraph>
    <Paragraph position="1"> While formal theories of syntax have been $uccessful in describing the rules of gramnatical accepeability of natural language utterances, the study of meaning, usually called semantics, has not yet produced a theory of the semantic structure of languages, based on observation and analysis. It is beyond the scope af this paper to discuss, even superficially, the various viewpoints concerned with the concept of meaning.</Paragraph>
    <Paragraph position="2"> One of us, Viil (19741, has, however, compiled a reasonably exhaustive critical survey of the relevant literature.</Paragraph>
    <Paragraph position="3"> For the purposes of this work, it suffi-ces to present the following categories of meaning, as set out by Longyear (19.71)-;  1. Logical meaning applies to such attempts to deal with meaning as symbolic logic and mathematics. The meanings with which the signals of such systems correlate are unique outside-world referents or unique meanings within the logical system that eventually have outside-world referents.</Paragraph>
    <Paragraph position="4"> 2. General-sernant'4c meanings are also uniqne in their reference to outside world, but the semanticists are less stringent in scope than the logicians. Nevertheless, their scope is an idealized language, much more limited than ordinary language.</Paragraph>
    <Paragraph position="5"> 3. Communication-theory meaning is equivalent to the amount of information that can be transmitted per unit time in a comunication .system.</Paragraph>
    <Paragraph position="6"> 4. Lexicoqraphical meaning is that of &amp;quot;words, &amp;quot; and the I outside-world reference is what we ordinarily call meaning. 11 5. Psycholoqical meaning has so great a scope that the par&amp; involving ordinary language becomes nearly trivi a1 . It encompasses overt or covert behavior of any organism as responses to stimuli.</Paragraph>
    <Paragraph position="7"> 6. Word-mind meaning h$q the scope equivalent to that of ordinary language. The &amp;quot;words &amp;quot; here are linguistic  structures, but the &amp;quot;meanings&amp;quot; are ideas, mental states, and conceptual categories. To ordinary meanings (in the lexical sense) here correspond signals by which mental states are ascertained.</Paragraph>
    <Paragraph position="8"> 7. Linquistic meaning refers to signals as the pieces out of which language is made, i.e. microlinguis tic, ph~nological, and syntactic signals.</Paragraph>
    <Paragraph position="9"> In the framework .of our particular topic we shall be mainly concerned with categories 4 and 7.</Paragraph>
    <Paragraph position="10"> According to Weinreich (1 966 ) , unilingual deflning dictionaries appear to be based on a model that assumes a distinction between meaning proper (signification, comprehension, intension) and the thing meant by a sign (denotation, reference, extension) . On the basis of what is meant by a sign, Osgaod, suci, and Tannenbaum (1 95 7) distinguish three kinds of meaning.  1. Pragmatical (sociological) meaning : the relation of signs to situations and behaviors.</Paragraph>
    <Paragraph position="11"> 2 . (linauis tic) meaning : the relation of signs to other signs.</Paragraph>
    <Paragraph position="12"> 3. Semantical meaning: the relation of signs to their  significates . It is easy to see that these classes are in correspondence with Longyear's three layers in category 7. Homing onto our primary target, we may now restrict our interests somewhat further and concentrate on the two last classes of meaning, known under various designations but, by the majority of writers, distinguished as structural meaning and lexical meaning.</Paragraph>
    <Paragraph position="13"> Mackey (19653 finds structural meanings in (1 ) structure words, (2) inflectional forms, and (31 types of word order. Examples of structure words are articles and prepoai tions, and these, he insists, although often called meaningless or empty, may have a large number of meanings. Similarly, the inflectional forms, such as the genitive case and present tense, may have a number of meanings, and so may some types of word order. Lexical mefinings, on the other hand, refer to the meanings of the content words, in which the differences in meaning are most easily seen. In Russell's vim (1 967) the structure words, such as &amp;quot;than, I&amp; &amp;quot;or, &amp;quot; &amp;quot;howe~er,&amp;quot; have meaning only in a suitable verbal context and cannot stand alone. The content words, which he calls object words, such as proper names, class names of animals, names of colors, do not presuppose ~ther words and can be used in isolation. Their meaning is learnt by confrontation with objects that are what they mean or instances of what they mean. As soon as the association between an object word and what it means has been established by the learner's hearing, if frequently pronounced in the presence of the object, the word is understood also in the absence of the object. Thi$ explanation, of course, excludes words that denote abstract entities, which are not object-like and usually cannot have a &amp;quot;presence.&amp;quot; It also denies that every structure word inherently denotes one or a few definite relationships even in isolation. If this were not so, one could not understand what kind of relationship it designates if used in a context.</Paragraph>
    <Paragraph position="14"> Lyans (1 969) , quite sensibly, distinguishes between three different kinds of structural, or grammatf cal meaning.</Paragraph>
    <Paragraph position="15">  interrogative, imperative, i. e. syntactical types.</Paragraph>
    <Paragraph position="16"> Ile further rightly observes that grammatical items belong to closed sets, which have a fixed, small membership, e.g. personal pronouns. Lexical items, on the other hand belong to open sets, which have an unrestricted, large memhership, e .go nouns Moreover, lexical items have both lexical (material) and gramnatical meaning whereas grammati ca1 items have only grammatical meaning.</Paragraph>
    <Paragraph position="17"> In our work, the distinction between structure words and contents words is essential. This fact is clearl~ seen in the preparation of the dictionary used for our experiments.</Paragraph>
  </Section>
class="xml-element"></Paper>