This book "asserts that the origin and spread of languages must be examined primarily through the time-tested techniques of linguistic analysis, rather than those of evolutionary biology" and "defends traditional practices in historical linguistics while remaining open to new techniques, including computational methods" and "will appeal to readers interested in world history and world geography."
Date: Fri, 13 Jan 2006 14:42:00 +0100 From: Rolf Kreyer Subject: Regularity in Semantic Change
AUTHORS: Traugott, Elizabeth C.; Dasher, Richard B. TITLE: Regularity in Semantic Change SERIES: Cambridge Studies in Linguistics 97 PUBLISHER: Cambridge University Press YEAR: 2005
Rolf Kreyer, Institut für Anglistik, Amerikanistik und Keltologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Germany
[A review of the first edition of this book, published in 2001, appears in http://linguistlist.org/issues/13/13-1975.html -- Eds.]
The book by Traugott and Dasher aims to show that semantic change can be described with reference to predictable unidirectional paths of change. In particular, the authors argue for an 'Invited Inferencing Theory of Semantic Change' (IITSC), the prime objective of which is ''to account for the conventionalization of pragmatic meanings and their reanalysis as semantic meanings.'' (35) Semantic change, thus, is seen as a product of language use. Accordingly, the authors make extensive use of contextualized data from diachronic corpora of (mainly) English and Japanese spanning over more than one thousand years to provide evidence for their claims.
The book is divided into six main chapters: the first gives an overview of the general framework underlying the findings of the authors. Chapter two discusses relevant previous and current work on semantic change. The following four chapters are detailed case studies relating to different areas of linguistic description, namely modal verbs, adverbials with discourse marker function, performative verbs and constructions, and social deictics. The book concludes with a summary of the major findings and with some directions for future research.
Chapter one describes in detail the aim of the whole book. The authors start off with the observation that ''the direction of semantic change is often highly predictable, not only within a language but also cross-linguistically.'' (4) To account for such regularities, the authors suggest an Invited Inferencing Theory of Semantic Change' (IITSC), which roughly can be sketched as follows: At the basis of semantic change is the individual who makes innovative use of a particular word with a particular coded meaning by exploiting 'invited inferences' ''in associated streams of speech.'' (38) Such a particular utterance token thus is used to convey a meaning by the speaker/writer which differs from the original coded meaning. If this meaning persists in the speech community and is also used by other speakers/writers the invited inference will become a 'generalized invited inference': although the original meaning is still the dominant one, an additional meaning based on the invited inference is associated with the word. Once the original meaning recedes to certain contexts or disappears altogether, the generalized invited inference has become semanticized as a new coded meaning of the word.
Basic to the authors' understanding of semantic change are the notions of 'subjectivity' and 'intersubjectivity.' The former refers to the explicit encoding of the ''SP/W's [speaker/writer's] point of view, for example in deixis, modality, and marking of discourse strategies. [... The latter] is most usefully thought of in parallel with subjectivity: as the explicit, coded expression of SP/W's attention to the image or ''self'' of AD/R [addressee/reader] in a social or epistemic sense, for example, in honorification.'' (21-2)
The main mechanism underlying semantic change, according to the authors, is subjectification, i.e. the development of meaning components in lexemes that increase subjectivity, or, in the authors' words ''the semasiological process whereby the SP/Ws [speakers/writers] come over time to develop meanings for Ls [lexemes] that encode or externalize their perspectives and attitudes as constrained by the communicative world of the speech event, rather than by the so-called ''real-world'' characteristics of the event or situation referred to.'' (30)
This mechanism also accounts for one of four general tendencies in semantic change, namely the development from non-subjective to subjective to intersubjective meanings, i.e. from meanings that merely express a particular state of the extralinguistic world, to meanings that allow to express the speaker or writer's general point of view, to meanings that serve to specifically express the speaker or writer's attitude towards the hearer or reader or their needs.
The remaining three paths of semantic change are from truth- conditional to non-truth conditional meaning, from contentful to procedural meaning, and from meanings that only have scope within the proposition to meanings that have scope over the whole discourse. The figure below summarizes these four pragmatic- semantic tendencies in semantic change (note that this diagram should only be read horizontally, not vertically; s-w: scope within, s-o: scope over) (cf. Traugott & Dasher 2005: 40):
It is important to note, though, that not every lexeme has to undergo these changes; each of the four lines of development are just possible kinds of changes. However, ''if a lexeme with the appropriate semantics undergoes change, it is probable that the change will be of the type specified,'' (281) even though not every lexeme will undergo all possible changes. In addition, the four paths are unidirectional, that is a change in the opposite direction is usually ruled out. How these general tendencies are instantiated in different semantic domains is discussed in chapters three to six of the book. Before that, however, in chapter 2 the authors discuss the scientific context in which their model of semantic change is placed.
Chapter 2 gives a brief outline of previous and current research on semantic change and discusses Bréal, major research of the early twentieth century (Meillet, Bloomfield, Saussure, the concept of semantic fields as discussed, among others, by Trier, Stern, or Berlin and Kay) and contemporary approaches, such as studies on metaphor, metonymy and invited inference, issues of grammicalization and historical pragmatics. With regard to these different approaches to semantic change, the authors consider their book to be ''a contribution especially to the interface between historical pragmatics and historical semantics, building on the various approaches [...] sketched in [...] chapter , especially Neogricean pragmatics.'' (104) Specifically, the authors contend that semantic change is dominated by what Horn (1984) has formulated as the R[elevance]-heuristic, i.e. ''say/write no more than you must, and mean more thereby'': ''[T]he R-heuristic leads to change because it evokes utterance meanings beyond what is said; in other words, it involves ''pragmatic strengthening''.'' (19)
Chapter three presents three case studies on the development of modal auxiliaries and particles, namely English 'must' and 'ought to' and Chinese 'de'. Of the more general paths of semantic change posited in chapter one, the authors find the following instantiations in the semantic change of modal verbs (cf. Traugott & Dasher 2005: 148):
As an example of the first path consider the discussion of the development of 'must'. In OE 'must' or 'mot-' usually had the meaning of ability or permission, as for instance in (1): (1) (c. 880 Boethius 5, 12.12 [Traugott & Dasher 2005: 123]) Mot ic nu cunnian hwon þinre fæstrædnesse, ... ''May I now inquire a little about your fortitude, ...''
In later OE and EarlyME 'mot-' acquired an additional deontic meaning component: (2) (c. 1000 AECHom I, 17 (App) 182.240 [Traugott & Dasher 2005: 124]) we moton eow secgan eowre sawle þearfe, licige eow ne licige eow. ''we must tell you about your soul's need, whether it please you or not.''
The epistemic use of 'must' occurs sporadically already in LateOE and ME. A fully semanticized and regular epistemic use, however, does not occur before the EarlyMdE period.
(3) (1586 Apr. 30, Dudley [Traugott & Dasher 2005: 129]) ... surely his expences cannott be lytle, albeyt his grefe must be more to have no countenance at all but his own estate. ''... surely his expenses can't be small, although it must be an even greater grief to him that he has no standing other than his own estate.''
In chapter 4 the authors analyse case studies on the development of adverbials in discourse marker function. Three classes are discussed: (1) discourse markers that signal local connectivity, i.e. between utterances, (2) discourse markers that have developed intersubjective meaning, i.e. that explicitly mark the speaker's or writer's attention to the addressee, and (3) discourse markers that signal global connectivity, i.e. that serve discourse structuring purposes on a more global level. As to the first group, the authors discuss English 'indeed,' 'in fact,' and 'actually,' and find that all three underwent similar changes in the history of English, namely a development from clause-internal adverbial to epistemic sentential adverbial to discourse marker. As examples of discourse markers with intersubjective meanings English 'well' and 'let's' are discussed. With the former, for instance, Traugott & Dasher find a widening of contexts in which 'well' in discourse marker function occurs: ''At first it is anchored in the speech of others than the narrator; then it comes to be preempted to the narrator/speaker/writer's perspective, and finally it develops meanings with strong orientation to AD/R's [addressee's or reader's] face.'' (176) A similar increase in intersubjectivity is also shown for 'let's'.
In addition, however, 'let's' has also undergone ''a shift from content meanings based in argument structure at the clausal level to pragmatic procedural meanings based in argument structure at the discourse level.'' (177) The final case study in chapter 4 involves Japanese 'sate': it originates in a clause-internal adverbial function (similar to English 'thus') and gradually extents scope over the whole clause and, finally, over larger portions of the discourse. Concomitantly, the authors claim, an increase in subjectification is also given which eventually leads to an increase in intersubjective meaning when 'sate' serves as a hedging device and as an epistolary formula. The findings on adverbials with discourse-marker function, again, are summarized in four correlated paths of directionality in semantic development (cf. Traugott & Dasher 2005: 187):
The next series of case studies in chapter 5 concerns the development of performative verbs and constructions, in particular directives and declaratives. As an example of the first, the authors consider the semantic changes in the English verb 'promise.' The verb ''seems to have been coined from the noun'' (205), which, borrowed from Latin shortly after 1400, had a spatial meaning, i.e. ''a promise was something ''sent forward''.'' (205) The verb 'promise' seems to have undergone changes along two paths: it is likely to have been used performatively when introducing sentential complements in the first person present tense. This shows an increase in subjectivity since ''it expresses the speaker's approbation of authority as an actor attempting to match world to word.'' (209) A further change leads to the use of the verb as epistemic parenthetical in LateME:
(4) (1469 Paston I, 542 [Traugott & Dasher 2005: 207]) He losyth sore hys tyme her, I promyse yow. ''he is wasting his time badly here, actually.''
This use also is an instance of intersubjectification ''since it explicitly pays attention to AD/R's [addressee's or reader's] image needs in the here and now of the speech event. [... It acknowledges] at the discourse level that AD/R might have doubts about SP/W's message.'' (209) In contrast to this use of the verb, the second path of change is most frequent in third-person contexts. The first change is that to an epistemic meaning of the verb, i.e. ''portend'', a change that also shows an increase in subjectivity since it makes an explicit statement about the belief-state of the speaker. This use, again, seems ''to be a precursor for the development in the eighteenth century of nonperformative raising uses with nonfinite complements,'' as in example (5):
(5) (1700 Congreve, Way of the World, Act I [Traugott & Dasher 2005: 208]) I have seen him. He promises to be an extraordinary person;
One of two examples of declaratives discussed in this book is Japanese 'aisatu', which ''entered Japanese as a Zen term for a particular type of religious training that involved speech events.'' (219) A first change in meaning, namely that from ''question and answer'' to ''answer'', according to the authors, is indicative of subjectification. The final change is its associations ''with the speech act of ''greeting,'' apparently first as an indirect way of describing this act and more recently as a declarative performative to name a particular speech act as one of greetings.'' (219) This, in the view of the authors, shows an increase in intersubjectification. As a summary of their findings the authors name the following paths of semantic change (cf. Traugott & Dasher 2005: 225):
The final set of case studies involves the development of social deictics, which ''directly encod[e...] within their semantic structures the conceptualized relative social standing (superiority/inferiority, (non) intimacy, in-group versus out-group status, etc.) of a participant either in the CDE [conceptualized described event] or in the CSE [conceptualized speech event] by ''pointing'' to that social standing from the deictic ground (perspective) of SP/W relative to AD/R and other elements of the CSE.'' (226) In the centre of the authors' observations stand a subclass of social deictics, namely referent and addressee honorifics for Japanese predicate items (verbs, adjectives and copula). The semantic development of these seems to encompass four distinct stages: -- These honorifics usually start off as nonhonorific lexemes without any social deictic meaning component and are only concerned with the conceptualized described event (CDE). (Stage 1) -- This is the referent honorific stage: Lexemes on this stage are still primarily concerned with the described event, but they have acquired an additional social deictic meaning. As a consequence, the world of the conceptualized speech event (CSE) 'informs', i.e. interacts and enriches, the described event, because the deictic pointing realized in the referent honorific ''is made from a grounding in the CSE, namely SP/W's conceptualization of his or her status vis-à- vis AD/R.'' (236) (Stage 2) -- Some of these referent honorifics may develop into addressee honorific lexemes (Stage 3), -- and even further into addressee honorific affixes. (Stage 4)
The difference between the first and the last two stages lies in their dependence vs. independence from the described event: while referent honorifics ''point to the social standing of at least one ''referent,'' i.e. participant in the CDE, addressee honorifics directly encode the social deictic positioning of AD/R relative to SP/W independently of their possible roles in the CDE.'' (240) The difference between the two kinds of addressee honorific elements lies in the fact that the first still retains non-social-deictic meaning which continues to inform the conceptualized described event. With addressee honorific affixes, however, the non-social-deictic component does not interact with the described event but ''is procedural rather than contentful: it plays a role in the marking of discourse structure by orienting the utterance in which it appears to the ongoing discourse [...].'' (241) The kinds of lexemes that undergo the change from stage 1 to stage 2 point to the conclusion that although ''[m]any of these patterns of semantic change may appear to represent metaphoric shifts across conceptual domains [...] it is consistently the case that new predicate honorifics in Japanese develop from those L[exeme]s and constructions that index social status marking as a GIIN [generalized invited inference].'' (244) In addition, the authors claim that the rise of social deictic meaning ''intrinsically involves the development of procedural meaning'' (245), subjectification and intersubjectification.
Summing up, Traugott & Dasher again posit four correlated paths of semantic development in social deictics, where RefHon = Referent honorific, AddHon = Adressee honorific (cf. Traugott & Dasher 2005: 276):
The appeal and the strength of Traugott and Dasher's book, in my view, lies in the central role that the authors ascribe to language use, pragmatic considerations and principles, and the speaker or writer as ''the prime negotiator (with the AD/R) of reference and of meaning in general [...].'' (7) This focus on actual language use and, consequently, the discussion of contextualized data allow the authors to investigate the influence of invited inference in processes of semantic change, which eventually leads to the formulation of the model of the Invited Inference Theory of Semantic Change (IITSC).
In addition to emphasizing the importance of language use in semantic change, and in addition to providing a convincing explanation of semantic change, a further merit of this study is the identification of regular correlated paths of semantic change. These generalizations are interesting and insightful and provide testable hypotheses for future research.
The evidence the authors adduce to substantiate their claims is usually convincing, and the wealth and range (as regards aspects of the linguistic system as well as languages) of examples is very impressive. However, one has to keep in mind, that the authors provide 'only' a set of case studies. That is the authors manage to show that for some lexemes, semantic change can be explained by IITSC. They leave open, however, how these findings relate to other lexemes of the respective semantic domains, i.e. what is the proportion of instances of semantic change in one domain that can be explained by IITSC? It would have been interesting to have an exhaustive analysis of, say, all modal verbs in English to see if the mechanisms and paths the authors describe can also account for other tokens. This would help to substantiate their claim that ''if a lexeme with the appropriate semantics undergoes change, it is probable that the change will be of the type specified.'' (281) A further point of criticism concerns the sparseness of frequency data in the study (sections 6.4 and 6.5 being notable exceptions). Since the authors make use, for instance, of the Helsinki Corpus it would have been useful to provide information on the distribution of the different meanings over time. In particular, a more extensive use of corpus data would have been helpful in describing the step from invited inference to generalized invited inference. What are the possible contexts that allow an invited inference of a particular kind, and do all of these contexts lead to the development of generalized invited inferences? Information on these aspects would definitely have made this study still more convincing. So, although the study is based on corpus data, it is not a corpus-study in the strict sense: the corpora used, for the most part, do not seem to have been analysed exhaustively but rather been used as a convenient source for illustrative examples.
Another remark concerns the 'balancing' of languages in the case studies within the different semantic domains. The first three sets of case studies, i.e. discussion of modal verbs, adverbials with discourse marker function and performative verbs and constructions are primarily concerned with English data, although some information on Chinese and/or Japanese data is also given. The analysis of social deictics, in contrast, heavily draws on Japanese data and only provides supplementary analyses of English examples. A more balanced choice of material would have made the cross-linguistic relevance of the authors' finding even clearer.
A final note should be made with regard to the fact that some of the case studies in this book are based on earlier publications of the authors: those that are familiar with the works of Traugott and Dasher, therefore, will find less new information (although still enough!) in this book than those which are unfamiliar with the authors previous work.
On the whole, the book by Traugott and Dasher is a highly stimulating and interesting read, which, on the basis of a large range of examples, provides an innovative and convincing analysis of semantic change. Particularly appealing is the book's focus on language use and pragmatic mechanisms. This, in my view, is a very promising approach to semantic change, and it is hoped that further research along these lines will be conducted.
Horn, Laurence R. 1984. Toward a new taxonomy for pragmatic inference: Q-based and R-based implicature. In D. Schiffrin (ed.), Meaning, Form, and Use in Context: Linguistic Applications; Georgetown University Round Table '84, 11-42. Washington DC: Georgetown University Press.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Rolf Kreyer is an Assistant Professor of Modern English Linguistics in the department of English, American and Celtic Studies of the University of Bonn, Germany. His research interests include corpus linguistics, syntax, and text linguistics. He is the author of "Inversion in Modern Written English. Syntactic Complexity, Information Status and the Creative Writer", which will soon be published by Gunter Narr. At present he is working on a corpus-linguistic study that aims to analyse the interaction of language use and grammar.