of clustered and overlapping readings concentrating around one or more salient readings. Third, prototypical categories are blurred at the edges; there may be entities whose membership of the category is uncertain, or at least less clear-cut than that of the bona fide members. And fourth, prototypical categories cannot be defined by means of a single set of criterial (necessary and sufficient) attributes. Although these four characteristics do not necessarily co-occur, they are sys- tematically related. The first and third are extensional in nature and involve cat- egory membership, whereas the second and fourth represent an intensional per- spective and involve definitions rather than members. Characteristics one and two refer to salience effects and differences of structural weight, whereas three and four focus on flexibility and demarcation problems. (In what follows, we will sometimes use the notion nonequality with regard to features one and two and nondiscreteness with three and four.) There is obviously much more to be said about the status of the four features and their relations (see Geeraerts, Grondelaers, and Bakema 1994), but for present purposes, this brief overview will suffice. Turning to historical se- mantics, we can now convert each of the four characteristics of prototypicality into a statement about the structure of semantic change. Modulations of Core Cases By stressing the extensional nonequality of lexical semantic structure, prototype theory highlights the fact that changes in the referential range of one specific word meaning may take the form of modulations on the core cases within that referential range. In other words, changes in the extension of a single sense of a lexical item are likely to take the form of an expansion of the prototypical center of that extension. If the referents in the range of application of a particular lexical meaning do not have equal status, the more salient members will probably be more stable (dia- chronically speaking) than the less salient ones. Changes will then take the form of modulations on the central cases: if a particular meaning starts off as a name for referents exhibiting the features ABCDE, the subsequent expansion of the category will consist of variations on that type of referent. The further the expansion ex- tends, the fewer features the peripheral cases will have in common with the pro- totypical center. A first layer of extensions, for instance, might consist of referents exhibiting features ABCD, BCDE,orACDE. A further growth of the peripheral area could then involve feature sets ABC, BCD, CDE,orACD (to name just a few). The Development of Radial Sets By stressing the intensional nonequality of lexical semantic structure, prototype theory highlights the clustered-set structure of changes of word meaning. This hypothesis shifts the attention from the extensional structure of an individual 990 stefan grondelaers, dirk speelman, and dirk geeraerts meaning of a lexical category to the intensional structure of the lexical item as a whole, that is, to the overall configuration of the various readings of the word. The hypothesis suggests that the structure of semasiological change mirrors the synchronic semantic structure of lexical categories, given that the latter involves family resemblances, radial sets, and the distinction between central and periph- eral readings. Semasiological change, then, involves the change of prototypically clustered concepts. This general statement can be broken down into two more specific ones. First, the structure of semasiological change as a whole is one of overlapping and interlocking readings; specifically, a novel use may have its starting point in several existing meanings at the same time. Second, there are differences in structural weight among the readings of an item; specifically, there are peripheral meanings that do not survive for very long next to more important meanings that subsist through time. Semantic Polygenesis By stressing the extensional nondiscreteness of lexical semantic structure, prototype theory highlights the phenomenon of incidental, transient changes of word mean- ing. That is to say, the synchronic uncertainties regarding the delimitation of a category have a diachronic counterpart in the form of fluctuations at the bound- aries of the item. In Geeraerts (1997: 62–68), a specifically striking example of such fluctuations is discussed under the heading ‘‘semantic polygenesis.’’ Semantic polygenesis involves the phenomenon that one and the same reading of a particular lexical item may come into existence more than once in the history of a word, each time on an independent basis. Such a situation involves what may be called ex- tremely peripheral instances of a lexical item: readings that are so marginal that they seem to crop up only incidentally and disappear as fast as they have come into existence. Specifically, when the same marginal meaning occurs at several points in time that are separated by a considerable period, we can conclude that the dis- continuous presence of that meaning is not due to accidental gaps in the available textual sources, but that the meaning in question must actually have come into existence independently at the two moments. Semantic Change from Subsets By stressing the intensional nondiscreteness of lexical semantic structure, prototype theory highlights the encyclopedic nature of changes in word meaning. That is to say, diachronic semantics has little use for a strict theoretical distinction between the level of senses and the level of encyclopedic knowledge pertaining to the entities that fall within the referential range of such senses. In semantic change, the ency- clopedic information is potentially just as important as the purely semantic senses lexical variation and change 991 (to the extent, that is, that the distinction is to be maintained at all). This view follows from a prototype-theoretical conception in general, and from the fourth feature mentioned above in particular, in the following way. If the meaning of a lexical item (or a specific meaning within a polysemous item) cannot be defined by means of a single set of necessary features that are jointly sufficient to distin- guish the category from others, the definition necessarily takes the form of a dis- junction of clustered subsets. If, for instance, there is no feature or set of features covering ABCDE in its entirety, the category may be disjunctively defined as the overlapping cluster of, for instance, the sets ABC, BCD, and CDE (and, in fact, others). Similarly (turning from a description based on an extensional perspective to a description undertaken from an intensional perspective), if no single combi- nation of features yields a classical definition of a category, it can only be properly defined as a disjunction of various groupings of the features in question. From a diachronic point of view, this means that semantic changes may take their starting point on the extensional level just as well as on the intensional level, or in the domain of encyclopedic information just as well as in the realm of semantic information. Even where a classical definition is possible, extensional subsets or intensional features with an ‘‘encyclopedic’’ratherthana ‘‘semantic’’statusmayplay a crucial role in processes of semantic change. That is to say, semantic extensions may start from a typical or otherwise salient example of a category, rather than from a ‘‘meaning’’ in the traditional sense. To round off the overview, it should be stressed that the aspects of semantic change enumerated here were not necessarily brought to diachronic semantics by proto- type theory or Cognitive Linguistics alone. What is indubitably new, however, is the fact that these known aspects of change can now be incorporated into a global model of lexical semantic structure. That is to say, from a descriptive point of view the importance of prototype theory probably resides less in the novelty of its ob- servations, taken separately, than in the fact that it brings them together in an overall model of the structure of lexical meaning. Further, of the four prototype-based mechanisms of change, the second has enjoyed most theoretical attention. Detailed examples of all types may be found in Geeraerts (1997). The radial set structure of semantic change is acknowledged and exemplified in Dirven (1985), Lewandowska-Tomaszczyk (1985), Casad (1992), Evans (1992), Goossens (1992), Nerlich and Clarke (1992), Rudzka-Ostyn (1992), Anstatt (1995), Dekeyser (1996), Kronenberg (1996), Maffi (1996), Cuyckens (1999), Soares da Silva (1999), De Mulder and Vanderheyden (2001), Eckardt (2001), Tissari (2001), Koivisto-Alanko (2002), and many other studies. The application of the model to the evolution of grammatical rather than lexical categories is illus- trated, among others, by Winters (1989, 1992a, 1992b), Melis ( 1990), Nikiforidou (1991), Kemmer (1992), Luraghi (1995), Cook (1996), and Aski (2001); see also De Mulder (2001). 992 stefan grondelaers, dirk speelman, and dirk geeraerts 3. From Semasiology to Onomasiology Given that Cognitive Linguistics is strongly involved with categorization as a basic cognitive function, a shift from the semasiological to an onomasiological per- spective is a natural one: from the point of view of the speaker, the basic act of categorization is, after all, the onomasiological choice of a category to express a certain idea. So, what are the contributions of Cognitive Linguistics to onoma- siological research? Before we can answer that question, we first have to chart the field of onomasiological research. Apart from the distinction between synchrony and diachrony, the conceptual map of onomasiology should be based on at least the following four distinctions: the distinction between structural and pragmatic onomasiology, the distinction between the qualitative and the quantitative aspects of lexical structures, the distinction between referential and nonreferential types of meaning, and the distinction between lexicogenetic mechanisms and sociolexi- cological mechanisms. Structural and Pragmatic Onomasiology The two elements that make up Baldinger’s description of onomasiology (see the quotation at the beginning of this chapter) are not equivalent. On the one hand, studying ‘‘a multiplicity of expressions which form a whole’’ (1980: 278) leads di- rectly to the traditional, structuralist conception of onomasiology, that is, to the study of semantically related expressions (as in lexical field theory or the study of the lexicon as a relational network of words interconnected by links of a hyponymical, antonymical, synonymous nature, etc.). On the other hand, studying ‘‘the desig- nations of a particular concept’’ (1980: 278 ) opens the way for a contextualized, pragmatic conception of onomasiology, involving the actual choices made for a par- ticular name as a designation of a particular concept or a particular referent. This distinction can be further equated with the distinction between an investigation of structure and an investigation of use, or between an investigation of langue and an investigation of parole. Qualitative and Quantitative Aspects The distinction between what may roughly be described as the qualitative versus the quantitative aspects of linguistic semantic structure may be introduced by con- sidering semasiological structures first. Qualitative aspects of semasiological struc- ture involve the following questions: which meanings does a word have, and how are they semantically related? The outcome is an investigation into polysemy, and the relationships of metonymy, metaphor, and the like that hold between the various lexical variation and change 993 readings of an item. Quantitative aspects of lexical structure, on the other hand, involve the question whether all the readings of an item carry the same structural weight. The semasiological outcome of a quantitative approach is an investigation into prototypicality effects of various kinds, as described above. The distinction between qualitative and quantitative aspects of semantic struc- ture transfers easily into the realm of onomasiology. The qualitative question then takes the following form: what kinds of (semantic) relations hold between the lexical items in a lexicon (or a subset of the lexicon)? The outcome is an investi- gation into various kinds of lexical structuring: field relationships, taxonomies, lexical relations such as antonymy, and so on. The quantitative question takes the following onomasiological form: Are some categories cognitively more salient than others; that is, are particular categories more likely to be chosen for designating things out in the world than others? Are certain lexical categories more obvious names than others? This type of ‘‘quantitative’’ research is relatively new. The best- known example to date is Berlin and Kay’s basic-level model (Berlin and Kay 1969; Berlin 1978), which involves the claim that there exists a particular taxonomic level which constitutes a preferred, default level of categorization. The basic level in a taxonomy is the level that is (in a given culture) most naturally chosen as the level where categorization takes place; it has, in a sense, more structural weight than the other levels (see also Schmid, this volume, chapter 5). The relationship between this type of ‘‘quantitative’’ onomasiology and the pragmatic perspective mentioned in the previous distinction probably does not need further clarification. A particular onomasiological structure (like a level in taxonomy) can be identified as a preferred level of categorization only by taking into account the pragmatic perspective, that is, the actual choices language users make from among a set of alternative possibilities. Referential and Nonreferential Types of Meaning The distinction between referential (denotational) and nonreferential (connota- tional) aspects of meaning will be clear enough in itself. It involves the distinction between the descriptive aspects of lexical expressions (the contribution they can make to the propositional content of sentences) and their emotive, stylistic,ordis- cursive value. Although there is a general bias in lexical semantics toward the study of referential rather than nonreferential meanings, this relative lack of attention is to be specifically regretted from an onomasiological perspective, because the ties between nonreferential meaning and onomasiology are perhaps even stronger than those between nonreferential meaning and semasiology. In fact, the very definition of nonreferential meaning involves the concept of onomasiological alternatives. Indeed, we invoke the notion of nonreferential mean- ing precisely when a word’s communicative value differs from that of a referen- tial synonym or when its communicative value cannot be defined in referential terms. The latter case involves the meaning of expressions like Hello! What this 994 stefan grondelaers, dirk speelman, and dirk geeraerts expression does (i.e., to perform the speech act of greeting) cannot be defined in purely referential terms; the expression does not describe a state of affairs or a process, but it performs an action. In the same way, the word yuck does not describe aversion, but expresses it. In cases such as these, we say that hello has a discursive meaning or that yuck has an emotive meaning. The words dead and deceased or departed, on the other hand, do have an identifiable referential value. At the same time, their communicative value is not identical: deceased and departed are less straightforward and slightly more euphemistic than dead—that is to say, although their referential values are equivalent, they differ in nonreferential value. Crucially, the distinction between referential and nonreferential meaning leads to the identification of the sociostylistic value of words. For instance, the intro- duction of the loan word Computer into German initially involves the spread of the concept ‘computer’. What lies behind this simultaneous introduction of a con- ceptual and a lexical innovation is a common expressive need on the part of the language users; that is, the driving force behind the spread of the concept ‘com- puter’ and the word Computer is basically just the growing familiarity of language users with this new piece of equipment. However, when the word Rechner is in- troduced as an alternative term for Computer, the concept ‘computer’ is already in place. Now, in order to get a grip on the factors behind the competition between Computer and Rechner, we have to take into account their nonreferential values as well, that is, the differences they exhibit in terms of their stylistic value, which may determine the preference for one or the other term. These values will neces- sarily have to include the sociolinguistic distribution of Computer and Rechner:ifit turns out that one or the other is preferred because it belongs to a prestigious variety of the language, then this sociolinguistic characterization of the item will go into its nonreferential meaning. Note that sociolinguistics as referred to here is to be taken in the broadest possible sense: whether a word is typical of a learned register, of a rural dialect, of an expert jargon, of a trendy youth culture, or of an upper-class sociolect are all aspects of its sociolinguistic character, and this sociolinguistic character is part and parcel of its nonreferential meaning. This implies, in other words, that the nonreferential value of lexical items involves not just their emotive, stylistic, or discursive value, as mentioned above, but their variational value at large, including all possible kinds of sociolinguistic characteristics. Lexicogenetic and Sociolexicological Mechanisms In light of the foregoing, it is now a relatively straightforward matter to explain the difference between lexicogenesis and sociolexicology. Lexicogenesis involves the mechanisms for introducing new ‘‘word form–word meaning’’ pairs—all the tra- ditional mechanisms, in other words, such as word-formation, word creation (the creation of entirely new roots), borrowing, blending, truncation, ellipsis, folk ety- mology, and others, which introduce new items into the onomasiological inventory of a language. Crucially, the semasiological extension of the range of meanings of an lexical variation and change 995 existing word is itself one of the major mechanisms of onomasiological change— one of the mechanisms, that is, through which a concept gets encoded by a lexical expression. In this sense, the study of onomasiological changes is more compre- hensive than the study of semasiological changes, since the former encompasses the latter. Traditionally, lexicogenetic mechanisms are sometimes discussed as if triggered by language as such. We might say, for instance, that German borrows Computer from English. But the language as such is obviously not an anthropomorphic agent: what happens is that individual language users act in a specific way (say, by using a loan word) and that these individual acts lead to changes at the level of the language as a whole—that is, at the level of the speech community. This phenomenon has revealingly been described by Keller (1990), who suggests that linguistic change may be characterized as an ‘‘invisible hand’’ process—a notion he borrowed from economics. As such, changes spread through a linguistic community as if guided by an invisible force, whereas the actual process involves a multitude of communi- cative acts. 1 The invisible hand metaphor, however, stops short of indicating pre- cisely how the transition from the individual level to the global level occurs. What exactly are the mechanisms that enable the cumulative effects? Logically speaking, two situations may occur: either the changes work in parallel, or they take place serially. The first situation occurs when members of a speech community are confronted with the same communicative, expressive problem and independently choose the same solution. The introduction of Computer as a loan from English into German (and many other languages) may at least to some extent have proceeded in this way. More or less simultaneously, a number of people face the problem of giving a name to the new thing in their native language; independently of each other, they then adopt the original name that comes with the newly introduced object. The second type occurs when the members of a speech community imitate each other. For instance, when one person introduces a loan word, a few others may imitate him or her, and they in turn may be imitated by others, and so on. In the same way, the overall picture of a traffic jam is one in which a great number of cars appear to be halted by an invisible hand, while what actually happens is a cumulative process of individual actions: when the first car brakes to avoid a dog running over the highway, the car behind it has to slow down to avoid an accident, and so on. Studying how onomasiological changes spread through a speech community is typically an aspect of sociolexicology, as it is meant here: in addition to identifying onomasiological mechanisms along traditional lexicogenetic lines, we need to study how these mechanisms are put to work and how they may lead to overall changes in the habits of the language community. In short, classifications of lexicogenetic mechanisms merely identify the space of possible or virtual onomasiological changes; sociolexicology studies the actual realization of the changes. Needless to say, the latter approach coincides with the pragmatic perspective (it concentrates on the actual onomasiological choices made by language users), and it crucially involves all the nonreferential values mentioned above (as factors that may influence these choices). 996 stefan grondelaers, dirk speelman, and dirk geeraerts 4. The Contribution of Cognitive Linguistics to Onomasiology Various approaches in lexical semantics have contributed in different ways to the study of onomasiology. Prestructuralist semantics—apart from coining the term onomasiology itself (Zauner 1903)—has introduced some of the basic terminology for describing lexicogenetic mechanisms. Although basically concerned with se- masiological changes, the major semasiological treatises from Bre ´ al (1897) and Paul (1880) to Stern (1931) and Carnoy (1927) do not restrict themselves to strictly semasiological mechanisms like metaphor and metonymy, but also devote atten- tion to mechanisms of onomasiological change like borrowing or folk etymol- ogy. Characteristically, there is a certain degree of overlap among the overviews given by Kronasser (1952) and Quadri (1952) of semasiological and onomasio- logical research, respectively. Attempts to classify lexicogenetic mechanisms con- tinue to the present day. Different proposals may be found in the work of, among others, Dornseiff (1966), Tournier (1985), Zgusta (1990), and Grzega (2002). It lies beyond the scope of the present chapter to systematically compare these proposals, but it may be noted that there is no single, universally accepted clas- sification. Structuralist semantics makes two important contributions to onomasiology. First, it insists, in the wake of Saussure himself, on the distinction between sema- siology and onomasiology. In the realm of diachronic linguistics, this shows up in Ullmann’s (1951, 1962) classification of semantic changes and in Baldinger’s (1964) argumentation for studying the interplay between semasiological and onomasio- logical changes. More importantly, the bulk of (synchronic) structuralist semantics is devoted to the identification and description of different onomasiological struc- tures in the lexicon, such as lexical fields, taxonomical hierarchies, lexical relations like antonymy and synonymy, and syntagmatic relationships. From the point of view of the classification presented above, structuralist semantics is mainly situated within the field of ‘‘qualitative’’ synchronic onomasiology: it concentrates on ono- masiological structures within the (synchronic) lexicon. Second, structuralist se- mantics has identified one of the possible explanatory factors for onomasiological change, namely, homonymic clashes (Gillie ´ ron and Roques 1912). Gillie ´ ron claims that homonymy is a pathological situation that calls for curative devices, namely, the therapeutic elimination of one of the homonyms. The principle of avoidance of homonymy derives from the idea that there exists an isomorphism between the form and the content of natural languages, a principle that is summarized in the maxim ‘‘one form, one meaning.’’ Although this isomorphic principle is presented as a structural cause for change, the most realistic way of interpreting it is to accept that it ultimately relies on communicative mechanisms: in some communicative situations, homonymy may lead to difficulties of understanding, and such hom- onyms may eventually be avoided by the language users. In this respect, avoidance lexical variation and change 997 of homonymy may be considered a first example of a pragmatic perspective in onomasiology. There are at least four important contributions that Cognitive Semantics has made to onomasiological research: a. Cognitive Semantics has drawn the attention to a number of ‘‘qualitative’’ onomasiological structures that did not come to the fore in the structur- alist tradition. This holds true, on the one hand, for the development of the Fillmorean frame model of semantic analysis (Fillmore 1977, 1985; Fillmore and Atkins 1992; Cienki, this volume, chapter 7). On the other hand, the seminal introduction of conceptual metaphor research in the line of Lakoff and Johnson ( 1980) (see also Grady, this volume, chapter 8) can be seen as the identification of figurative lexical fields: the ensembles of near-synonymous metaphors studied as conceptual metaphors consti- tute fields of related metaphorical expressions (just like ordinary seman- tic fields consist of ensembles of near-synonymous lexical items). b. Cognitive Semantics introduces a ‘‘quantitative’’ perspective into the study of onomasiological structures. As mentioned above, basic-level research in the line of Berlin and Kay introduces the notion of salience (which is well known in Cognitive Semantics through the semasiological research into prototypicality) into the description of taxonomical structures: basic levels are preferred, default levels of categorization. Further research (Geeraerts, Grondelaers, and Bakema 1994) has established that the concept of ono- masiological salience may be further refined: their notion of entrenchment, defined over individual concepts rather than taxonomic levels, is a gen- eralization of the notion of onomasiological salience as represented by the notion of basic level (see Schmid, this volume, chapter 5). c. Cognitive Semantics introduces a ‘‘quantitative’’ perspective into the study of lexicogenetic mechanisms. Within the set of lexicogenetic mecha- nisms, some could be more salient (i.e., might be used more often) than others. Superficially, this could involve, for instance, an overall prefer- ence for borrowing rather than morphological productivity as mechanisms for introducing new words, but from a cognitive semantic perspective, there are other, more subtle questions to ask: do the way in which novel words and expressions are being coined reveal specific (and possibly pre- ferred) ways of conceptualizing the onomasiological targets? An example of this type of research (though not specifically situated within a cognitive semantic framework) is Alinei’s work (e.g., 1996) into the etymological patterns underlying the European dialects: he argues, for instance, that taboo words in the European dialects may be motivated either by Chris- tian or Islamic motifs or by pre-Christian, pre-Islamic heathen motifs; the quantitative perspective then involves the question whether one of these motifs is dominant or not. Within Cognitive Semantics properly speaking, 998 stefan grondelaers, dirk speelman, and dirk geeraerts this type of approach is represented by the search for dominant (or even universal) conceptual metaphors for a given domain of experience. A case in point is the work of Ko ¨ vecses (1990; 2000). On a broader scale, the etymological research project described by Koch and Blank (Koch 1997; Blank and Koch 1999) intends to systematically explore motivational pref- erences in the etymological inventory of the Romance languages. In com- parison with much of the metaphor-based research, the approach put forward by Blank and Koch takes into account all possible pathways of lexicalization (and not just metaphor). d. Cognitive Semantics highlights the crucial role of a usage-based socio- lexicological approach to the study of lexical change. Onomasiological change cannot be understood unless we take into account pragmatic onomasiology: changes are always mediated through the onomasiological choices made at the level of usage. Words die out because speakers refuse to choose them, and words are added to the lexical inventory of a language because some speakers introduce them and others imitate these speak- ers; similarly, words change their value within the language because peo- ple start using them in different circumstances. Lexical change, in other words, is the output of processes that are properly studied in the context of pragmatic onomasiology. To repeat a point made earlier, this pragmatic, usage-based perspective automatically takes the form of a socio- lexicological investigation: in choosing among existing alternatives, the individual language user takes into account their sociolinguistic, nonref- erential value, and conversely, the expansion of a change over a language community is the cumulative effect of individual choices. In this sense, it is only through an investigation into factors determining these individual choices that we can get a grasp on the mechanisms behind the invisible hand of lexical change. Figure 37.1. The central position of usage-based onomasiology lexical variation and change 999 . the extension of a single sense of a lexical item are likely to take the form of an expansion of the prototypical center of that extension. If the referents in the range of application of a particular. namely, the therapeutic elimination of one of the homonyms. The principle of avoidance of homonymy derives from the idea that there exists an isomorphism between the form and the content of natural. geeraerts meaning of a lexical category to the intensional structure of the lexical item as a whole, that is, to the overall configuration of the various readings of the word. The hypothesis suggests that the