认知语言学60页 Cognitive Linguistic1

cognitive linguistics的定义
认知语言学（Cognitive Linguistics）是一种语言学理论，旨在研究语言的认知特征和语义的认知基础。

它探讨语言的认知结构、概念和思维对语言表达的影响。

认知语言学认为语言习得和使用都基于人类认知能力，语言是人类思维的产物，语言通过表达和传递概念和经验来构建和组织人类认知世界。

认知语言学的研究范围包括词汇语义、句法和语用等方面，致力于揭示语言和认知的密切关系，并通过研究语言的认知机制来深入理解人类思维和语言使用的本质。

Lecture One Prototypes, Categories and Categorization Importance of categorization (Lakoff 1987:5-11):“There is nothing more basic than categorization to our thought, perception, action, and speech.” Every time we see something as a kind of thing, we are categorizing. Whenever we reason about kinds of things, we are employing categories. Any time we either produce or understand any utterance of any reasonable length, we are employing categories: categories of speech sounds, of words, of phrases and clauses, as well as conceptual categories. The above discussion involves two major types of categorization, the categorization of natural things and that of language (Taylor 2021) --- the latter can be referred to as metacategorization. “Without the ability to categorize, we could not function at all, either in the physical world or in our social and intellectual lives. An understanding of how we categorize is central to any understanding of how we think and how we function, and therefore central to an understanding of what make us human.”From the time of Aristotle to the later work of Wittengenstein, categories were assumed to be abstract containers, with things either inside or outside the category. Things were assumed to be in the same category if and only if they had certain properties in common. And the properties they had in common were taken as defining the category. This classical theory became part of the background assumptions taken for granted in most scholarly disciplines over the centuries.Within cognitive psychology, categorization has become a major field of study, thanks to the pioneering work of the psychologist Eleanor Rosch, who made categorization an issue. Rosch observed that studies demonstrated that categories, in general, have best examples (called “prototypes”) and that all of the specifically human capacities do play a role in categorization. Prototype theory is changing our idea of the most fundamental of human capacities --- the capacity to categorize --- and our idea of what the human mind and human reason are like. The approach to prototype theory suggests that human categorization is essentially a matter of both human experience and imagination --- of perception, motor activity, and culture on the one hand, and of metaphor, metonymy, and mental imagery on the other.Theory of prototypeThe basic claim of cognitive linguistics on categories is that cognitive categories, unlike classical categories, are prototype-based categories; or the formation of categories centres around prototypes.There are two ways in which to understand the term ‘prototype’. We can apply the term to the central member, or perhaps to the cluster of central members, of a category. The prototype can be understood as a semantic representation of the conceptual core of a category. On this approach, we could say, not that a particular entity is the prototype, but that it instantiates the prototype. (Taylor, 2021: 59)Below is the naissance and development of the theory of prototype-based categories.1.2.1Berlin and Kay’s finding of focal coloursCategorization of colours is a result of arbitrariness or motivation?According to structuralist linguistics, it is a result of arbitrariness.Something about arbitrariness: T aylor 2021: 5-6.There are a number of implications for the study of colour terms which follow from the structualist approach to word meaning. Among these are the following:a)All colour terms in a system have equal status.b)All referents of a colour term have equal status.c)The only legitimate object of linguistic study is the language system, not individual terms in a system.Cognitive linguists put forward an alternative approach to categorization of colours, ., there are focal colours compared with nonfocal ones.According to the linguist-anthropologists Berlin and Kay in their work of Basic Colour Terms (1969), the characteristics of basic colour terms, derived from an investigation of the colour terms in98 languages are as follows. Basic colour terms area)n ot subsumed under other terms.b)morphologically simple.c)not collocationally restricted.d)of frequent use. (Taylor 2021: 8)Berlin and Kay make two especially interesting claims (ibid.: 9). The first concerns so-called “focal”colours. Two colour samples might well be categorized as the same by speakers of one language, but as different by speakers of anther. If, on the other hand, people are asked to select good examples of the basic colour terms in their language, cross-language (and within-language) variability largely disappears. Secondly, they noted that the 98 languages in their survey appeared to select their basic colour terms from an inventory of only 11 focal colours. Besides, the languages did not select randomly from this inventory, instead, they followed the following implicational hierarchy:black, whiteredyellow, blue, greenbrownpurple, pink, orange, grayGiven the focality of colour categories --- whether this be the consequence of neurological processes of perception, of environmental factors, or of both --- the structuralist account of colour terminology turns out to be grossly inadequate. Two characteristics of colour terms are at variance with the assumptions of structuralism:a)C olour categories have a centre and a periphery.b)Because of the primacy of focal reference, colour terms do not form system, in the Saussurian sense.In brief, colour terminology turns out to be less arbitrary than the structuralists maintained. Colour is “a prime example of the influence of underlying perceptual-cognitive [and perhaps also environmental: J. T.] factors on the formation and reference of linguistic categories”.What follows are Rosch’s findings:(1)Focal colours are perceptually more salient than non-focal colours.(2)Focal colours are more accurately remembered in short-term memory and moreeasily retained in long-term memory.(3)The names of focal colours are more rapidly produced in colour-naming tasks andareacquired earlier by children.All in all, focal colours appear to possess a particular perceptual-cognitive salience, which is probably independent of language. According to Kay and McDaniel (1978), the salience of focal colours reflects certain physiological aspects of man’s perceptive mechanisms. Another line of explanation relates focal colours to certain universally occurring phenomena like day and night (white and black), the sun (red), vegetation (green), the sky (blue) and the ground (brown) (Wierzbicka 1990).Rosch replaced Berlin and Kay’s term “focus” with prototype.1.2.2 Prototypical shapes, organisms and objectsThe “good forms” (a notion as proposed by gestalt psychology), . squares, circles and equilateral triangles, are assumed to be perceptually salient among geometrical shapes. See Figure . Natural prototypes have a crucial function in the various stages involved in the formation and learning of categories.The goodness (or typicality) of category members:Not only are there good (“proto typical”) and bad (“non-proto typical”) examples of reds and squares, but there are also good and bad examples of organisms and objects such as dogs, cars and houses. See Figure .Category membership involves different degrees of typicality. Rosch’s main concern was to prove that categories are formed around prototypes, which function as cognitive reference points.Two different types of boundaries and transition zones: (1) “vague entity” or “vagueness”: concrete entities without clear-cut boundaries in reality --- this is the case with knee and other body parts; with fog, snow and similar weather phenomena; and with landscape forms like valley or mountain; (2) “fuzzy category boundaries” or fuzziness: cognitive categories with vague boundaries.There are cases where vagueness and fuzziness coincide. This is true of the second type of entities. Entities like mountains are vague because they are not clearly delimited as individual entities; the cognitive category MOUNTAIN (or KNEE or FOG) is fuzzy because it odes not have clear boundaries either.All types of concrete entities and natural phenomena like colours are conceptually organized in terms of prototype categories, whose boundaries do not seem to be clear-cut, but fuzzy.W. Labov’s series of experiments in volving cups and cup-like containers (Labov 1973, 1978) on the fuzzy nature of category boundaries: See Figures , and .The nature of cognitive categories based on Labov’s findings: four points on p. 19.Cognitive categories are stored in our mind as mental concepts and signaled by the words of a language, but there is no one-to-one relation between categories (or concepts) and words. In fact it is quite normal that one word denotes several categories, or in conventional terminology, that words are polysemous.1.2.3 Prototypical grammatical categoriesPrototype effect not only occurs in the field of nonlinguistic conceptual structures, but also in linguistic structures. The theory of prototype-based category also applies itself to the analysis of grammatical categories. For example:Noun: book, man, tomato, rice, body, water, spirit, state, have a try, take a look, listening and speaking, good-for-nothing, forget-me-not, etc.Verb: run, eat, go, don’t father me, etc.Word: mother vs. mother-in-law; husband vs. ex-husband; the wait-and-see policy; There is a sort of oh-what-a-wicked-world-this-is-and-how-I-wish-I-could-do-something-to-make-it-better-and- nobler expression about Montmorency that has been known to bring the tears into the eyes of pious ladies and gentlemen (Wang Yin, 2021: 348-349); etc.There exists a prototypical scale as is illustrated by the figure below:prototype peripheryExercises, see .1.2.4 Principle of family resemblanceL. Wittgenstein’s discussion of the category GAME (p. 25): nothing is common to different games, but there are similarities and relationships among them. His conclusion was that games are connected by a network of overlapping similarities, which he called family resemblances. On a somewhat more abstract level, the principle of family resemblances has been defined as:A set of items of the form AB, BC, CD, DE. That is, each item has at least one, and probablyseveral, elements in common with one or more other items, but no, or few, are common to all items. (Rosch & Mervis, 1975: 575)The principle of family resemblances opens up an alternative to the classical view that attributes must be common to all category members, that they must be “category-wide”. This could be most impressively demonstrated with the analysis of “superordinate categories” like GAME, FURNITURE, VEHICLE, FRUIT, etc. Superordinate categories largely depend on family resemblances.The explanatory potential of the underlying principle of family resemblances: family resemblances can explain why attributes contribute to the internal structure of the category even if they are not common to all category members, . if they are not essential features according to the classical view. See Figure .1.2.5 Internal category structure and gestaltExamples of man’s judgment of things:Man’s judgment of most organisms and object in contrast with that of a zebra and an elephant: ;A series of interviews conducted in the context of the categorization of houses (Schmid, 1993: 121ff): p. 32The series of interview support the assumption that categorization and goodness ratings involve two stages: the perception of an object as a whole as the first step (the so-called holistic perception), and a kind of decomposition of the perceived whole into individual properties or attributes as a second (optional) step.The idea of “perceived whole” comes close to the notion of gestalt as advocated by gestalt psychologists, whose major claim is that gestalt perception can be traced back to “gestalt laws of perceptual organization”, or “gestalt principles”, of which the most important are:• “principle of proximity”: p. 33. (Example, see Zhao YF:• “principle of similarity”: ibid. (Example, see Zhao YF:• “principle of closure”: ibid. (Part of an object is shielded from being seen)• “principle of continuation”: ibid. (Interrupted circle, triangle, line, etc.)The roles played by shape, parts, and function of objects in the gestalt structure:Shape is just one aspect of the gestalt of organisms and objects. A central role in providing an object with a gestalt is played by its constituent parts. Consider the cottages (a) and (b) in Figure . It is the parts of an object that play an important role in establishing a prototypical gestalt. Parts do not only contribute to the overall shape of an object, but are also related to its function, which for most artifacts is the raison d’être of the object. The existence of most parts of an object is motivated by the particular purpose the object serves. Consider the examples of chair (p. 35), cottage (p. 36) and teddy bear (p. 37).Conclusion: even in the domain of concrete objects we are not just confronted with a single kind of gestalt prototype but have to take into account that certain lexical categories require richer visual representations, which in turn suggest richer underlying gestalt prototypes.The discussion of gestalt has shown mainly two things:(1)Gestalt perception seems indeed to play an important part in categorization andgoodness ratings.(2)The role of gestalt in the categorization of objects and organisms need not becompletely left to intuition, it can be studied by making selective use of the principles of gestalt psychology and by considering additional aspects like parts and function.As to the cognitive status of categories, prototypes, attributes and gestalt, see .Exercises, .1.2.6 Four characteristics of prototypicalityThe following four characteristics are typical of prototypicality:1) Prototypical categories exhibit degrees of typicality; not every member is equally representative for a category.2) Prototypical categories exhibit a family resemblance structure, or more generally, their semantic structure takes the form of a radial set of clustered and overlapping readings.3) Prototypical categories are blurred at the edges.4) Prototypical categories cannot be defined by means of a single set of criterial (necessary and sufficient) attributes. (Li Fuyin, 1999: 230)Context-dependence and cultural models1.3.1 Context-dependence of prototypesFour examples: p. 43.What turns out the most likely member of a certain category depends on the context. Depending on the context, the prototype shifts. The ‘context-dependent’ categories which are thus evoked may be completely different from the non-contextualized prototypes elicited in goodness-of-example experiments.The prototypes in these examples ., the prototypical members) are context-dependent prototypes, and the category they belong to is a context-dependent category.The context-dependent category structure is different from the structure that was obtained in non-contextualized goodness-of-example ratings. The context not only determines the choice of the category prototype, but it also leads to an adjustment of the position of other category members.The relation between context and attributes: When it is viewed in terms of attributes, context seems to have a twofold effect: First, the context can change the weight of attributes that seem to be relevant for a certain category; second, the context can emphasize attributes that are not prominent and even introduce new attributes which would not be mentioned at all in non-contextualized attribute-listing experiments. In certain contexts, previously peripheral examples are equipped with large bundles of heavily weighted attributes and turned into good examples or even prototypes, while well-established good examples are reduced to the status of marginal members.1.3.2 Context, situation and cognitive modelsThe context can completely reshuffle the positions of members within the category structure. But what is the context?The four traditional definitions of the term ‘context’(pp. 45-46): 1) linguistic context: the linguistic material preceding and following a word or sentence; 2) the set of background assumptions necessary for an utterance to be intelligible (according to language philosophers and pragmalinguists); 3) the situation where an utterance is embedded (in discourse-oriented approaches to language); 4) Malinowsky: situational context and cultural context in addition to linguistic context.The cognitive approach to ‘context’: the notion should be considered a mental phenomenon.Langacker defines his central notion of domain ‘as a c ontext for the characterization of semantic unit’ (1987a: 147).The distinction between ‘context’ and ‘situation’: The former belongs to the field of mental phenomena, while the latter refers to some state of affairs in the ‘real world’.Their definitions: The term situation is defined as the interaction between objects in the real world, whereas the term context as the cognitive representation of the interaction between cognitive concepts (or categories). See Figure .The cognitive representation, or context, does not remain an isolated mental experience, but is associated in at least two ways with related knowledge stored in long-term memory. On the one hand, context-specific knowledge about the categories involved is retrieved (p. 47). On the other hand, the currently active context calls up other contexts from long-term memory that are somehow related to it (ibid.).The sum of the experienced and stored contexts (or cognitive representations) for a certain field by an individual is referred to as cognitive model. Two important properties of cognitive models: First, cognitive models are basically open-ended; second, cognitive models are not isolated cognitive entities, but interrelated (p. 48). In line with recent theories of the mind which emphasize its so-called ‘connectionist’ architecture, one can argue that cognitive models combine to build networks.Three aspects of cognitive models: their incompleteness, their tendency to build networks, and the fact that they are omnipresent.In every act of categorization we are more or less consciously referring to one or several cognitive models that we have stored. We can neither avoid the influence of cognitive models nor function without them.1.3.3 Cultural modelsIn view of the fact that cognitive models represent a cognitive, basically psychological, view of the stored knowledge about a certain field and that descriptions of cognitive models are based on the assumption that many people have roughly the same basic knowledge about things, cognitive models are not universal, but depend on the culture in which a person grows up and lives.The culture provides the background for all the situations that we have to experience in order tobe able to form a cognitive model. So, cognitive models for particular domains ultimately depend on so-called cultural models, which in reverse can be seen as cognitive models that are shared by people belonging to a social group or subgroup. (The definition of ‘cultural model’: a view of cognitive models highlighting the fact that they are intersubjectively shared by the members of a society or social group.)The relation between cognitive models and cultural models: They are just two sides of the same coin. While the term ‘cognitive models’ (related to cognitive linguistics and psycholinguistics) stresses the psychological nature of these cognitive entities and allows for inter-individual differences, the term ‘cultural models’ (belonging to sociolinguistics and anthropological linguistics) emphasizes the uniting aspect of its being collectively shared by many people.The effects of diverging cultural models on the structure of a category: Different cultural models are apt to lead to different prototypical examples of the same category or concept. See Figure .Other examples: the English and French prototypes of the category FIRST MEAL OF THE DAY (see Figure ; the British and American prototypes of the category FIRST FLOOR.Cultural models are not static but changing.Levels of categorization1.4.1 Basic level categoriesScientific classifications may be fascinating in their complexity and rigidity, but they are not necessarily suitable for human categorization. According to the so-called folk taxonomies, there are principally three levels of categorization: basic level categories (middle/intermediate level categories), superordinate categories (higher level categories), and subordinate categories (lower level categories). The folk taxonomies suggest that we approach hierarchies from the centre, that we concentrate on basic level categories such as dogs and cars and that our hierarchies are anchored in these basic level categories.The principle underlying hierarchical structure is the notion of class inclusion, . the view that the superordinate class includes all items on the subordinate level:animalmammals birds reptilesdogs cats cows lions elephants mice etc.terriers bulldogs alsatians poodles etc.Similar hierarchies exist for man-made objects like vehicles: .It seems that the whole range of concrete entities in the world can be hierarchically ordered according to the principle of class inclusion. The detailed classifications (or taxonomies) which have been developed in many scientific fields may simply appear to be an extension of the basic human faculty of categorization.Scientific classificationsThe prime example of scientific taxonomies is the classification of plants and animals based on the proposals made by the Swedish botanist Linnaeus.More general in scope and closer to our linguistic concerns is the classification proposed by M. Roget in his Thesaurus of English Words and Phrases.What such classifications have in common is that they do not seem to be tailored to the human mind in at least two respects. Firstly, scientific taxonomies consist of too many levels. Secondly, they do not consider the fact that we are in constant contact and involvement with the organisms and objects of the world around us. As a consequence of this interaction, organisms and objects are evaluated and assume different degrees of importance in our eyes. (Hence, such terms as “basic level” relative to “non-basic lev el” --- subordinate and superordinate levels.) In contrast with these subjective judgments, scientific classifications aim to be as objective as possible; they do not favour any one of the items classified or a certain taxonomic level.Category names on a middle level like dog, cat, car or truck are the ones that are first learned by children; they tend to be the shortest names in hierarchies and they are used most frequently (Brown, 1958; 1965). Scientific classification or taxonomies should always be objective but everyday categorization or folk taxonomies are not necessarily objective.1.4.2 Superordinate categoriesHow do other types of cognitive categories differ from basic level categories and how does the status of these categories affect our notion of hierarchy? This section focuses on the discussion of superordinate categories, with subordinate categories to be dealt with in the next section.The structure of superordinate categories and the notion of parasitic categorizationTo start with the most obvious deficiency of superordinate categories, there is no common overall shape and, consequently, no common underlying gestalt that applies to all category members. However, this does not mean that we cannot approach the objects categorized as FRUIT or VEHICLE holistically. If you were asked to provide a picture of these categories, you would….To provide a picture of these categories, one would ‘borrow’ the gestalt properties of the superordinate category from the basic level categories involved --- a first case of what will be called parasitic categorization.See Figure : the bulk of the attributes of the superordinate category FURNITURE are borrowed from the rich attribute inventories of the basic level categories CHAIR and TABLE --- a clear case of parasitic categorization. The fact that we draw the attributes for the superordinate category FURNITURE from basic level categories seems to be mainly due to the family resemblances which can be observed between category members.The non-basic status of superordinate categories is also reflected linguistically: 1) Many words for superordinate categories such as FURNITURE, VEHICLE, MUSICAL INSTRUMENTS do not belong to the simple one-syllable type; 2) superordinate terms do not normally come to mind fist, and they are learnt by children only after basic level words have been acquired.Salient attributes and the functions of superordinate categoriesOne of the main reasons for creating superordinate lexical categories is the highlighting of salient general, and mostly functional, attributes. See Figure .Inextricably linked with the highlighting effect and equally important is a second function, the collecting function of superordinate categories: the faculty to assemble a large number of categories under one label which makes the whole set of categories available for easy handling. See Figure .Experiential hierarchies and logical hierarchiesSuperordinate categories can be characterized in the following way:(1)They are subsidiary because they rely on basic level categories for their gestalt and forthe majority of their attributes; this can be called parasitic categorization.(2)Superordinate categories do have one or several category-wide attributes, which are infact their raison d’être.(3)Inextricably bound up with the highlighting effect is the collecting function ofsuperordinates. This means that… (p. 84).(4)The experiential hierarchies emerging from everyday categorization . the folktaxonomies) are neither complete nor fully consistent.ExercisesDo items 2, 4 and 7.1.4.3 Subordinate categoriesWe use subordinate terms like poodle or terrier and not basic level terms like dog when we want to be more specific. This specificity determines the way in which we categorize on the subordinate level, and it is also responsible for the fact that subordinate categories are often expressed by compounds and other composite terms.The most frequent type of lexical category apart from basic level categories are subordinate categories.Subordinates: characteristics of category structureAs Rosch observed, only some of the attributes are specific to the subordinate category in question, while the majority are identical with the attributes of the respective basic level category, that is, are borrowed from the basic level categories; this can again be regarded as a case of parasitic categorization. Compare Figure .There is a crucial difference which concerns the quality of the attributes available for direct categorization. In the case of superordinates this is a set of salient general attributes that are also shared by all the respective basic level categories; in contrast, the attributes available for direct categorization of subordinate categories are specific, . they specify the category in question; they are not normally shared by other categories.Subordinate categories are not only rendered by “simple” words like daisy and dime, but also (and even more often) by composite forms, which include compound forms such as blackbird, apple juice, or wheelchair, and also syntactic groups like black bean or black hair.。