The Grammaticalization of Verbs. Verbs as Sources of Grammatical Change

Contents

Contents

1. Introduction

2. Grammaticalization and its Meaning
2.1. The Evolution of Grammaticalization
2.2. The Rise of Grammaticalization Research
2.3. Recent Trends in Grammaticalization

3. Where Does Grammar Come From?
3.1. Mechanisms of Change: Reanalysis and Analogy
3.1.1. Reanalysis
3.1.2. Induction, Deduction and Abduction
3.1.3. Reanalysis and Grammaticalization
3.1.4. Analogy

4. Characteristics of Grammaticalization
4.1. Grammaticalization From a Diachronic and a Synchronic Perspective
4.2. Grammaticalization Chains
4.2.1. Overlapping (“Layering”)
4.2.2. Ambiguity
4.2.3. Asymmetry
4.2.4. Unidirectionality
4.2.5. Generalization
4.2.6. Frequency
4.2.7. Decategorialization
4.2.8. Specialization
4.2.9. Increase in Bondedness
4.3. Degrammaticalization

5. Motivations for Grammaticalization
5.1. Creativity versus Routinization
5.2. Inference
5.2.1 From Less to More Subjective
5.3. Metaphor and Metonymy
5.3.1. Metaphorical Extensions
5.3.2. Metonymic Strategies

6. From Source to Target – Basicness as a Relevance Factor

7. Auxiliary Verbs
7.1. Auxiliation Chains
7.1.1. Verb-to-TAM Chains
7.2. Stages of Auxiliation
7.3. Sit, Stand and Lie as Aspectual Markers
7.3.1. The Evolution of the Sit/Stand/Lie Aspectual Structure: Functional Need?
7.3.2. The Shift from Locative to Temporal Meaning
7.4. The Future: It Comes, It Goes, It Has to Be
7.4.1. Pathways of Future
7.4.2. From Desire to Prediction
7.4.3. From Motion-in-Space to Progress-in-Time
7.4.5. Obligation Futures
7.5. The Case of Used to

8. From Verb to Preposition
8.1. Prerequisites and Conditioning Factors
8.1.1. European Languages
8.1.2. Serial Verb Languages
8.2. Semantic, Morphological and Phonological Changes Involved
8.2.1. Between Verb and Preposition
8.2.2. Coalescence and Phonological Erosion
8.3. Source and Target Domains of Deverbal Prepositions

9. The Evolution of Complementizers
9.2. Evidence from Other Languages
9.3. Reanalysis at Work
9.4. Universals versus Substrate

10. A New Quotative Marker: English Be Like
10.1. A Teenage Phenomenon?
10.2. A Notoriously Polyfunctional Item
10.3. Origin and Evolution of Like
10.4. Subjectification

11. Conclusion

12. Summary (Zusammenfassung)
1. Einleitung
2. Die Entstehung Grammatischer Formen und Konstruktionen
3. Kognitive Prozesse: Metapher und Metonymie
4. Grammatikalisierungskanäle
5. Grammatikalisierungsprozesse und Unidirektionalität
6. Verben als lexikalische Quellen für Grammatikalisierungsprozesse

13. Bibliography

Buchstaller, Isabelle (2002), He Goes and I’m Like: the New Quotatives Re-Visited. Internet

Proceedings of the University of Edinburgh Postgraduate Conference.

http://www.ling.ed.ac.uk/~pgc/archive/2002/pgc02-programme.html.

Figures

Abbreviations

1. Introduction

Languages don’t change; people change languages

(William Croft, 1990)

How and why do grammatical structures evolve? This question has been extensively discussed by linguists concerned with the study of grammaticalization, a discipline which may be defined as that part of language theory which focuses on “the interdependence of langue and parole”, dealing with the organization of categories and of coding (Traugott and Heine 1991b:1) or as Haspelmath (1999:1044) puts it, “grammaticalization shifts a linguistic expression further towards the functional pole of the lexical-functional continuum”.

The phenomenon under investigation is a universal one for there is a striking cross-linguistic consistency of the lexical sources of particular targets, i.e. grammatical forms. These regularities of grammaticalization pathways are interpreted as reflections of universal aspects of human cognition and perception. Precisely how grammatical forms evolve out of lexical structures is the main issue of this paper, and it will be argued that grammatical structures are shaped by discourse in an ongoing process. Following this view, grammar is therefore merely the label used for “certain categories of observed repetitions in discourse”. (Hopper 2002:156) Following this, frequent repetition in discourse plays a crucial role in the development of grammatical forms. The famous dictum, “grammars code best what speakers do most” is a central postulate of all discourse-based approaches to grammaticalization, and points to the assumption that grammars reflect coding mechanisms for those speech functions which speakers most often perform. (DuBois 1985:362-63) Analogically, entities of high frequency are candidates which are most likely to enter grammaticalization paths.

Furthermore, it will be argued that basicness is an inherent characteristics of most source concepts. It has been observed that for any given grammatical domain, there is only a limited number of lexical items that are likely to be sources for grammaticalization. Most of these constitute very basic human concepts and activities, depending on the socio-cultural situation in which the language is spoken. Since verbs form the core element of every sentence, expressing different conditions such as states, changes, activities, achievements etc., they provide a suitable source for grammatical targets. In Heine and Kuteva’s (2002) World

Lexicon of Grammaticalization, a book summarizing the most salient generalizations on the change of grammatical constructions, 53 out of 173 source concepts are verbs.

This paper consists of two major parts, each divided into chapters. The first section sets out the theoretical background for the study of grammaticalization and the second outlines some common pathways of verbs as sources for grammaticalization.

Chapter 2 gives an overview of the research history of grammaticalization in order to provide a better understanding of this tradition.

Chapter 3 describes reanalysis and analogy, the two mechanisms involved in linguistic change and demonstrates how they interact within grammaticalization.

Chapter 4 details the characteristics of grammaticalization chains. The specific functional processes involved in the transition from lexical to grammatical material, i.e. semantic, phonetic and morphosyntactic changes, are set out and illustrated. The chapter concludes with examples of so called degrammaticalization, that is, cases of the reversed development (from grammatical to lexical items).

Chapter 5 attempts to locate the motivating factors behind grammaticalization. It will be demonstrated that although both metaphor and metonymy play a crucial role, the role of individual communicative needs must be taken into consideration.

Chapter 6 provides a bridge to the second part of the paper by listing some of the main verbal sources.

Chapter 7 deals with the evolution of auxiliaries, which encompasses the majority of verbal pathways of grammaticalization. The verb-to-tense/aspect/modality-chain is illustrated by (i) posture verbs which typically become aspectual markers, and (ii) by verbs of desire, motion and obligation that typically evolve into future markers.

In Chapter 8 European languages and languages with serial verb constructions are compared in regard to their potential of providing deverbal prepositions.

Chapter 9 is concerned with the evolution of complementizers, which often develop out of verba dicendi, such as say.

Finally, Chapter 10 illustrates the recent development of be like to one of the most popular quotatives in English speaking countries. It may not involve the grammaticalization of a verb, but it is an interesting case of a frequent lexical item taking on new functions in a very short period of time.

2.Grammaticalization and its Meaning

Grammaticization, grammatization or grammaticalization? A variety of different terms are used for a phenomenon that has attracted a lot of attention in the past decades. The disagreement regarding its name is reflective of the diversity of perspectives to this subject. Thought to be less than a hundred years old, the term ‘grammaticalization’ is used for at least two different, yet related processes. The one most commonly referred to is the evolution of a grammatical morpheme out of a lexical morpheme. For example, positional verbs such as sit, stand and lie may become markers of durative aspect. This type of development is labelled the ‘lexical item morpheme model’ and originates in Meillet’s account of grammaticalization which will be presented in the following chapter. The second and more recent tradition is associated with Talmy Givón’s ‘syntacticization’, a process “by which loose, paratactic, ‘pragmatic’ discourse structures develop - over - time into tight, ‘grammaticalized’ syntactic structures.” It considers grammaticalization to be a syntactic, discourse-pragmatic phenomenon. (Givón 1979:208) Syntacticization and the rise of grammatical morphology often go hand in hand and both models are therefore sometimes regarded as mutually dependent parts of the same process. (Givón 1979:220-22)

This chapter gives a short survey of the history of grammaticalization, its study and its major scholars. More detailed surveys can be found in Heine et al. (1991a), Hopper and Traugott ([1993]2003) and C. Lehmann (1995).

2.1. The Evolution of Grammaticalization

Although there are links going back to von der Gabelentz, von Humboldt and even to Horne Tooke and Condillac, grammaticalization as a term was coined in 1912 by the French linguist Antoine Meillet, who defined it as “le passage d’un mot autonome au rôle d’élément grammatical”. (Meillet 1958:131) His article L’évolution des formes grammaticales is the first full-length paper on grammaticalization and marks the beginning of a perspective of grammaticalization which still prevails today. Meillet’s notion of grammaticalization hints at the study of the history of particular grammatical forms. (Hopper 1991:18) In his view, the transition from lexical items (mots principaux) to morphemes fulfilling grammatical function (mots accessoires) is gradual: “Et il y a tous les dégrés intermédiaires entre les mots principaux et les mots accessoires.” (Meillet 1958:135)

According to Meillet, new grammatical forms emerge through two processes. First, through analogy, whereby new paradigms come into being through formal resemblance to already existing paradigms, e.g. the replacement of plural shoen by shoes through analogy to the established plural –s. Secondly, they can develop through a process known as grammaticalization, defined as the “attribution of a grammatical character to an erstwhile autonomous word”. Meillet illustrates this process with the French verb être ‘to be’, whose meaning ranges from a full existential ontological sense, as in je suis celui qui suis ‘I am the one who is’, to a structure with a less fully locative sense in je suis chez moi ‘I am at home’, to an almost redundant sense, as in je suis malade ‘I am ill’, and finally to a purely grammatical function as a tense-aspect auxiliary in je suis parti ‘I left’. (Hopper and Traugott [1993]2003:19, Meillet 1958:131) Later in the same article, he goes even further by pointing out that,

Whereas analogy may renew forms in detail, usually leaving the overall plan of the system untouched, the ‘grammaticalization’ of certain words creates new forms and introduces categories which had no linguistic expression. It changes the system as a whole. (Meillet 1958:133)

How does such a change come about? In Meillet’s opinion, the main reason is a loss of expressivity in frequently used collocations, whose function may be rejuvenated through new collocations that perform the same role. Loss of expressivity is often accompanied by the weakening of phonological form and concrete meaning. Obviously, Meillet’s account of grammaticalization is strongly influenced by the “classical” attitude toward language that equates change with deterioration. (Hopper and Traugott [1993]2003:24-25) Still, his paper is considerably rich in its insights and even though subsequent works on grammaticalization have modified Meillet’s view, it still presents a challenging concept around which to create related modern ideas.

2.2. The Rise of Grammaticalization Research

The mid 20th century saw the rise of structuralism with its insistence on the clear separation of langue and parole, and of synchrony and diachrony. Being viewed as a diachronic phenomenon only, the topic of grammaticalization was hardly dealt with, as the primary interest amongst linguists was devoted to synchronic typology. (Heine et al. 1991a:10) However, some Indo-europeanists were still interested in this then ‘unfashionable’ topic and among the most influential figures were Jerzy Kurylowicz and Emile Benveniste ¹. Kurylowicz ([1965] 1975:52) provided the by now classical definition of the term ‘grammaticalization’:

Grammaticalization consists in the increase of the range of a morpheme advancing from a lexical to a grammatical or from a less grammatical to a more grammatical status, e.g. from a derivative formant to an inflectional one.

In the context of questioning the autonomous syntactic theory and amid general dissatisfaction with rigid structuralism, grammaticalization became a major theme of general linguistics in the 1970s. (Hopper and Traugott [1993]2003:26) Grammaticalization was seen as a means of explaining language change in a less ‘static’ manner than structuralism and generative transformational grammar would have allowed. Meillet’s often cited words, un système où tout se tient, suddenly made sense because it came to be accepted that not only the sentence as the basic unit of language, but also the processes of grammaticalization needed to be taken into account in order to explain linguistic changes. (Lehmann 1994:176) Inspired by typological works, Talmy Givón introduced new parameters for the explanation of linguistic behaviour: “In order to understand current morphologies and morphotactics of a language, one must construct specific hypotheses about the syntactic order and transformational structure of the language at some earlier stage of its historical development.” (Givón 1971:394) Based on the famous slogan “Today’s morphology is yesterday’s syntax” he demonstrated how certain verb forms in a number of African languages developed from collocations of pronouns and independent verbs into stems with affixes. (Hopper and Traugott [1993]2003:26) Later on, Givón focussed on another paradigm case of linguistic evolution, one that can be described roughly as, “Today’s syntax is yesterday’s pragmatic discourse”. Discourse pragmatics came to be recognized as an important means for understanding language structure in general and the development of syntactic structures (‘syntacticization’) and grammatical categories in particular. (Heine 1991a:13) In Givón’s view, the motivation for syntacticization is caused by many communicative factors pertaining to the immediate situation in which communication takes place, i.e. the degree of time pressure or the amount of shared presuppositional background. ² (Givón 1979:268) The newly evolved syntactic structures are in time eroded through morphologization, lexicalization and phonological attrition but instead of an increasing syntacticization over time, the result is cyclic waves (Givón 1979:208-09):

Discourse Syntax Morphology Morphophonemics Zero

This line of research opened a new era in grammaticalization studies, based on the view that grammaticalization was not just the reanalysis of lexical as grammatical material but the reanalysis of discourse patterns as grammatical patterns. In search for language universals, many linguists applied the idea of grammaticalization to general problems of synchronic description. Li and Thompson (1974) explained this view by demonstrating that the shift from SVO to SOV in Chinese is the result of a process in which verbs assume a grammatical function. If the first verb (V1) in a sequence S-V1-O-V2 grammaticalizes to a case marker, the result will be an SOV structure and V2 becomes the only verb in the sentence.

Cross-linguistic work by Joseph H. Greenberg provided more information about the connection of synchronic syntax and grammaticalization in that it admitted that diachronic principles are also involved in the explanation of synchronic generalizations. (Greenberg 1978)

In 1976 Ronald W. Langacker articulated the concept of “cognitive grammar”, a theory of linguistic structure that viewed grammar as neither generative nor constructive, but intrinsically symbolic. He described grammar as an inventory of symbolic units with semantic and phonological content, which are used to construct appropriate expressions. (Langacker 1988a:4-5) As opposed to orthodox generative theory he did not see grammar as an autonomous dimension of linguistic structure but as being linked directly to the lexicon and semantics: “A description of grammatical structure that makes no reference to meaning is ultimately no more revealing than a dictionary providing only a list of undefined forms”. (Langacker 1988b:49) Langacker (1977:65) compared languages in their diachronic aspect to “gigantic expression-compacting machines” that, …require as input a continuous flow of creatively produced expressions formed by lexical innovation, by lexically and grammatically regular periphrasis, and by the figurative use of lexical or periphrastic locutions. The machine does whatever it can to wear down the expression fed into it. It fades metaphors by standardizing them and using them over and over again. It attacks expressions of all kinds by phonetic erosion. It bleaches lexical items of most of their semantic content and forces them into service as grammatical markers. It chips away at the boundaries between elements and crushes them together into smaller units. The machine has a voracious appetite. Only the assiduous efforts of speakers – who salvage what they can from its output and recycle it by using their creative energies to fashion a steady flow of new expressions to feed back in – keep the whole thing going.

2.3. Recent Trends in Grammaticalization

The first modern work providing a comprehensive survey of studies in grammaticalization up to the 1980s is Christian Lehmann’s Thoughts on Grammaticalization. ³ Lehmann introduced a set of “parameters” according to which degrees of grammaticality can be measured synchronically and diachronically (cf. Figure 1) His concept is built on the fact that the grammaticality of a linguistic sign is converse to its autonomy and grammaticalization minimizes autonomy. Consequently, the degree of autonomy of a sign tells us about its degree of grammaticalization. In this theory, three aspects are relevant for determining the autonomy, namely “weight” (i.e. the property that renders a sign distinct from the members of its class), “cohesion” (i.e. the factor inherent in relations with other signs, that detract from autonomy) and “variability” (i.e. the mobility or shiftability of a sign with respect to other signs). These three aspects are associated with “the fundamental aspects of every linguistic operation”, Saussure’s ([1916]1986) two axes: (i) the paradigmatic axis (selection of a sign on one slot) and (ii) the syntagmatic axis (combination of signs in sequence). The weight of a sign, viewed paradigmatically, is its integrity, i.e. its substantial size, semantic- and phonologicalwise. Viewed syntagmatically, it is the structural scope, that is, the extent of the construction which it enters or forms, that constitutes the weight of a sign. In a paradigm, the cohesion of a sign with others is called its paradigmaticity, i.e. the degree to which it enters a paradigm and is integrated within it. Syntagmatically, this cohesion is called bondedness, the degree to which it depends on other signs. The possibility of using other signs in its stead or of omitting it altogether is called the paradigmatic variability. The syntagmatic variability of a sign refers to the extent to which it can be shifted within the construction. In his view, the process of grammaticalization consists of these six parameters and cannot exist independently of them. It is a combination of decrease in weight, increase in cohesion and a decrease in paradigmatic and syntagmatic variability. (Lehmann 1995:122-24)

Abbildung in dieser Leseprobe nicht enthalten

Figure 1 The parameters of grammaticalization (taken from Lehmann 1995:123)

In the 1990s increasing attention was paid to semantics and pragmatics and the elements underpinning semantic change. In particular, the role of metaphor and metonymy in the flow of speech was a much debated issue. In 1991, Heine, Claudi and Hünnemeyer made an important contribution to this approach with Grammaticalization: A Conceptual Framework. Providing rich exemplification from African languages, the authors described grammaticalization as a result of problem solving, in which speakers use existing forms for new functions. Lexical items from concrete domains are pressed into service to express concepts in more abstract domains. The transition from source to target structures is seen as a metaphorical abstraction. (Heine et al. 1991a:48) The development of English be going to from a verb of motion to a future time marker corresponds to the metaphor TIME IS SPACE. The domain of spatial movement is used as a metaphorical vehicle to refer to the domain of deictic time. Sentences like (2) have developed historically from sentences like (1) since the 15th century. (Heine et al. 1991a:46-47)

(1) Henry is going to town.
(2) It is going to rain.

Verbs indicating movement towards a goal (e.g. go, come), have become future markers in languages around the world, such as in Spain, Africa, America, Asia and the Pacific. (Bybee 2002:148)

Certain typical pathways of grammaticalization are observable among different languages, for example (Heine and Kuteva 2002):

(3) - main verb auxiliary

- say complementizer

- positional verbs, e.g. sit, stand copula be

- durative verbs, e.g. keep, exist, live continuous

- give benefactive

- verbs of motion, e.g. go, come future

- do continuous, causative

Elizabeth C. Traugott contributed a new framework of grammaticalization by concentrating on pragmatic principles of meaning change in the process of grammaticalization. She argued that grammatical change is usage-driven and that there are semantic-pragmatic factors leading to a (unidirectional) change of meaning, specifically changes from concrete meanings to more abstract ones. (Hopper and Traugott [1993]2003:33). For example, the definite article the derived in early modern English from the non-proximal demonstrative that in Old English. Once the new form the developed, the demonstrative was partially relieved of this function and came to denote (i) the speaker’s distance from the objects in the situation outside the text and (ii) the speaker’s evaluative distance. The definite article gained an expressive function in addition to that of its cohesive one because it developed a participant-oriented function, namely that of presenting what is talked about as if though it were known to the hearer, e.g. The woman was walking down the street. (Traugott 1982:250-52) Similar processes have taken place in many other languages such as Hungarian, Bizkaian Basque and several pidgins and creoles. (Heine and Kuteva 2002:109-11) More recently, Traugott has refined her framework by emphasizing the importance of strengthening of informativeness, conversational implicatures and metonymy concerning the question as to how such meaning changes come about. (Traugott and König 1991)

The importance of conventionalization of implicature for semantic change is also recognized by Bybee, Perkins and Pagliuca in The Evolution of Grammar: Tense, Aspect, and Modality in the languages of the World (1994), based on a sample of 76 languages whose TAM systems were surveyed from both a formal and semantic point of view. Arguing against the widespread claim that languages develop grammatical categories because they need them, the authors suggested that grammatical evolution is driven by the communicative necessity to be more concrete and specific. (Bybee et al. 1994:297-300)

In recent times, the nature and status of grammaticalization was approached from different directions. On the one hand, important works on grammaticalization theory were published in order to manifest its independent status. Hopper and Traugott’s comprehensive work Grammaticalization ([1993]2003) drew a thorough picture of the state of the art including studies in creolization. The authors argued that grammaticalization involves morphosyntactic change initiated by pragmatic and semantic changes. Ten years later, it was updated, and grammaticalization, formerly described as a process, was re-defined as a change in which “lexical items and constructions come in certain linguistic contexts to serve grammatical functions and, once grammaticalized, continue to develop new grammatical functions.” (Hopper and Traugott [1993]2003:231) On the other hand, more and more expressions of concern about the status of grammaticalization were noted. In 2001, an entire volume of Language Sciences was dedicated to the topic ⁴, putting forward the claims of detractors that grammaticalization is “no distinct theory” but “derivative of other kinds of language change” and “merely involves other kinds of changes and mechanisms of change which are well understood and are not limited to cases involving grammaticalization”, finally concluding that “there is no such thing as grammaticalization”. (Campbell 2001:116-17, Newmeyer 1998:226)

Until today, a wide spectrum of themes have been the subject of disagreement among scholars, including questions such as the following: (i) How far can grammaticalization be considered a universal diachronic process or mechanism of change and in how far is it conditioned by synchronic factors? (ii) What is the role of the speaker in grammaticalization? (iii) Does grammaticalization itself provide a cause for change or is it an epiphenomenon, that is, a conglomeration of causal factors and mechanisms which elsewhere occur independently? (iv) If it is epiphenominal, how do we explain that similar pathways so often occur in known cases of grammaticalization? (v) Is grammaticalization unidirectional? (vi) What is the nature of the parameters guiding grammaticalization? (vii) What phenomena in language are not examples of grammaticalization? (viii) To what extent do cases of grammaticalization in contact situations contribute to the discussion of what motivates grammaticalization changes?

Most of these questions are still of relevance today, including the question of directionality and the difference between grammaticalization and lexicalization. Counterexamples to grammaticalization were advanced, as were phenomena that have much in common with the classical type of grammaticalization and yet lack some perceived crucial components. This has led to numerous publications for and against unidirectionality and to publications trying to define The Limits of Grammaticalization (Ramat and Hopper 1998).

Summing up, grammaticalization has become a very popular topic in recent years, not only among functionally oriented grammarians. ⁵ Subsequently, many different approaches were put forward in order to get at the very nature of grammaticalization which Traugott defines “in the most neutral terms possible, (…) a subset of crosslinguistically recurring changes, that involve correlations across time between semantic, morphosyntactic (and sometimes also) phonological changes”. (Traugott 2001) Despite these developments, a “standard model” of grammaticalization is difficult to define. As Bisang (2003) pointed out, the creation of a special theory of grammaticalization should not be of main interest, but rather the different factors leading to the evolution of grammatical structures. Though universals exist, there are factors that vary across languages, making research in this field so challenging.

3. Where Does Grammar Come From?

Back in 600 B.C., the Egyptian king Psametich II sent some young children to a lonely island where they were raised in isolation and were taught a language without grammar. After some time they were left alone and later generations developed their own grammar different to Egyptian grammar but still typical for a human language. Experiments like these have provoked questions about where grammar actually comes from and how it evolves.

Two major views are persistent in this respect. Chomskyan linguists or nativists, on the one hand, would argue for the existence of an innate grammar. This model proposes that human beings are genetically endowed with Universal Grammar, which consists of unchanging “principles” characterizing the fundamental structure of language and restricting the class of attainable grammars, and of “parameters” that are constructed by experience and which define possible variation. Thus differences between languages across space and time are explained as results of different settings of the parameters in the process of language acquisition. (Hopper and Traugott [1993]2003:45). The language learner is conceived as a language acquisition device, a passive being with a rich language-specific genetic endowment that selects relevant stimuli according to internally present criteria. (Lightfoot 1991:2)

The anti-nativist view on the other hand, denies the existence of an innate grammatical coding and sees language as existing in a state of constant flux, adapting itself to the language learner’s spontaneous assumptions about communication, learning and social interaction. (Hopper and Traugott [1993]2003:46). Following this view, language is not an isolated neural capacity and therefore language competence is not autonomous of a person’s other cognitive abilities or social skills. Grammar develops automatically when people communicate and may be regarded a by-product of everyday conversation. Grammatical items evolve out of lexical ones in order to facilitate complex constructions and tighten loose discourse structures. (Haspelmath 2002:2+9, Givón 1979:208, Heine and Reh 1984:81). In Hopper’s opinion ( 1987:148), grammar, if existent at all, is „always emergent but never present“ and “it’s forms are not fixed templates, but are negotiable in face-to-face interaction in ways that reflect the individual speakers’ past experience of these forms, and their assessment of the present context”. (Hopper 1987:142) Lichtenberk (1991a:76) adopts the same view:

Languages may strive toward greater regularity and iconicity by eliminating anomalies and variations; at the same time, however, new patterns emerge elsewhere n the grammar, introducing new anomalies and new variation. Grammars are always noncomplete.

Whereas Chomskyans generally study competence in isolation from performance, functionalists argue that competence evolves from performance and that as a result, the two are intertwined. In his Cours de linguistique générale Saussure writes “La langue est à la fois l’instrument et le produit de la parole”. Accordingly, “the study of grammaticalization challenges the concept of a sharp divide between langue and parole, and focuses on the interaction of the two”. (Traugott and König 1991:189)

3.1. Mechanisms of Change: Reanalysis and Analogy

3.1.1. Reanalysis

According to Langacker, reanalysis is “a major mechanism of syntactic evolution which we must understand in depth if we wish to understand how and why syntactic change occurs” and he defines it as “change in the structure of an expression or class of expressions that does not involve any immediate or intrinsic modification of its surface manifestation”. (Langacker 1977:57-58) In other words: the hearer understands a form to have a structure and meaning that are different from those of the speaker, e.g. when [ Hamburg ] + [ er ] “piece of fast food from Hamburg” is understood as [ Ham ] + [ burger ] and later on, new words are coined with burger, like cheeseburger, fishburger. (Hopper and Traugott [1993]2003:50) Shakespeare used nuncle ‘uncle’ (Nuncle, give me an egg, and I’ll give thee two crowns, King Lear, I, 4., 153), which was derived from a reanalysis based on the final – n of the possessive pronouns mine and thine before it was lost: mine + oncle mine noncle my nuncle. This form has survived and can be found today in certain dialects. (Campbell 1998:103) Another example is the borrowing process of Italian l’alicorno ‘the unicorn’ to French la licorne ‘the unicorn’, in which the initial a of alicorno was reanalyzed as part of the French feminine definite article la. (Detges and Waltereit 2002:158-59)

Reanalysis occurs not only with single words but also with syntactic constructions. In present-day English, the sequence try + VERB is reanalyzed as AUXILIARY + VERB, as I’ll try and contact her. In this form, try is used differently than in I’ll try to contact her because try and are treated as a single word: First, try and and are intonationally and phonetically bound (‘try-ən’). Second, only try is possible in this sequence, but not tried, trying, tries: * He tries and contacts her. And finally, it is not possible to separate try and and, e.g. with adverbs: * I’ll try hard and contact her. Furthermore, try and in this context signals the agent’s inability to achieve the complement verb and the speaker’s lack of confidence in the agent’s success, and therefore exposes modal-like qualities. (Hopper and Traugott [1993]2003:50) Another case of reanalysis is be going to from be + main verb + progressive aspect + purposive preposition to tense marker. Further, got was reanalyzed as a finite verb. When the form have got ‘possess’ underwent auxiliary contraction to the point of deletion (4) it was only a small step to consider non-finite, past participial got as a finite verb synonymous to have (5):

(4) I have got a question I’ve got a question I got a question
(5) a. I don’t got any questions.
b. Do you got any questions?

This form is widely used in American English but it is not Standard English. The reanalysis of got as a present-tense form has even led to the creation of a third singular gots in some varieties of English, especially African American Vernacular English. ⁶

3.1.2. Induction, Deduction and Abduction

Every case of reanalysis is the result of the following process: the hearer observes a result (i.e. the verbal activity of the speaker), then he invokes a law (grammatical rules) and infers that something may be the case (formulation of the hearer’s linguistic structure). For example, the French-speaking hearer matches the sound chain l’alicorno with French NPs, as la liaison ‘the binding’, adopts it to his own language and the result is la licorne. This is called abduction, termed by Charles S. Peirce and it is one of the three types of reasoning, that is, the human ability to reason from the form of what is said to the intent of what is said. (Hopper and Traugott [1993]2003:42) The other two reasoning types are induction (reasoning proceeds from observed cases and results to establish a law) and deduction (the hearer applies a law to a case and predicts a result). Of these three modes of reasoning it is only abduction that can originate new ideas. (Andersen 1973:775)

3.1.3. Reanalysis and Grammaticalization

It is important to note that reanalysis is not identical with grammaticalization. They are rather “two independent principles showing a number of significant correlations”. (Heine and Reh 1984:97) Whereas grammaticalization typically involves unidirectionality, irreversibility and gradualness, reanalysis may be bi-directional, reversible and abrupt. Both are consequences of communicational needs. The purpose of reanalysis is understanding, that is, it is typically a listener’s strategy. In contrast, grammaticalization is the result rather of communicative strategies on the part of the speaker. (Detges and Waltereit 2002:152) It could be said, that reanalysis is the more fundamental phenomenon as it occurs in any type of functional change and thus not every instance of reanalysis is a case of grammaticalization. (Detges and Waltereit 2002:190)

3.1.4. Analogy

As was said above, Meillet made a distinction between grammaticalization (in the sense of reanalysis) and analogy. The first refers to rule changes, the latter to rule spread (i.e. generalization), that is, surface manifestations are modified according to already established regularities. Only reanalysis can create new grammatical structures. (Hopper and Traugott [1993]2003:64) Typically, reanalysis is covert, i.e. it is not directly observable and it operates on the syntagmatic axis. By comparison, analogy involves paradigmatic organization and is overt, i.e. it makes the unobservable changes of reanalysis observable and thus plays an important role in the study of grammaticalization. The interaction of the two mechanisms is illustrated in Figure 2. In Stage I be going to is the progressive of the directional verb with a purposive clause. Stage II is the stage of the future auxiliary with an activity verb resulting from reanalysis. Stage III is the consequence of extension through analogy to all verbs, including stative verbs. Finally, in Stage IV be going to is reanalyzed as gonna.

Abbildung in dieser Leseprobe nicht enthalten

Figure 2 Schema of the development of the auxiliary be going to (taken from Hopper and Traugott [1993]2003:69)

4. Characteristics of Grammaticalization

The younger rises when the old doth fall.

(William Shakespeare, King Lear)

Grammaticalization is the type of language change in which an expression moves away from the lexical pole and toward the grammatical pole. Two variants are identified. The grammatical formative may evolve out of something other than a grammatical formative or out of a grammatical formative with a weaker degree of grammatical function:

Abbildung in dieser Leseprobe nicht enthalten

The development of lexical items into grammatical ones is an “expected, natural process”. (Givón 1973:921) This shift is conceptualized as a pathway along which forms evolve. This so-called “cline of grammaticality” is illustrated as follows: (Hopper and Traugott [1993]2003:7)

content item grammatical word clitic inflectional affix ( zero)

This transition is usually gradual, continuous and unidirectional, i.e. the opposite direction away from the grammatical pole and toward the lexical pole is rather rare. (Geurts 2000:781)

One aspect of grammaticalization is a “specific evolutional processes, in particular a decrease in semantic and phonetic complexity and syntagmatic variability”. (Heine and Reh 1984:96). On the other hand however, there is an increase in abstractness, generalization and pragmatic specification and strengthening through inferencing. (Traugottt 1996:183) This chapter concerns the typical traits and effects of grammaticalization, and the phonetic, morpho-syntactic and semantic processes involved.

4.1. Grammaticalization From a Diachronic and a Synchronic Perspective

Similar to language, grammaticalization has been approached from two perspectives: that of its structural change in time (“diachronic” or historical) and that of its structure at a single point in time (“synchronic”). As a result, the rigid distinction between synchrony and diachrony has been called into question not only by work on discourse analysis and language in use but also by the study of grammaticalization. (Hopper and Traugott [1993]2003:2)

Viewed diachronically, grammaticalization is thought of “as that subset of linguistic changes whereby lexical material in highly constrained pragmatic and morphonsyntactic contexts becomes grammatical, and grammatical material becomes more grammatical”. (Traugott 1996:181) Such an approach considers the sources and targets of grammatical forms and the steps of change. Typically, “the source meaning uniquely determines the grammaticization path that the gram will travel in its semantic development”. (Bybee et al. 1994:9) In most cases, multiple grammatical senses developing from the same source are assumed to be related to the core meaning of the source item. For example, in many languages the same verb expresses possession and existence. In Mandarin Chinese, you ‘have, exist’ expresses possessive and existential meaning. Both meanings are linked for the relation of possession between two entities is a kind of relation of existence as well. That A possesses B is equivalent to that B exists within the domain of A. (Her 1991:383)

When viewed from a synchronic perspective, grammaticalization is hypothesized to be a syntactic, discourse pragmatic phenomenon. One line is concerned with cross-linguistic work exploring universal paths of coding the same kind of discourse-pragmatic structure. Another deals with the study of alternative uses of the same form including the task of identifying a cline from more to less categorial or prototypical uses. (Traugott 1996:185) For example, the German verb scheinen is realized in different types of constructions. It may function inter alia as an intransitive main verb (6a), as a copula (6b) and as a parenthetical verb (6c). (Diewald 2001:94)

(6) a. Die Sonne scheint
‘The sun is shining’

b. Sie scheint verärgert (zu sein)
’She seems to be annoyed’

c. Er ist scheints nicht zu Hause
’He seems not to be at home’

In the 16th century, scheinen developed from a main verb with an adverbial function to a copula verb. At first, it appeared only with the infinitive verb sein ‘to be’, as in (7a). Later on, the particle zu was added and became obligatory (7b); scheinen was generalized and thus came to be used with other verbs (7c). (Diewald 2001:100-03)

(7) a. Es scheinet wol ein geringer gottesdienst sein ( Martin Luther)
b. Und ob ich zwar scheinte nur ein zwerg zu seyn gegen meinen feind
c. (…) sie schien sich nach dem Fremden umzusehen (Johann Wolfgang von Goethe)

It is very likely that the development of scheinen was influenced analogically by the grammaticalization of modal verbs such as müssen, können, dürfen, sollen and wollen until the 17th century. (Diewald 2001:107)

4.2. Grammaticalization Chains

Basic to work on grammaticalization is the concept of “chains” or “clines” that reflect the process of change and help to reconstruct it. A grammaticalization chain is a structure where sequent stages of the lexical-to-grammatical development of linguistic entities are linked successively. (Kuteva 2001:10)

Usually, forms do not shift abruptly from one category to another, but go gradually through a series of transitions that tend to be similar in type across languages ⁷. For example, body part nouns like back often come to denote spatial relationships (in/at the back of) and then develop into adverbs and prepositions and even case affixes. (Hopper and Traugott [1993]2003:6) A cline for this kind of change looks like this:

lexical noun relational phrase adverb preposition case affix

At this point it is useful to distinguish between clines or grammatical chains and “grammaticalization channels”. Best conceived as the route taken by linguistic signs when being in change, grammaticalization channels constitute the alternative options available to languages for developing new grammatical categories. For example, English has used two channels of grammaticalization to introduce future markers, i.e. the domain of obligation and desire (shall, will) and the domain of motion (go). A grammaticalization chain concerns the internal structure of a channel, i.e. what has happened on the way from a verb or auxiliary to a future morpheme, or more precisely, how the process from motion through intention and immediate future and finally to future is structured. (Heine et al. 1991a:221-22)

In many cases, one morpheme provides the source for more than one grammaticalization pathway and the result is a so-called polygrammaticalization chain. (Kuteva 2001:117) For example, besides denoting future time, the verb go has acquired a whole range of grammatical functions worldwide and has been the source for tense-aspect functions, andative markers, conjunctions, imperatives, allative prepositions and purpose clause markers. (Heine and Kuteva 2002:155-65) The English verb get can denote possession, movement, causation, obligation and change-of-state among other senses. (Gronemeyer 1999:1) Another relevant example is the German verb werden that was grammaticalized to a future and a passive auxiliary. (Diewald 1997:113)

4.2.1. Overlapping (“Layering”)

Naturally, the transition from one concept to another one involves overlapping of the adjacent concepts because the process is gradual rather than discrete. Between the concepts is an intermediate stage (the so-called split, according to Heine and Reh 1984:57) where both exist side by side: (8) A A,B B This condition is also called layering. As new layers emerge within a functional domain, the older layers are not immediately discarded but may remain and interact with the newer ones. (Hopper 1991:22) A prominent example is the future tense in English and its nuances: (Leech 1971:56)

Abbildung in dieser Leseprobe nicht enthalten

Layering or overlapping takes place not only within a functional domain but may also refer to one single morpheme. When an expression is transferred from source to target concept, that is, when a lexical item splits into a lexical and a new grammatical item, the result is multiple meaning as long as one and both functions remain. (Heine 1993:48) This special case of

overlapping is called divergence (Hopper 1991:24) and is best illustrated by English to keep, which has two co-existing meanings:

(10) a. They keep the money.
b. They keep complaining.

Originally, keep was used as a main verb only in contexts such as (10a) and later it became an auxiliary extending its use to contexts like (10b). Furthermore, this transfer illustrates the typical transition from concrete, less abstract meanings to more abstract meanings. In (10a) keep refers to visible, tangible complements like money, whereas in (10b) complements like complaining are more abstract in that they cannot be touched etc. (Heine 1997:7)

4.2.2. Ambiguity

Grammaticalization chains are regarded as “continua without any clear-cut internal boundaries” and accordingly grammaticalization is defined as “an evolutional continuum” that cannot be segmented into discrete units. (Heine et al. 1991a:16, Heine et al. 1991b:163) Ewe megbé provides evidence for the overlapping model as it denotes a semantic ambiguity and refers at the same time to more than one of these four categories, e.g. sentence (11) has either a spatial or a temporal meaning:

(11) é le megbé ná-m
3SG be behind PREP-1SG
(a) ‘He is behind me (spatially)’
(b) ‘He is late (= he could not keep pace with me)’

The life span of such a so-called hybrid form can be short, but may also extend over hundreds of years. The grammaticalization of the German demonstrative der to a relative pronoun started in Old High German ⁸ and by the 16th century, it was still a hybrid form. In modern Standard German, the split between demonstrative and relative pronoun is clear, but the hybrid still survives in certain contexts, as in (12): (Heine et al. 1991a:232-33)

(12) der mir gefällt ist zu groß DEM/REL to.me pleases is too big‘The one that appeals to me is too big’

4.2.3. Asymmetry

As Ewe megbé in (11) and German der in (12) show, a linguistic item does not always develop symmetrically regarding its meaning and morphosyntax. Asymmetry between cognitive and linguistic structure is another characteristic of grammaticalization chains. (Heine et al. 1991b:178) In (11) the meaning of megbé has shifted from body part noun to a spatio-temporal entity but the morphology is still that of the noun, i.e. the morphosyntax was not yet affected by the conceptual shift. As Givón (1975:86) points it out, “morphological and syntactic behaviour is likely to lag behind the more progressive semantic reanalysis, and thus quite often represent vacuous relics of the older semantic situation.” This asymmetrical condition raised the question of whether two instances of the same form are cases of polysemy or homonymy. There is still disagreement among scholars. Following Persson (1988), Lichtenberk (1991b:476) introduces “heterosemy” as the appropriate term to refer to cases in which two or more meanings or functions are derived from the same ultimate source: “There is heterosemy if a verb, a directional particle, and an aspect marker all ultimately descend from the same historical source.”

Asymmetrical cases challenge “one function - one form” principles of linguists who are reluctant to recognize the possibility of multiple meanings. However, any serious approach to language has to recognize the possible coexistence of the old and the new meaning within the synchronous system. (Kinberg 1991:324-26)

4.2.4. Unidirectionality

Grammaticalization processes are seen as being basically unidirectional, that is, irreversible. Grammaticalization can be compared to biological processes such as growth, maturation and ageing, assuming that growing up implies becoming taller and it is not to be expected that a child starts becoming shorter and shorter and finally returns to the womb. Similarly, it is very unlikely that the French future tense endings would separate from the verb, become auxiliaries and then would end up as full verbs meaning “to keep”, which is the presumed etymology of Latin habere, the source of the French future. In the following section, the functional processes and their concomitant effects involved in the evolution of a grammatical item are discussed in detail. Like links in a chain, these processes provide an overlapping structure similar to a grammaticalization chain, because they are often inseparable from each other and go hand in hand.

4.2.5. Generalization

Morphemes of increased generality are most likely to grammaticalize. Once they start down the road of grammaticalization, these items become more general throughout the process of change. Grammaticalization typically involves generalization, that is, an increase in the polysemies of a form. As grammaticalization progresses, meanings expand their range through the development of various polysemies but the reversal, i.e. narrowing of meaning, is rare in grammaticalization. ⁹ Thus, grammatical items are characteristically polysemous. For example, English –s inflections denote nominal plural and third-person-singular verbal marker, or –d inflections are used in past tense and past participle constructions. (Hopper and Traugott [1993]2003:100-03)

A grammatical morpheme is generalized when it loses components of its original meaning and becomes more general and abstract. For instance, will lost its volitional aspect and be going to lost its spatial movement component when they were grammaticalized to future markers. (Bybee 2002:156) Concerning be going to, the restriction that the subject moves in space toward a goal is no longer existent, that is, it has been lost and the meaning became more general because the subject can in any sense (spatial or not) be on a course toward a particular endpoint in the future. (Bybee et al. 1994:5) This process is also called bleaching, semantic reduction, semantic attrition or desemanticization. From a diachronic perspective, it constitutes the gradual decrease in semanticity by the loss of certain properties. Synchronically, it is a specific kind of polysemy. (Lehmann 1995:127)

Sometimes a linguistic item loses its original meaning through the process of generalization. For example, Latin adripare “to reach the shore” gave French arriver and Italian arrivare, first with a general meaning “to arrive” and later with an even more general meaning “to happen”. (Vincent 2001:20)

4.2.6. Frequency

When a morpheme is generalized through grammaticalization, their frequency increases because they become appropriate in a growing range of contexts. The original meaning of be going to was the only possible interpretation in Shakespeare’s time (13), but its use was expanded throughout the centuries (14-15) and thus it occurred more frequently. (Bybee 2002:146)

(13) movement: We are going to Windsor to see the King
(14) intention: We are going to get married in June
(15) future: These trees are going to lose their leaves

When be going to had its literal meaning of a subject moving to a location, the subject position could be filled only by a noun phrase denoting an animate, mobile entity. The verb following the phrase had to be a dynamic verb. Then the phrase grammaticalized and “the number of different types appropriate for subject position expanded to include non-animate and non-mobile entities and the verb position expanded to include a broader range of predicates”. (Bybee 2003:605)

As grammatical items are habitualized and become more frequent, they are automatized: “Grammaticalization is the process of automatization of frequently occurring sequences of linguistic elements”, as Bybee (2002:152) puts it. This process has two consequences. Firstly, the identity of the component units is gradually lost, and secondly, the whole chunk is reduced in form both morphologically and phonologically. (Bybee 2002:153) Thus, going to is repackaged as gonna. Other examples include gotta (got to), wanna (want to), hafta (have to), [spostə] (supposed to), [yustə] (used to) and I know watcha thinking, gotcha (got you),[aI kN] (I can); among the most familiar contracted forms are I’ll, I’ve, I’m, won’t, you’re etc. Contracted forms usually have a higher frequency in speech and are the more casual register forms. (Bybee 2003:615, Hopper and Traugott [1993]2003:128) The familiarity effect is crucial for the spread of a sound change, since “the changes in question occur more often in casual speech, words that are used more often in casual speech will be more often subjected to the change”. (Bybee 2003:616)

[...]

---

¹ In his article “Mutations of Linguistic Categories” (1968), Benveniste, who was a student of Meillet, repeated much of Meillet’s account of the evolution of auxiliary verbs out of lexical verbs such as ‘have’. He even coined the word ‘auxiliation’. However, neither did he use the term ‘grammaticalization’ nor did he refer at any point in his paper to Meillet’s work. (Hopper and Traugott [1993]2003:26)

² Members of the Pidgin community come from different cultural communities, thus they share little pragmatic prepsuppositional background. The same applies for children and adults, as the child lacks the shared background of knowing of the adult regarding the world. Time pressure arises for both Pidgin community and the child because urgent functions and tasks need to be performed but they have no shared mode of communication with their surrounding human community. (Givón 1979:225-27)

³ It was published as a working paper in 1982 and revised as a monograph in 1995.

⁴ Grammaticalization: A Critical Assessment. Language Sciences 23, 2001.

⁵ Recent work on various sign languages has concentrated on grammaticalization changes as well and typical pathways of change have been discovered, e.g. in American Sign Language the verb finish has evolved into an aspectual marker for perfective action. This development is attested in many languages all over the world. ( Meir 2003:110)

⁶ http://dictionary.reference.com/search?q=get

⁷ Back in 1871, Darwin recognized that “the formation of different languages and of distinct species, and the proofs that both have been developed through a gradual process, are curiously parallel”.

⁸ c. 800 – 1100.

⁹ A change in the opposite direction was Latin trahere “to drag” that gave French traire “to milk” and Italian trarre “to pull”. The original general meaning of present-day English lust “sexual desire” is only retained in German Lust “delight or desire of any kind”. (Vincent 2001:21)