Languages & Linguistics
Inference
Inference refers to the process of drawing conclusions or making educated guesses based on available evidence or information. In linguistics, inference plays a crucial role in understanding the meaning of language, as it involves using contextual and linguistic cues to fill in gaps and make sense of ambiguous or incomplete information. This process is essential for comprehension and communication in language.
Written by Perlego with AI-assistance
Related key terms
1 of 5
5 Key excerpts on "Inference"
eBook - PDFMeaning as Explanation
Advances in Linguistic Sign Theory
- Ellen Contini-Morava, Barbara S. Goldberg, Ellen Contini-Morava, Barbara S. Goldberg(Authors)
- 2011(Publication Date)
- De Gruyter Mouton(Publisher)
As William Diver often points out, it counts on the participation of human intelligence (Diver 1984: 1, 1990; 1). It is true that language is not identical with thought, and forms of thinking do not coincide with forms of language. What has not been paid sufficient attention is the fact that linguistic systems work in close Logical Inference in interpreting direct message sentences 467 collaboration with thinking. The mental activities of cognition and ratio-nal thinking are infinitely complicated and varied, and they are in a state of constant change. Their tool, language, however, has to be simple enough to be mastered by everyone and immutable enough to be the common property of a community. This difficulty is overcome by devel-oping linguistic systems which, because they have their own form and meaning, are used to hint at the mental activities the speaker wants to communicate, instead of directly giving them a form. As the meanings which are capable of being expressed by linguistic systems are much more limited than the range of messages a speaker may want to communicate, the linguistic expression must necessarily be much more general and rough hewn than the message the speaker wants to get across. This is, however, compensated for by the abstract and abbreviated nature of lin-guistic meaning, which allows participants in communication to work out the intended message by using their inferential faculty on the basis of the linguistic form used and shared world and contextual knowledge. There are two kinds of Inference relevant to language use: linguistic and pragmatic. Linguistic Inferences aim at realizing grammatical, lexical, and sentential meaning. These Inferences are conducted under the guid-ance of linguistic systems, and, therefore, can nearly always guarantee agreement between speaker and hearer. This is what we mean by saying that the language of a community ensures mutual understanding.
eBook - PDF- R. Dietrich, C.F. Graumann(Authors)
- 2014(Publication Date)
- North Holland(Publisher)
Language Processing in Social Context, R. Dietrich and C.F. Graumann (Editors) © Elsevier Science Publishers B.V. (North-Holland), 1989 Inference in Language Understanding: What, When, Why and How Alan Garnham Laboratory of Experimental Psychology University of Sussex, Brighton, United Kingdom 1 Inference in Language Understanding: What, When, Why and How Since the late 1960th it has been a commonplace in cognitive psychology and artificial intelligence (AI) that listeners and readers make many Inferences in their attempts to understand discourse and text. Inferences serve a variety of functions in text comprehension. Among other things, they can be used to iden-tify an unclearly pronounced word, to resolve a lexical ambiguity, to determine the referent of a pronoun, and to compute an intended message from a literal meaning. This paper focuses on one particular function of Inference—linking informations from different parts of a text in order to establish its literal mean-ing. It explores in detail some of the questions about Inference making that must be answered by any viable psychological theory of text comprehension. In particular it addresses the four WH-questions of the title: What Inferences do people make? When in the comprehension process are they made? Why do people make those Inferences? How are they made? The answers to these ques-tions are, of course, interrelated. Although I will address each one separately, I will not be be able to answer any one of them completely until I have answered them all. To some extent, answers to these questions come from common sense and from previous research in experimental psychology and AI. However, although there is some truth in answers from these sources, they are, to a greater or lesser extent, misleading. Common sense is never a very good source of psycholinguis-tic theories—we simply do not have conscious access to most of the processes 153
- Jeffrey Heinz, Colin de la Higuera, Menno van Zaanen(Authors)
- 2022(Publication Date)
- Springer(Publisher)
The development of hardware and software that can make discoveries and learn has changed—and continues to change—our society and our lives. If the systems that underly natural languages can be learned by such machines and programs—if these logics, rules, constraints, and structures can be automatically acquired—then virtually all of the above tasks will be solvable automatically by machines. Grammatical Inference goes to the heart of this enterprise. The “grammar” in “grammatical Inference” refers to any aspect of the logics, rules, constraints, and structures that compose the sys- tems underlying natural language. Grammars are models of these systems of knowledge. “Inference” refers to rational steps made in acquiring knowledge from observations and prior assumptions about those observations. At its core, grammatical Inference is a method of inquiry that tries to understand 2 1. STUDYING LEARNING the computations involved in making Inferences about grammars from observations under a variety of different learning scenarios. The purpose of this book is to introduce computational linguists to the major results of this field and to its way of thinking. While the field of grammatical Inference has much to offer computational linguistics, there is no doubt that computational linguists can make contributions to the field of grammatical Inference as well. The notion of grammar adopted here is broad enough that it can be used for any generative system, including non-linguistic ones in other fields. For example, there can be grammars for DNA or RNA sequences, for the order in which messages should be sent over a computer network, or for the structure of web pages. In this book, however, we will either discuss situations that deal with natural language data or discuss topics that are of general nature (and hence valid for all types of data). 1.1 AN OVERVIEW OF GRAMMATICAL Inference Grammatical Inference takes a cue from formal language theory.
eBook - PDF- Anna Duszak(Author)
- 2011(Publication Date)
- De Gruyter Mouton(Publisher)
From this one might conclude that more inferential activity is required of the reader of scientific texts than of the reader of popular scientific texts. This conclusion can be further supported by research results con-cerning text-connectors. It has been shown that text-connectors are marked explicitly in popular science, but not as often in science, because the latter readers are able to infer the connection thanks to their knowl-edge of the subject (Myers 1991: 22; Ventola-Mauranen 1992: 463). 344 Merja Koskela In the present paper I intend to elaborate further the idea of Inference as a textual relation in scientific and popular scientific texts. First I will discuss how the concept of Inference can be defined. Second, I will present some results of my previous research, and based on them I will suggest a preliminary categorization of types of Inference that can be found in Swedish scientific and popular scientific texts. Third, I will discuss whether there is a difference between science and popular science with respect to the different types. 2. Inference as a means of creating coherence in a text The concept of Inference was originally introduced in formal logic, but it is often used in psychology as well. In cognitive psychology and artificial intelligence Inference is defined as a process of filling in the missing con-nections between the surface structure fragments of the text by recourse to context and knowledge about the world. This has been called the text-based view (Collins-Brown-Larkin 1980: 386; cf. Clark 1992: 4 f.). According to another, model-based view, Inference can be seen as a process of synthesizing the surface structure fragments in the text with an underlying model, a frame, that organizes the text as a coherent whole in the mind of the reader. In this view, Inference is controlled by the reader's expectations regarding the contents of the text (Collins-Brown-Larkin 1980: 386; cf. Rothkegel 1991: 193).
eBook - PDF- Rainer Bäuerle, Christoph Schwarze, Arnim von Stechow(Authors)
- 2012(Publication Date)
- De Gruyter(Publisher)
Inferences - The Base of Semantics?* Christopher U. Habel, Berlin 1. Introduction This conference is devoted to Meaning, Use and Interpretation of Language. Therefore, I believe that models of human language understanding are a main topic of this conference. One aim of this paper is to discuss some aspects of language understanding from the viewpoint of Artificial Intelligence (AI) and Cognitive Science (CS), i. e. especially from a viewpoint which can be charac-terized by the method of computer modelling or simulation of human be-havior, particularly of human language production and understanding. 1 The main goal of this paper is to demonstrate that the problems investigated and the solutions proposed by CS and AI are - relevant and of interest to semanticists (no matter whether they are linguists, logicians or philosophers), - related to those treated or proposed by semanticists (of kinds mentioned above). In other words: the contrasting views of language understanding - on the one hand, that of logico-linguistic semantics, on the other hand, that of AI and CS - are closely related, and furthermore, there is a need to integrate these views. 2 I will argue for an integrated view with examples from two problem areas: definite reference and word meaning. I will also give emphasis to two theoretical entities: 'the speaker-listener's knowledge' and 'Inferences'. The latter idea will be shown as the basic notion of my model of language understanding. * This paper is a revised version of the paper Grundzüge einer inferentiellen Semantik I read at the Konstanz colloquium. I received many helpful suggestions from the participants. Parts of the first - German - draft of this paper were prepared during my work with the project Automatic Construction of Semantic Nets at the Technical University of Berlin. Further-more, my research on 'referential nets' has been supported by the DFG. I would particularly like to thank my colleagues Helmar Gust and Claus-Rainer Rollinger.
Index pages curate the most relevant extracts from our library of academic textbooks. They’ve been created using an in-house natural language model (NLM), each adding context and meaning to key research topics.
