Current research on grammaticalization argues that grammatical markers are generally derived from content words (or lexical expressions); but earlier research by Brugmann (1904) and Bühler (1934) showed that grammatical markers are also commonly derived from (spatial) deictics (or demonstratives). The present paper provides an overview of this research focusing on Bühler’s two-field theory of pointing and naming. In this theory, there are two basic types of linguistic expressions, deictics (or ‘pointing words’) and symbols (or ‘naming words’), that are functionally and diachronically independent of each other. The paper argues that Bühler’s two-field theory can be seen as an alternative to the standard model of grammaticalization in which all grammatical markers are ultimately based on content words. Elaborating this approach, it is argued that the grammaticalization of deictic expressions involves a different mechanism of change than the grammaticalization of content words and that the two developments give rise to different types of grammatical markers.
In this contribution it is argued that grammaticalization and related types of change arise as a side-effect of strategic, rhetorical language use by speakers. The outcome of these processes is determined by the strategy for which the underlying lexical items are used, rather than by the lexical content of these items or by pre-set characteristics of types of change. This is illustrated with the three grammatical uses that derived from lexical bien in French: modal particle, concessive conjunction and discourse particle. This approach to grammaticalization and other changes is also shown to shed new light on high-level generalizations such as persistence, subjectification and pragmaticalization, which need to be complemented with lower-level generalizations about the rhetorical strategies that are driving them.
Grammaticalization, and especially degrammaticalization, are polysemous terms. The term ‘grammaticalization’ has been used to refer to changes from a lexical item to a grammatical item (‘primary grammaticalization’), and from a grammatical item to a ‘more grammatical’ item (‘secondary grammaticalization’). Likewise, the term ‘degrammaticalization’ may refer to changes from a grammatical item to a lexical item (‘primary degrammaticalization’), and from a ‘more grammatical’ to a ‘less grammatical’ item (‘secondary degrammaticalization’), as well as to a number of other types of changes. This paper discusses Lehmann’s parameters of grammaticalization as a taxonomic tool, arguing that the parameters can be used to describe both grammaticalization and degrammaticalization changes, as well as to identify different subtypes of grammaticalization and degrammaticalization, with special emphasis on degrammaticalization.
In this paper, we propose a ‘constructionist’ scenario of grammaticalization that links successive diachronic stages of a grammaticalization process with particular construction types. Relying on earlier work, especially on the model of relevant context types in grammaticalization proposed in Diewald (2002), we develop the model further and introduce an additional developmental stage into the model, i.e. the stage of paradigmatic (re-)integration. This fourth stage restricts the model to grammaticalization processes and sharpens the notion of grammaticalization. In particular, we advocate the view that the fourth stage can and must be formulated as an essential criterion for differentiating grammaticalization from lexicalization and from other types of semantic change.
“Ghost” inflectional morphology that has lost its grammatical function but remains as phonetic material has been argued to have undergone lexicalization (since the inflection becomes an unanalyzable part of the lexical item and emerges as “more lexical”) and/or degrammaticalization (since the inflection loses grammatical function and is hence “less grammatical”); if seen as the natural consequence of an inflection having attained advanced grammatical status, it may also be understood as degrammaticalization. Focusing on comparative -er (near), superlative -est (next), adverbial genitive -s (e.g. once, towards, sideways), and adverbial dative -um (whilom), this paper distinguishes between changes affecting different parts of a construction (i.e. the host words and the inflectional endings) and argues that the inflections are subject to neither lexicalization nor (de)grammaticalization, but are instances of “petrification”.
Research within constructional approaches to language has begun to develop a framework for modelling diachronic variation, but the precise relationship between diachronic construction grammar and grammaticalization has yet to be fully worked out (Noël 2007). The article includes discussion of some central issues in construction grammar which are of significant relevance to researchers interested in grammaticalization and lexicalization, with suggestions for some of the ways in which key topics in grammaticalization research – such as unidirectionality, reanalysis and analogy – might be explained using a constructional model. In order to illustrate this thesis, I draw on a range of data from the history of English (particularly, quantifiers and degree modifiers, composite predicates and possessives).
In this article we propose that the notion of grammaticalization as a gradual process can be extended to language typology. Within the same language family, grammaticalization phenomena can be ongoing in one language and have reached a stage further down the cline in another language. This entails that diachronically languages of the same family may grammaticalize at a different speed. ‘Degree of grammaticalization’ is operationalized as degree of paradigmatization: the reduction of alternatives is viewed as representing a higher degree of paradigmaticity and of grammaticalization. We make our case for three important Romance languages, French, Italian and Spanish. It is argued that in Romance we find the following cline of grammaticalization: French > Italian > Spanish, i.e. with French being most grammaticalized, Spanish least, and Italian in between.
This paper offers a qualitative and quantitative study of the development of periphrastic tense and aspect constructions in the Irish language. It is argued that the modern periphrastic perfect construction arose to fill a gap left in the tense marking system by the generalization of the former perfect to preterite senses. It is suggested that this development was facilitated by the evolution of a periphrastic, prepositional progressive, which grammaticalized to specify the aspectual differences in continuous aspect marking. A similar development is envisaged for the comparable, but not identical, Welsh aspectual system.
Historical linguistics has begun to explore the way new insights into change can be gained from a constructional approach to language. In this work we focus on a Spanish construction used for events caused accidentally by a human participant marked dative. The construction enters fairly late in the language and thus raises the question of how the new pairing of form and meaning emerged. Under our proposal, the development originates in an old voice pattern expressing spontaneously occurring events and leads to the new construction through a series of gradual extensions to distinct event types. Most significantly, as a result of these extensions the input construction becomes more schematic, which is to say, it grammaticalizes (Trousdale this volume).
The paper provides a functional, usage-based analysis of the genesis and diffusion of object clitic doubling in Spanish, which is seen as an agreement phenomenon, and hence as a working example of grammaticalization. It has been claimed that doubling arises from the so-called topic-shift construction (Givón 1976), but historical and contemporary data support an alternative proposal that pays more attention to frequency effects and takes into account the often neglected high proportion of clitic-only objects in discourse. Our analysis suggests a strong correlation between being an object encoded frequently by means of a clitic/affix and developing agreement It also points to the accessibility of the discourse referents as the main factor in determining the grammatical form of direct and indirect objects in Spanish.
This study focuses on “the many careers of negative polarity items”, taking a diachronic perspective on NPIs in general and on scalar NPIs in particular. Its main thesis is that scalar NPIs are prototypical NPIs. The downward entailing contexts of NPIs can be explained and made cognitively accessible by the pragmatic mechanisms associated with scalar NPIs, viz. the capacity to evoke alternatives (ALT) and the scalar interpretation of these alternatives (SCALE). NPIs with standard contexts of distribution are, or are otherwise tied to, scalar expressions, while NPIs with an idiosyncratic range of contexts are not. The diachronic development of core NPIs crucially involves the loss, change or replacement of ALT and/or SCALE. Non-scalar elements can also become NPIs, but non-prototypical ones, viz. escort particles, analogy NPIs, or mimicry NPIs.
Current research on grammaticalization argues that grammatical markers are generally derived from content words (or lexical expressions); but earlier research by Brugmann (1904) and Bühler (1934) showed that grammatical markers are also commonly derived from (spatial) deictics (or demonstratives). The present paper provides an overview of this research focusing on Bühler’s two-field theory of pointing and naming. In this theory, there are two basic types of linguistic expressions, deictics (or ‘pointing words’) and symbols (or ‘naming words’), that are functionally and diachronically independent of each other. The paper argues that Bühler’s two-field theory can be seen as an alternative to the standard model of grammaticalization in which all grammatical markers are ultimately based on content words. Elaborating this approach, it is argued that the grammaticalization of deictic expressions involves a different mechanism of change than the grammaticalization of content words and that the two developments give rise to different types of grammatical markers.
In this contribution it is argued that grammaticalization and related types of change arise as a side-effect of strategic, rhetorical language use by speakers. The outcome of these processes is determined by the strategy for which the underlying lexical items are used, rather than by the lexical content of these items or by pre-set characteristics of types of change. This is illustrated with the three grammatical uses that derived from lexical bien in French: modal particle, concessive conjunction and discourse particle. This approach to grammaticalization and other changes is also shown to shed new light on high-level generalizations such as persistence, subjectification and pragmaticalization, which need to be complemented with lower-level generalizations about the rhetorical strategies that are driving them.
Grammaticalization, and especially degrammaticalization, are polysemous terms. The term ‘grammaticalization’ has been used to refer to changes from a lexical item to a grammatical item (‘primary grammaticalization’), and from a grammatical item to a ‘more grammatical’ item (‘secondary grammaticalization’). Likewise, the term ‘degrammaticalization’ may refer to changes from a grammatical item to a lexical item (‘primary degrammaticalization’), and from a ‘more grammatical’ to a ‘less grammatical’ item (‘secondary degrammaticalization’), as well as to a number of other types of changes. This paper discusses Lehmann’s parameters of grammaticalization as a taxonomic tool, arguing that the parameters can be used to describe both grammaticalization and degrammaticalization changes, as well as to identify different subtypes of grammaticalization and degrammaticalization, with special emphasis on degrammaticalization.
In this paper, we propose a ‘constructionist’ scenario of grammaticalization that links successive diachronic stages of a grammaticalization process with particular construction types. Relying on earlier work, especially on the model of relevant context types in grammaticalization proposed in Diewald (2002), we develop the model further and introduce an additional developmental stage into the model, i.e. the stage of paradigmatic (re-)integration. This fourth stage restricts the model to grammaticalization processes and sharpens the notion of grammaticalization. In particular, we advocate the view that the fourth stage can and must be formulated as an essential criterion for differentiating grammaticalization from lexicalization and from other types of semantic change.
“Ghost” inflectional morphology that has lost its grammatical function but remains as phonetic material has been argued to have undergone lexicalization (since the inflection becomes an unanalyzable part of the lexical item and emerges as “more lexical”) and/or degrammaticalization (since the inflection loses grammatical function and is hence “less grammatical”); if seen as the natural consequence of an inflection having attained advanced grammatical status, it may also be understood as degrammaticalization. Focusing on comparative -er (near), superlative -est (next), adverbial genitive -s (e.g. once, towards, sideways), and adverbial dative -um (whilom), this paper distinguishes between changes affecting different parts of a construction (i.e. the host words and the inflectional endings) and argues that the inflections are subject to neither lexicalization nor (de)grammaticalization, but are instances of “petrification”.
Research within constructional approaches to language has begun to develop a framework for modelling diachronic variation, but the precise relationship between diachronic construction grammar and grammaticalization has yet to be fully worked out (Noël 2007). The article includes discussion of some central issues in construction grammar which are of significant relevance to researchers interested in grammaticalization and lexicalization, with suggestions for some of the ways in which key topics in grammaticalization research – such as unidirectionality, reanalysis and analogy – might be explained using a constructional model. In order to illustrate this thesis, I draw on a range of data from the history of English (particularly, quantifiers and degree modifiers, composite predicates and possessives).
In this article we propose that the notion of grammaticalization as a gradual process can be extended to language typology. Within the same language family, grammaticalization phenomena can be ongoing in one language and have reached a stage further down the cline in another language. This entails that diachronically languages of the same family may grammaticalize at a different speed. ‘Degree of grammaticalization’ is operationalized as degree of paradigmatization: the reduction of alternatives is viewed as representing a higher degree of paradigmaticity and of grammaticalization. We make our case for three important Romance languages, French, Italian and Spanish. It is argued that in Romance we find the following cline of grammaticalization: French > Italian > Spanish, i.e. with French being most grammaticalized, Spanish least, and Italian in between.
This paper offers a qualitative and quantitative study of the development of periphrastic tense and aspect constructions in the Irish language. It is argued that the modern periphrastic perfect construction arose to fill a gap left in the tense marking system by the generalization of the former perfect to preterite senses. It is suggested that this development was facilitated by the evolution of a periphrastic, prepositional progressive, which grammaticalized to specify the aspectual differences in continuous aspect marking. A similar development is envisaged for the comparable, but not identical, Welsh aspectual system.
Historical linguistics has begun to explore the way new insights into change can be gained from a constructional approach to language. In this work we focus on a Spanish construction used for events caused accidentally by a human participant marked dative. The construction enters fairly late in the language and thus raises the question of how the new pairing of form and meaning emerged. Under our proposal, the development originates in an old voice pattern expressing spontaneously occurring events and leads to the new construction through a series of gradual extensions to distinct event types. Most significantly, as a result of these extensions the input construction becomes more schematic, which is to say, it grammaticalizes (Trousdale this volume).
The paper provides a functional, usage-based analysis of the genesis and diffusion of object clitic doubling in Spanish, which is seen as an agreement phenomenon, and hence as a working example of grammaticalization. It has been claimed that doubling arises from the so-called topic-shift construction (Givón 1976), but historical and contemporary data support an alternative proposal that pays more attention to frequency effects and takes into account the often neglected high proportion of clitic-only objects in discourse. Our analysis suggests a strong correlation between being an object encoded frequently by means of a clitic/affix and developing agreement It also points to the accessibility of the discourse referents as the main factor in determining the grammatical form of direct and indirect objects in Spanish.
This study focuses on “the many careers of negative polarity items”, taking a diachronic perspective on NPIs in general and on scalar NPIs in particular. Its main thesis is that scalar NPIs are prototypical NPIs. The downward entailing contexts of NPIs can be explained and made cognitively accessible by the pragmatic mechanisms associated with scalar NPIs, viz. the capacity to evoke alternatives (ALT) and the scalar interpretation of these alternatives (SCALE). NPIs with standard contexts of distribution are, or are otherwise tied to, scalar expressions, while NPIs with an idiosyncratic range of contexts are not. The diachronic development of core NPIs crucially involves the loss, change or replacement of ALT and/or SCALE. Non-scalar elements can also become NPIs, but non-prototypical ones, viz. escort particles, analogy NPIs, or mimicry NPIs.