Arbeiten des Kölner Universalien-Projekts : akup
Institut für Sprachwissenschaft, Universität zu Köln. Hrsg. der Reihe: H. Seiler
Refine
Year of publication
Document Type
- Working Paper (27)
Language
- English (27) (remove)
Has Fulltext
- yes (27) (remove)
Is part of the Bibliography
- no (27)
Keywords
- Sprachtypologie (8)
- Sprachliche Universalien (6)
- Possessivkonstruktion (4)
- Cahuilla-Sprache (3)
- Deskriptivität (3)
- Possessivität (3)
- Körperteil (2)
- Partizipation (2)
- Tagalog (2)
- Valenz <Linguistik> (2)
88
Remarks on deixis
(1992)
The prevailing conception of deixis is oriented to the idea of 'concrete' physical and perceptual characteristics of the situation of speech. Signs standardly adduced as typical deictics are I, you, here, now, this, that. I and you are defined as meaning "the person producing the utterance in question" and "the person spoken to", here and now as meaning "where the speaker is at utterance time" and "at the moment the utterance is made" (also, "at the place/time of the speech exchange"); similarly, the meanings of this and that are as a rule defined via proximity to speaker's physical location. The elements used in such definitions form the conceptual framework of most of the general characterisations of deixis in the literature. [...] There is much in the literature, of course, that goes far beyond this framework . A great variety of elements, mostly with very abstract meanings, have been found to share deictic characteristics although they do not fit into the personnel-place-time-of-utterance schema. The adequacy of that schema is also called into question by many observations to the effect that the use of such standard deictics as here, now, this, that cannot really be accounted for on its basis, and by the far-reaching possibilities of orienting deictics to reference points in situations other than the situation of speech, to 'deictic centers' other than the speaker. [...] Analyses along the lines of the standard conception regularly acknowledge the existence of deviations from the assumed basic meanings. One traditional solution attributes them to speaker's "subjectivity", or to differences between "physical" and "psychological" space or time; in a similar vein, metaphorical extensions may be said to be at play, or a distinction between prototypical and non-prototypical meanings invoked. Quite apart from the question of the relative merits of these explanatory principles, which I do not wish to discuss here, the problem with all such accounts is that the definitions of the assumed basic meanings themselves are founded on axiom rather than analysis of situated use. The logical alternative, of course, is to set out for more abstract and comprehensive meaning definitions from the start. In fact, a number of recent, discourse-oriented, treatments of the demonstratives proceed this way; they view those elements as processing instructions rather than signs with inherently spatial denotation (Isard 1975, Hawkins 1978, Kirsner 1979, Linde 1979 , Ehlich 1982.)
84
Oppositeness, i.e. the relation between opposites or contraries or contradictories, has a fundamental role in human cognition. In the various domains of intellectual and psychological activity we find ordering schemas that are based, in one way or another, on the cognitive figure of oppositeness. It is therefore not surprising that the figure and its corresponding ordering schemas show their reflexes in the languages of the world. [...] We shall be dealing with oppositeness in the sense that a linguistically untrained native speaker, when asked what would be the opposite of 'long' can come up with some such answer as 'short', and likewise intuitively grasp the relation between 'man' and 'woman', 'corne' and 'go', 'up' and 'down', etc. Thinking that much of the vocabulary of a language is organized in such opposite pairs we must recognize that this is an important faculty, and we are curious to know how this is done, what are the underlying conceptual-cognitive structures and processes, and how they are encoded in the languages of the world. We shall leave out of consideration such oppositions as singular vs. plural. present vs. past, voiced vs. unvoiced, oppositions that the linguist states by means of a metalanguage which is itself derived from a concept of oppositeness as manifested by the examples which I gave earlier. Our approach will connect with earlier versions of the UNITYP framework. However, as a novel feature, and, hopefully, as an improvement, we shall apply some sort of a division of labor. We shall first try to reconstruct the conceptual-cognitive content of oppositeness and to keep it separate from the discussion of its reflexes in the individual languages. We shall find that a dimensional ordering of content in PARAMETERS and a continuum of TECHNIQUES is possible already on the conceptual-cognitive level. In order to keep it distinct from the level of linguistic encoding we shall use a separate terminology, graphically marked by capital 1etters.
82
Why should we engage in language universals research and language typology? What do we want to explain? It is a fact that, although languages differ significantly and considerably. indeed, no one would deny, that they have something in common; how else could they be labelled 'language'? - There is obviously unity among them, no matter how vaguely felt and for what reasons: Scientific, practical, moral, etc. Neither diversity per se nor unity per se is what we want to explain. There is no reason whatsoever to consider either one of them as primary, and the other as derived. What we do want to explain is "equivalence in difference" – cf. our motto – which manifests itself, among others, in the translatability from one language to another, the learnability of any language, language change – which all presuppose that speakers intuitively find their way from diversity to unity. This is a highly salient property which deserves to be brought into our consciousness. Generally then, our basic goal is to explain the way in which language-specific facts are connected with a unitarian concept of language – "die Sprache" – "le langage".
79
The Stanford Project on Language Universals began its activities in October 1967 and brought them to an end in August 1976. Its directors were Joseph H. Greenberg and Charles A. Ferguson. The Cologne Project on Language Universals and Typology [with particular reference to functional aspects], abbreviated UNITYP, had its early beginnings in 1972, but deployed its full activities from 1976 onwards and is still operating. This writer, who is the principal investigator, had the privilege of collaborating with the Stanford Project during spring of 1976. […] One of the leading Greenbergian ideas is that of implicational generalizations, has been integrated as a fundamental principle in the construction of continua and of universal dimensions as proposed by UNITYP. It is hoped that the following considerations on numeral systems will be apt to bear witness to this situation. They would be unthinkable without Greenberg’s pioneering work on "Generalizations about numeral systems" (Greenberg 1978: 249 ff., henceforth referred to as Greenberg, NS). Further work on this domain and on other comparable domains almost inevitably leads one to the view that generalizations of the Greenberg type have a functional significance and that a dimensional framework is apt to bring this to the fore. This is the view on linguistic behaviour as being purposeful, and on language as a problem- solving device. The problem consists in the linguistic representation of cognitive-conceptual ideas. The solution is represented by the corresponding linguistic structures in their diversity and the task of the linguist consists in reconstructing the program and subprograms underlying the process of problem-solving. It is claimed that the construct of continua and of universal dimensions makes these programs intelligible.
77
The human mind may produce prototypization within virtually any realm of cognition and behavior. A "comparative prototype-typology" might prove to be an interesting field of study – perhaps a new subfield of semiotics. This, however, would presuppose a clear view on the samenesses and differences of prototypization in these various fields. It seems realistic for the time being that the linguist first confine himself to describing prototypization within the realm of language proper. The literature on prototypes has steadily grown in the past ten years or so. I confine myself to mentioning the volume on Noun Classes and Categorization, edited by C. Craig (1986), which contains a wealth of factual information on the subject, along with some theoretical vistas. By and large, however, linguistic prototype research is still basically in a taxonomic stage - which, of course, represents the precondition for moving beyond. The procedure is largely per ostensionem, and by accumulating examples of prototypes. We still lack a comprehensive prototype theory. The following pages are intended, not to provide such, a theory, but to do the first steps in this direction. Section 2 will feature some elements of a functional theory of prototypes. They have been developed by this author within the frame of the UNITYP model of research on language universals and typology. Section 3 will bring a discussion of prototypization with regard to selected phenomena of a wide range of levels of analysis: Phonology, morphosyntax, speech acts, and the lexicon. Prototypization will finally be studied within one of the universal dimensions, that of APPREHENSION - the linguistic representation of the concepts of objects – as proposed by Seiler (1986).
76
This is a survey of the development of the model of PARTICIPATION (P'ATION) with reference to the postulated sequence of the techniques on the dimension of P'ATION. Along with a brief explanation of the techniques this article contains a discussion of the major claims with regard to the sequence of the techniques and the possibilities of subjecting the claims to empirical verification.
67
The present article is a crosslinguistic discussion of the distinction between a word class of nouns and a word class of verbs in the UNI TYP framework of the dimension of PARTICIPATION (for a first overall sketch of PARTICIPATION see Seiler 1984). According to this framework the noun/verb-distinction (henceforth N/V-D) must be regarded as a gradable, continuous phenomenon ranging from the stage of a clear-cut distinction with no overlap to almost a non-distinction. Although there is no question that most, if not all, languages do differentiate between nouns and verbs, it is also quite apparent that the languages do so to a different degree and by different means, and that it only makes sense to use the terms "noun" and "verb" in different languages when one actually has a common functional denominator in mind (see below). After a general introduction to the notion of a noun/verb-continuum (chapter 1) the reader will be presented with a survey of languages as diverse as German. English, Russian, Hebrew, Turkish, Salish. and Tongan (see chapter 2) in support of the continuum hypothesis. In chapter 3 the facts are coordinated in an overall pattern of regularities underlying the Increase or decrease of categorical restrictions between the respective word classes. Also, chapter 3 raises the issue to what degree a N/V-D can be considered a matter of certain lexemes or a matter of the morphosyntactic environment of certain lexical units. Lastly, we shall seek for an answer to the question why it is not a necessary requirement for languages to draw a sharp distinction between a word class of nouns and a word class of verbs.
65
The aim of this contribution is to embed the question of an antinomy between "integral" vs. "partial typology", inscribed as the topic of this plenary session, into the comprehensive framework of the dimensional model of the research group on language universals and typology (UNITYP). In this introductory section I shall evoke some cardinal points in the theory of linguistic typology, as viewed "from outside", viz. on the basis of striking parallelisms with psychological typology. Section 2 will permit a brief look on the dimensional model of UNITYP. In section 3 I shall present an illustration of a typological treatment on the basis of one particular dimension. In section 4 I shall draw some conclusions with special reference to the "integral vs. partial" antinomy.
62
As a traditional notion of fundamental importance in linguistics and philosophy (logic), "predication" is fraught with controversial issues. It is thus difficult to delimit the scope of this paper without becoming involved in some major issue. The following distinctions seem to me to be plausible on an intuitive basis. Evidence for why they are useful and legitimate will be found in the body of the paper. The discussion will focus on morphosyntactic predication […].
61
Ergativity in Samoan
(1985)
Most typological and language specific studies on so- called ergative languages are concerned with case marking patterns, particularly split ergativity, with the organization of syntactic relations as defined by syntactic operations such as coreferential deletion across coordinate conjunctions, Equi-NP-deletion and relativization , and with the notion of subject, but usually neglect the notion of valency, though the inherent relational properties of the verb , i. e. valency, play a fundamental role in the syntactic organization of sentences in ergative as well as in other languages . The following investigation of ergativity in Samoan aims to integrate the notion of valency into the description of semantic and syntactic relations and to outline the characteristic features of Samoan verbal clauses as far as they seem to be relevant to recent and still ongoing discussions on linguistic typology and syntactic theory. The main points of the definition of valency […] are: Valency is the property of the verb which determines the obligatory and optional number of its participants, their morphosyntactic form, their semantic class membership (e.g. ± animate, ± human) , and their semantic role (e.g. agent , patient , recipient). All semantic properties and morphosyntactic properties of participants not inherently given by the verb and therefore not predictable from the verb, are not a matter of valency. Valency is not a homogenous property of the verb, but consists of several exponents which show varying degress of relevance in different languages or different verb classes within a single language.
58
Grammatical relations, particularly the notions of transitivity, case marking, ergativity, passive and antipassive have been a favourite subject of typological research during the last decade, but surprisingly, the notion of valency has been of marginal interest in cross-linguistic studies, though the syntactic and semantic status of participants is, to a great extent, determined by the relational properties of the verb. Valency is the property of the verb which determines the obligatory and optional number of its participants, their morphosyntactic form, their semantic class membership (e.g. ± animate, ± human) ,and their semantic role (e.g. agent, patient, recipient). The valency inherently gives information on the nature of the semantic and syntactic relations that hold between the verb and its participants. If a verb is combined with more participants than allowed or less than required, or if the participants do not show the required morphosyntactic form or class membership, the clause is ungrammatical. In other words, it is not sufficient to consider only the number of actants as a matter of valency, but it is only acceptable if all semantic and morphosyntactic properties of the relation between a verb and its participants that are predictable from the verb are included. The predictability of these properties results from their inherent givenness, and it does not seem reasonable to count some inherently given relational properties as a matter of valency, but not others (compare Helbig (1971:38f) and Heidolph et ale (1981:479) who distinguish between the quantitative, syntactic and semantic aspect of valency).
56
The present paper is an attempt to describe a particular semantic domain in Thai, that of local relations, in terms of a gradual interconnection of what traditional descriptions usually regard as distinct and isolated categories. It is based on the well-known observation that isolating languages like Thai typically display a high degree of 'multifunctionality', or else of syntactic 'versatility' of very many lexical items. […] The semantic area studied in the following pages yields a clear systematic interconnection of three different categories, viz. that of nouns – as the focal instance of maximum syntactic independence –, that of verbs – as, conversely, the focal instance of maximally relational concepts –, and, as an intermediary category between these two, that of prepositions which the system lexically feeds from both these opposite ends. The examples given in the course of this paper have been obtained from published grammatical literature, from Thai texts, and from informants.
50
According to the present state of research, there seems to be no language which shows possessive classifiers and possessive verbs corresponding to English "to have" at the same time. In classifier languages predicative possession is expressed by verbless clauses, i.e. by existential clauses ("there is my possessed item"), equative clauses ("the possessed item is mine" "that is my possessed item") or by locative expressions ("the possessed item is near me"), in which the classifier in the case of non-inherent possession marks the nature of the relationship. While most Melanesian languages, as for instance Fijian, Lenakel, Pala and Tolai are classifier languages, Nguna, a Melanesian language spoken in Vanuatu, only shows traces of the Melanesian possessive classifier system, but, in contrast to the other Melanesian languages, it has a possessive verb, namely 'peani' "to have". In order to show how the Nguna possessive constructions deviate from the common Melanesian type, we shall start with a brief description of the Melanesian possessive constructions in general, and that of Fijian in particular.
47
The approach outlined in the present paper is based on observations made with African languages. Although the 1000-odd African languages display a remarkable extent of structural variation, there are certain structures that do not seem to occur in Africa. Thus, to our knowledge, an African language having anything that could be called an ergative case or a numeral classifier system has not been discovered so far. It may turn out that our approach can, in a modified form, be made applicable to languages outside Africa. This , however, is a possibility that has not been considered here. The present approach is based essentially on diachronic findings in that it uses observations on language evolution in order to account for structural differences between languages. Thus, it has double potential: apart from describing and explaining typological diversity it can also be material to reconstructing language history.
46
At the end of last year, I designed an inquiry about the present state of linguistic typology in the form of a questionnaire. It was an attempt to cover the whole field by formulating the questions which seemed most relevant to it. This questionnaire is reproduced, without modifications, following this preface. In the first days of this year, it was sent to 33 linguists who I know are working in the field. The purpose was to form, on the basis of responses received, a picture of convergences and divergences among trends of present-day linguistic typology. The idea was also to get an objective basis for my report on "The present state of linguistic typology", to be delivered at the XIII. International Congress of Linguistics at Tokyo, 1982.
44
Possessive constructions are grammatical constructions which contain two nominals and express that the referent of one of these nominals belongs to the other. The kind of relationship denoted by possessive constructions is not only that of ownership (1), as the term "possessive" might suggest, but also that of kinship (2), bodypart relationship (3), part/whole relationship (4) and similar relationships [...]. The following investigation will start with possessive constructions on phrase level, i.e. possessive phrases, and then deal with possessive constructions on clause level.
43
The basic question is whether POSSESSOR and POSSESSUM are on the same level as the roles of VALENCE, two additional roles as it were. My research on POSSESSION has shown (Seiler 1981:7 ff.) that this is not the case, that there is a difference in principle between POSSESSION and VALENCE. However, there are multiple interactions between the two domains, and these interactions shall constitute the object of the following inquiry. It is hoped that this will contribute to a better understanding both of POSSESSION and of VALENCE.
42
In this study I want to show, above all, that the linguistic expression of POSSESSION is not a given but represents a problem to be solved by the human mind. We must recognize from the outset that linguistic POSSESSION presupposes conceptual or notional POSSESSION, and I shall say more about the latter in Chapter 3. Certain varieties of linguistic structures in the particular languages are united by the fact that they serve the common purpose of expressing notional POS SESSION. But this cannot be their sole common denominator. How would we otherwise be able to recognize, to understand, to learn and to translate a particular linguistic structure as representing POSSESSION? There must be a properly linguistic common denominator, an invariant, that makes this possible. The invariant must be present both within a particular language and in cross-language comparison. What is the nature of such an invariant? As I intend to show, it consists in operational programs and functional principles corresponding to the purpose of expressing notional POSSESSION. The structures of possessivity which we find in the languages of the world represent the traces of these operations, and from the traces it becomes possible to reconstruct stepwise the operations and functions.
39
In my Cahuilla Grammar (Seiler 1977:276-282) and in a subsequent paper (Seiler 1980:229-236) I have drawn attention to the fact that many kin terms in this language, especially those that have a corresponding reciprocal term in the ascending direction – like niece or nephew in relation to aunt – occur in two expressions of quite different morphological shape. The following remarks are intended to furnish an explanation of this apparent duplicity.
30
One of the striking features in modern Newari noun phrases is the wide usage of a set of affixes found in combination with the various elements that may expand a noun into an endocentric construction. At first sight such affixation would appear as a linking device by which the subordinate constituents of a noun phrase are tied to their head noun. Closer investigation, however, reveals a more complex picture which I have attempted to outline in the following paragraphs. The results of this inspection lead to the conclusion that the pattern of affixation displayed in Newari mirrors the close interaction of two converse functional principles: both the syntagmatic function of nominal determination on the one hand and a paradigmatic function – the formation of certain types of lexicalized expressions in Newari – formally tie in with each other by the application of one common technique.
27
The language of the Cahuillas shows two systems of expressions referring to kinship, which could be termed, respectively, as labeling-relational and as descriptive-establishing. […] Descriptive terms show two properties: 1. They are analysable into constituent elements so as to recognize the connection between the term and the proposition. 2. They are distinguishable from the proposition: a. by a special formal element […], in Cahuilla the absolutive suffix. b. by a narrowing or specialization in the meaning. A term which is not descriptive, i.e. which is not connected with a proposition, I shall call "label", "1abeling": It does not say anything about the object but is assigned to it just as a label is attached to a thing […].
24
Using Ultan's theory of descriptivity grading as a starting point, I will attempt to capture this differential utility in terms of [...] criteria of literalness, explicitness and syntactic complexity. I will first briefly present his System and investigate some generalizations which he has proposed on the basis of his study of body part terminologies in numerous languages. I will apply his theory to nouns in this and four other semantic domains, in three North American Indian languages. I will test his generalizations and propose some new ones. I will then present an alternative system of descriptivity grading and compare the results of its application with those of Ultan's system. In the final section I will suggest another methodology for quantification. An appendix at the end of the paper lists all of the descriptive lexical items mentioned, graded according to both systems.
23
The basic idea I want to develop and to substantiate in this paper consists in replacing – where necessary – the traditional concept of linguistic category or linguistic relation understood as 'things', as reified hypostases, by the more dynamic concept of dimension. A dimension of language structure is not coterminous with one single category or relation but, instead, accommodates several of them. It corresponds to certain well circumscribed purposive functions of linguistic activity as well as to certain definite principles and techniques for satisfying these functions. The true universals of language are represented by these dimensions, principles, and techniques which constitute the true basis for non-historical inter-language comparison. The categories and relations used in grammar are condensations – hypostases as it were – of such dimensions, principles, and techniques. Elsewhere I have outlined the theory which I want to test here in a case study.
21
In an earlier paper, I proposed a system for evaluating the relative descriptivity of lexical items in a consistent manner in terms of the interrelations of three metrics. The first of these, including five possible degrees of descriptivity, is based on the premise that the sum of the meaningful parts of a given form is or is not equal to the meaning of the whole. The second, also composed of five degrees, is based on paraphrase-term relations in which the logical quantifiers: all, some and no, are applied to the terms of the paraphrase in one test and to the meaningful parts of the term (linguistic form) in the reversibility test. Both tests are applied in the form of logical propositions. The third metric, with three degrees, deals with the relative explicitness of the meaningful parts of a given form: explicit, implicit or neither. […] This system was then tested in a pilot study involving the fairly limited and semantically homogeneous lexical domain of body-part terms in a specific language, Finnish. The purpose of the present paper is to subject comparable data from other languages to the same kind of analysis and compare the results in order to ascertain whether the generalizations arrived at with the Finnish data also hold for the other languages or, more specifically, which of these generalizations are more or less universal and which language or language-type specific? The additional languages to be examined here are: French, German, Ewe, Maasai and Swahili.
20
These notes grew out of my preoccupation with writing a grammar of a particular language, Cahuilla, which is spoken in Southern California and belongs to the Uto-Aztecan family. [...] The Introduction to the Grammar as a whole – of which two sections are reproduced here in a modified version – tries to integrate the synoptic views of the different chapters into a series of comprehensive statements. The statements cluster around two topics: 1. A presentation of Cahuilla as a type of language. 2. Remarks on writing a grammar.
18
Actually, the title should include intralinguistic variation along with the interlinguistic one. For variation within one and the same language is the thing which directly presents itself to the observation while it still remains to be demonstrated that phenomena in different languages can be regarded as variants to be assigned to one and the same invariant principle. There are two senses in which the terms of variant, variation are used in the following remarks: one, which has just been mentioned, concerns the assignment of variants to some definite invariant. The other implies the possibility of gradient transitions and opposes the notions of discreteness and of yes-or-no. I shall not try here to reconcile these two senses and I trust that what I intend to show will become intelligible nevertheless. Henri Delacroix (1924:126f) has reformulated an old hypothesis which seems worth exploring in connection with the search for language universals: "Une langue est une variation historique sur le grand thème humain du langage." It remains to be seen what "le grand thème" or rather "les grands thèmes" are about and what particular language-specific properties could be shown to be variants of one and the same theme. One such major theme which we shall now investigate is the interrelation between, on one side, a word or a sequence of words, and, on the other, a sentence. As this for us is not only a syntactic but also a semantic problem, we might rephrase the anti thesis as that between a term or sequence of terms and a proposition. Two alternative views on the nature of this interrelation seem conceivable: A. The interrelation is yes-or-no, i. e. an element or a string of elements either constitutes a term (sequence of terms) or a proposition. B. The interrelation is of gradient nature, i. e. we find intermediary stages. Both alternatives are appropriate, but under different circumstances.
16
Three quantificational approaches to the measurement of lexical descriptivity are proposed, based on: the semantic sum of the parts of a lexeme is equal to the whole, paraphrase-term and term-paraphrase congruence, explicitness of semantic elements of a construction. Combination of all possible values into tripartite sets and then into equipollent groups results in a system composed of 12 grades. This system was tested with a semantic domain of the Finnish lexicon: body-part terms. The descriptivity indices for each lexical item were correlated with natural divisions of the body, construction-motivation types (form, function, location), grammatical construction types (endo- and exocentric compounds, derived forms, metaphors), and loanwords. These comparisons result in a number of grade profiles whereby specific descriptivity grades are characteristically associated with one or more types of body section, construction motivation, and grammatical construction. Diachronic and synchronic evidence points overwhelmingly to a process of semantic narrowing in the development of descriptive words and labels from phrases or sentences.