Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, LINGUISTICS ( (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 22 May 2017


Summary and Keywords

Linguistic change not only affects the lexicon and the phonology of words, it also operates on the grammar of a language. In this context, grammaticalization is concerned with the development of lexical items into markers of grammatical categories or, more generally, with the development of markers used for procedural cueing of abstract relationships out of linguistic items with concrete referential meaning. A well-known example is the English verb go in its function of a future marker, as in She is going to visit her friend. Phenomena like these are very frequent across the world’s languages and across many different domains of grammatical categories. In the last 50 years, research on grammaticalization has come up with a plethora of (a) generalizations, (b) models of how grammaticalization works, and (c) methodological refinements.

On (a): Processes of grammaticalization develop gradually, step by step, and the sequence of the individual stages follows certain clines as they have been generalized from cross-linguistic comparison (unidirectionality). Even though there are counterexamples that go against the directionality of various clines, their number seems smaller than assumed in the late 1990s.

On (b): Models or scenarios of grammaticalization integrate various factors. Depending on the theoretical background, grammaticalization and its results are motivated either by the competing motivations of economy vs. iconicity/explicitness in functional typology or by a change from movement to merger in the minimalist program. Pragmatic inference is of central importance for initiating processes of grammaticalization (and maybe also at later stages), and it activates mechanisms like reanalysis and analogy, whose status is controversial in the literature. Finally, grammaticalization does not only work within individual languages/varieties, it also operates across languages. In situations of contact, the existence of a certain grammatical category may induce grammaticalization in another language.

On (c): Even though it is hard to measure degrees of grammaticalization in terms of absolute and exact figures, it is possible to determine relative degrees of grammaticalization in terms of the autonomy of linguistic signs. Moreover, more recent research has come up with criteria for distinguishing grammaticalization and lexicalization (defined as the loss of productivity, transparency, and/or compositionality of former productive, transparent, and compositional structures).

In spite of these findings, there are still quite a number of questions that need further research. Two questions to be discussed address basic issues concerning the overall properties of grammaticalization. (1) What is the relation between constructions and grammaticalization? In the more traditional view, constructions are seen as the syntactic framework within which linguistic items are grammaticalized. In more recent approaches based on construction grammar, constructions are defined as combinations of form and meaning. Thus, grammaticalization can be seen in the light of constructionalization, i.e., the creation of new combinations of form and meaning. Even though constructionalization covers many apects of grammaticalization, it does not exhaustively cover the domain of grammaticalization. (2) Is grammaticalization cross-linguistically homogeneous, or is there a certain range of variation? There is evidence from East and mainland Southeast Asia that there is cross-linguistic variation to some extent.

Keywords: analogy, constructionalization, economy, gradualness, iconicity, language contact, pragmatic inference, reanalysis, schematization, subjectification

1 Grammaticalization—the Basics

1.1 Grammaticalization—a First Glance

Many grammatical categories are expressed by markers which are quite obviously related to lexical items. A good example is the English future marker will, which goes back to the Old English verb willan ‘want.’ Similarly, the verb go marks future in a sentence like She is going to visit her friend. In other examples, this time from the domain of the noun, we find body-part terms like back for expressing spatial orientation as in She stands at the back of the church.

Such phenomena are not limited to English. In fact, they are extremely widespread across the world’s languages. In modern Greek, the future marker Ɵa is derived from the old Greek verb Ɵέ‎lein ‘want’. The Chinese future marker yào is identical to the verb yào ‘want,’ and in Swahili, the prefix -ta- in a future form like ni-ta-nunua [1SG-FUT-buy] ‘I will buy’ corresponds to a reduced form of the verb root taka ‘want’. Verbs with the meaning of ‘go’ recurrently mark the future not only in other Indo-European languages like French (1) or Spanish but also in a large number of other languages as for example in Igbo (2).





Similarly, body-part terms with the meaning of ‘stomach’, ‘back’, ‘forehead/front’, ‘face’, ‘head’, etc. are frequently used in the context of spatial orientation. The following example from Bambara illustrates the use of kɔ́nɔ ‘stomach’:



Examples of this type are well known and they have been described in historical linguistics since the 19th century. In spite of this, it was only at the beginning of the 20th century that the French linguist Meillet (1912) coined the term ‘grammaticalization’ (in French: grammaticalisation) for processes of innovation in which autonomous words become grammatical agents (in French: agents grammaticaux). But even the existence of the term ‘grammaticalization’ did not have any immediate impact on the research agenda of linguistics at that time. This only happened about half a century later when Kuryłowicz (1965) introduced his famous definition:

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 . . .

(Kuryłowicz, 1965 [1975, p. 52])

Since then, grammaticalization (sometimes also called ‘grammaticization’) has become a prominent topic in linguistics that has led to an extensive body of literature that reaches from detailed descriptions of individual languages up to broad typological generalizations. As was the case with many other fields of linguistic research, grammaticalization has also been the subject of controversial discussions that even went as far as to its complete deconstruction by Newmeyer (1998) (see Section 4; see also Campbell, 2001). The Oxford Handbook of Grammaticalization (Narrog & Heine, 2011) provides a comprehensive survey of the different approaches and the findings of the field. In another survey, the World Lexicon of Grammaticalization (Heine & Kuteva, 2002), a large number of lexical items are listed together with the different grammatical categories they can express cross-linguistically.

In the past two decades, research on grammaticalization has profited a lot from corpus-based studies and from statistical microanalysis of results of change (see, e.g., Krug, 2000; Hoffmann, 2005, on English, Post, 2007, on Thai and Chinese, and edited volumes like Lindquist & Mair, 2004). Hilpert (2008) showed with his data-driven collocational analysis based on Gries and Stefanowitsch (2004) how linguistic items collocate differently in processes of grammaticalization. While English be going to collocates typically with transitive and agentive verbs on its way to a future marker, the Dutch verb gaan ‘go’ starts out from typical collocations with nonpunctual intransitive verbs. Research like this reveals that what looks similar on the surface as we observe it today may come about in processes of grammaticalization that vary cross-linguistically. The interesting question here is to what extent variation is possible.

1.2 Clines and Universal Properties of Grammaticalization

It is quite a remarkable linguistic fact that certain lexical items with similar meaning develop into markers of similar grammatical categories around the globe (e.g., ‘want’, ‘go’, ‘body part’ as described in section Grammaticalization—a First Glance). But this is not the whole story. Linguists working on grammaticalization argued for even more powerful generalizations concerning individual stages involved in processes of grammaticalization and their diachronic sequence. In this view, processes of grammaticalization develop along certain clines from more lexical to more grammatical properties. These clines have the structure of a continuum with its focal points that are arranged between an initial and a terminal pole. Processes of grammaticalization evolve along the individual positions of that continuum. The term ‘cline’ is not the only term for this type of diachronic generalization. ‘Pathway’ is another common term. Heine, Claudi, and Hünnemeyer (1991) use the notions of ‘channel’ and ‘chain’ in their approach. No matter what term is used, it is important to notice that clines are the result of cross-linguistic extrapolations and should not be confounded with neurological pathways.

The notion of a cline is already adumbrated in the above quotation from Kuryłowicz (1965 [1975, p. 52]) when he talks about developments ‘from a less grammatical to a more grammatical status’. Since his work, a lot of clines have been claimed in the literature. Givón (1971, 1979) introduced a very prominent cline which describes the different grammatical levels through which processes of grammaticalization develop. Starting out from discourse, what used to be a topic structure at an earlier stage may develop into a syntactic structure with the topic becoming a subject at a later stage. From there, it may further grammaticalize into a morphological pattern (agreement marker) that eventually gets further reduced into a morphophonemic structure that may ultimately end in a zero-marker:



Another cline is based on nouns that belong to the semantic domain of person (e.g., body parts) (Heine et al., 1991). At a next stage, the use of these lexical items extends to objects in general, among them non-human/inanimate objects as in John’s back > the back of the house, and from there to further stages defined as activity, space, time and quality.



Part of this cline is illustrated by the followng example from Ewe, in which the body-part term megbé ‘back’ is used as a noun in (6a) and as a locative marker expressing space in (6b) and time in (6c):



At the last stage of quality, body-part terms get a specific metaphorical meaning, which is lexical in most languages. A frequent metaphor moves from situating an event in the timescale to judgments concerning the degree of progress (advanced vs. old-fashioned). An example of this development is English back, which has a spatial meaning (to move back) and a temporal meaning ‘back in time’ (three years back) but also occurs in words like backward, which can have the metaphorical meaning of ‘underdeveloped, dimwitted’ as in the technology is backward but the design is beautiful.

Clines like the ones in (4) and (5) have a number of general properties which are claimed to be characteristic of grammaticalization (see also Bisang, 2016):

  1. (i) Cyclicity: They are realized in stages, step-by-step.

  2. (ii) Universality: The sequence of the individual stages follows universal clines.

  3. (iii) Unidirectionality: The grammaticalization clines are not reversible. Thus, morphology > syntax in (4) or time > space in (5) are predicted to be impossible.

  4. (iv) Source determination: The clines start out from a source concept that determines their further development (e.g., body-part terms, verbs with the meaning of ‘want’, ‘go’, etc.; see Heine & Kuteva, 2002).

  5. (v) Retention of earlier meaning: If a pathway develops from meaning A to B, there is an intermediate stage in which a linguistic item can have both meanings: A > {B/A} > B. Thus, semantic nuances of the source concept or the source construction are retained at later stages. This fact is sometimes also observed with processes of lexicalization. Thus the status of this criterion as an exclusive property of grammaticalization needs further empirical checking.

  6. (vi) As a consequence of (v), processes of grammaticalization are gradual and develop over a certain period of time with different nuances between A and B (Traugott & Trousdale, 2010).

  7. (vii) Co-evolution of meaning and form: The change from a more concrete to a more abstract grammatical meaning is reflected in the form of the linguistic item involved. Thus, increasing grammatical meaning goes with a corresponding loss in formal substance. As will be discussed in section “Areal Differences and Grammaticalization”, the extent to which this is true is questionable to a certain extent.

1.3 Motivations of Grammaticalization—Functional and Generative Approaches

What drives grammaticalization? There are two basic motivations discussed in the literature. One of them is the competition between the two motivations of economy and iconicity, the other one is based on Universal Grammar.

Linguistic utterances are generally determined by two competing motivations operating in a speech-act situation between speaker and hearer. On the one hand, they have to be explicit and distinctive enough for avoiding misunderstandings. On the other hand, they should be economic in the sense that a certain concept is expressed with minimal effort. Haiman (1983) discusses these two motivations in terms of economy vs. iconicity. While economy has to do with choosing the simplest way to express a given concept, iconicity is based on a certain isomorphism between a concept and the way in which it is formally expressed. These two forces together are responsible for the co-evolution of meaning and form (see Section 1.2, point (vii), and Section 6). If one assumes that grammaticalization develops from more concrete to more abstract meaning, this reduction of meaning is iconically reflected in the reduction of linguistic form. This process is further enhanced by economy and its general preference of minimizing articulatory effort on the phonological side. Moreover, higher levels of semantic abstraction come with less notional distinctions. Therefore, the formal cues for abstract grammatical categories can be more reduced than for more concrete meanings.

Explanations based on Universal Grammar focus on unidirectionality (see Section 1.2, point (iii); Roberts & Roussou, 2003; van Gelderen, 2004). In the minimalist approach (Chomsky, 1995), heads with a grammatical function generally take higher positions in the tree structure than lexical items. Thus, unidirectionality is motivated by the fact that lexical items that take on a grammatical function are raised to higher positions. At an early stage in one generation of speakers, a lexical item that gets interpreted as a grammatical marker moves from its original lower position to the higher position associated with its new function. If that item additionally loses some of its formal substance (e.g., erosion of phonological substance), it will be taken as a cue for a change in parameter setting in the linguistic input to the next generation of speakers. If that happens, the linguistic item in question will no longer have to move from its lower position to the higher position which corresponds to its more grammatical function—it will be generated directly at that higher position. Thus, grammaticalization is explained as a change from movement to merger in the minimalist framework.

1.4 Grammaticalization vs. Lexicalization

The term lexicalization is used in rather divergent ways in linguistics. In the context of how to distinguish lexicalization from grammaticalization, it is conceived as a process in which ‘an originally productive, transparent, compositional formation loses its productivity, transparency and/or compositionality’ (Himmelmann, 2004, p. 28). Brinton and Traugott (2005, p. 96) define this process as follows:

Lexicalization is the change whereby in certain linguistic contexts speakers use a syntactic construction or word formation as a new contentful form with formal and semantic properties that are not completely derivable or predictable from the constituents of the construction or the word formation pattern. Over time there may be further loss of internal constituency and the item may become more lexical.

Grammaticalization and lexicalization share important properties. In both cases, constructions and their components are subject to new morphosyntactic analyses, to the conventionalization of new meanings, and, eventually, to the reduction of form in terms of cline (4). Two prominent examples of lexicalization are fossilization and univerbation. Fossilization reduces morphologically complex forms to unanalizable wholes, i.e., formerly productive formatives are reanalyzed as parts of the root. A good example is the English verb set, which is the causative form of sit and is historically derived from the causative suffix *-eja. In processes of univerbation, frequent collocations of independent lexemes are fused into single words as in the case of blackboard, cupboard, airship, or steamboat.

For a long time there were no good criteria for distinguishing the two processes. Two more recent proposals by Himmelmann (2004) and Boye and Harder (2012) have significantly changed that situation.

Himmelmann’s (2004) proposal focuses on host-class expansion, i.e., the number of the classes of elements a marker is in construction with. He mentions the example of articles that developed out of demonstratives. Articles expand their host-class when they occur with classes of nouns demonstratives do not combine with regularly (e.g., proper names, nouns designating unique concepts; Himmelmann, 2004, pp. 32–33). While grammaticalization is characterized by the increasing number of classes of elements a lexical item can be combined with, lexicalization is characterized by host-class reduction. Thus, the adjective black can be combined with a large number of nouns in syntax (e.g., black animal, black house, black car, black shoes) but it can construct only with one other lexeme in the case of univerbation to produce a new lexical item with its specific meaning (blackboard or other specific concepts like blackbird or blackjack). Similarly, the class of linguistic items a marker can be combined with decreases significantly in processes of fossilization until it even loses its identifiability and becomes part of the root. What happens in both cases of lexicalization is ‘essentially confined to the two elements entering into the special collocation’ (Himmelmann, 2004, p. 35).

The proposal of Boye and Harder (2012) is based on the ability of a linguistic item to be used in discourse structures. Highly grammaticalized items cannot occur in focus positions, nor can they be addressed independently by wh-words in English. In contrast, the products of lexicalization can take on both functions.

2 Grammaticalization and Its Measuring—Autonomy and the Co-Evolution of Meaning and Form

The claim that grammaticalization is gradual and leads from lexical items to markers increasingly associated with grammatical categories implies that grammaticalization can be measured or that it is at least possible to determine relative degrees of grammaticalization within a given set of linguistic units. Autonomy as discussed by Lehmann offers a good criterion:

The autonomy of a sign is converse to its grammaticality, and grammaticalization detracts from its autonomy. Consequently, if we want to measure the degree to which a sign is grammaticalized, we will determine its degree of autonomy.

(1995, pp. 121–122)

The autonomy of a sign is determined by the three parameters of weight, cohesion and variability, which are defined as follows:


A linguistic sign needs a certain weight in order to be distinguished from other elements of its class—a weight which provides it with a certain prominence in the syntagm. Reduction of weight implies increase of grammaticalization.


A linguistic sign has a certain rigor or intensity with which it can contract systematic relations with other signs. The higher that rigor or intensity is, the less autonomous and the more grammaticalized it is.


A linguistic sign has a certain degree of mobility. The more restricted that mobility is, the less autonomous and the more grammaticalized it is.

Each of these parameters has a paradigmatic and a syntagmatic dimension. Thus, Lehmann’s (1995) approach is finally based on the following six parameters for measuring grammaticalization:


Paradigmatic weight (= Integrity)

Syntagmatic weight (= Structural scope)


Paradigmatic cohesion (= Paradigmaticity)

Syntagmatic cohesion (= Bondedness)


Paradigmatic variability

Syntagmatic variability

Lehmann (1995) does not use autonomy for measuring grammaticalization in terms of absolute and exact figures. In spite of this, autonomy offers good criteria for determining relative degrees of grammaticalization. This will be briefly illustrated for the parameters of integrity (paradigmatic weight), paradigmatic variability, and syntagmatic variability.

For a linguistic sign to be autonomous it needs a certain integrity or substance to be distinguishable from other signs. This substance is either of phonological or of semantic nature. On the phonological side, the loss of integrity is associated with reduction of phonological substance (attrition; see the Swahili verb taka ‘want’ and its morphological form -ta- in the function of a future marker). As is shown from more recent phonetic/phonological research, the universal status of attrition in processes of grammaticalization is questionable. Schiering (2005) shows in his cross-linguistic study that attrition strongly covaries with stress-based phonologies and is strongly limited in syllable-based phonologies and mora-based phonologies. Wichmann (2008, 2011) points out that loss of prosodic prominence and loss of independence in intonational structure is at least as important as attrition. On the semantic side, loss of integrity comes with desemanticization, i.e., the loss of concrete meaning (e.g., ‘want’) to the benefit of abstract grammatical meaning (e.g., ‘future tense’).

Paradigmatic variability is about ‘the freedom with which the language user chooses a sign’ (Lehmann, 1995, p. 137). The grammar of a language may force its speakers to express a certain value of a grammatical category, i.e., speakers may have to select either a past or a non-past marker in a given language if that language has a binary tense system of the type [±past]. In other languages, no such information needs to be provided even if the language has grammatical tense. Thus, a high degree of grammaticalization is typically associated with obligatoriness in the sense that a speaker has to specify a certain grammatical category in a given grammatical environment with one of the values present in the grammatical system of a language (but see Section 6 for a critical view on that).

Syntagmatic variability is associated with flexibility in word order. Grammaticalization is increasing if the freedom in word order gets reduced. Thus, the Italian auxiliary ha ‘has’ in Maria ha scritto una lettera [Mary have:3SG written INDEF letter] ‘Mary has written a letter’ has a fixed position, while there are basically no word-order restrictions in the Latin constructions that preceded this sentence (e.g., habet scriptam Maria epistulam [have:3SG:PRS written Mary letter:ACC] ‘Mary has written a letter’). Thus the verb ‘have’ is more grammaticalized in Italian than in Latin.

3 Grammaticalization and Pragmatics—Subjectification, Reanalysis, and Analogy

Discourse and pragmatic inference are important driving forces of grammaticalization that frequently initiate grammaticalization processes. In Hopper and Traugott’s (2003) analysis, metaphoric and metonymic inferences trigger the two mechanisms of analogy and reanalysis, which are frequently discussed in different approaches to grammaticalization. As will be seen below, their relevance for grammaticalization is controversial. Other inference-based models are discussed under the headings of invited inference (Traugott, 2002) and subjectification and intersubjectification (Traugott, 2010).

Metaphoric inferences are generally defined in terms of ‘understanding and experiencing one kind of thing in terms of another’ (Hopper & Traugott, 2003, p. 84). In the literature on grammaticalization, metaphoric inference operates across conceptual domains. A good example is the cline in (5), in which a body-part term is seen in the light of an object or a spatial notion is transferred to the domain of time. For clines like (5) to be realized, metaphoric inference must be enhanced by analogy as a mechanism of rule extension. At an initial stage, there is a category A that is associated with X (A:X). To take an example from (5), the category space is associated with a body-part term (space:body part). If analogy establishes similarities of A with B, X is not only associated with A but also with B (A:X = B:X). If that similarity goes from space to time in our example (5), body-part terms can be used in both domains by way of rule extension in terms of analogy (space:body part = time:body part).

While metaphoric inference is paradigmatic in nature, metonymic inference is syntagmatic and must be seen in the context of other morphosyntactic units that are contiguous to the unit in question. Metonymic inference induces mechanisms of reanalysis, i.e., the assignment of a new morphosyntactic analysis to a given linguistic structure. This can again be illustrated by looking at body-part terms, more specifically at the example of body-part terms in the function of prepositions. This cross-linguistically frequent phenomenon is illustrated by a slightly manipulated version of English in (7b). In (7a), the lexical item back is analyzed as a noun in the head-position of a noun phrase (the construction refers to the rear part of the house). In (7b), exactly the same phonological surface is interpreted as a prepositional phrase with its head back of.



Thus, one and the same surface form that was first interpreted as a noun phrase in (7’a) is reanalyzed as a prepositional phrase in (7’b):



Analogy and reanalysis have both been illustrated by the example of body-part terms. This shows that both mechanisms and their corresponding metaphoric and metonymic inferences can join forces in processes of grammaticalization. Sometimes, they operate simultaneously, sometimes in cyclic sequences (e.g., analogy > reanalysis > analogy, etc.).

The relevance of analogy and reanalysis is assessed differently in the existing literature. Traugott (2011, p. 22) distinguishes four main positions. In the first one, the two mechanisms are independent but they interact as in the scenario outlined above. The second position understands grammaticalization as a subtype of reanalysis and is favored by approaches based on Universal Grammar (see Section 1.3). A somewhat opposite view is the third view taken by Haspelmath (1998, p. 315), who argues that reanalysis is irrelevant for grammaticalization because it is not unidirectional and because it is driven by ambiguous input structures that play no role in grammaticalization. Finally, Fischer (2007) claims that analogy is the dominant force in online processing. In concrete speech situations, innovations that may end up in successful grammaticalization are driven by analogy judgments in ongoing speech.

Grammaticalization often starts when a given construction occurs in a specific context that invites the hearer to draw a particular inference. This is what is described in Traugott’s (2002) Invited Inference Theory of Semantic Change (for other approaches in that context, see Diewald & Smirnova, 2012; Heine, 2002). If that same construction occurs more frequently in a similar context, the invited inference eventually gets conventionalized and develops into a generalized invited inference. At the final stage of grammaticalization, that generalized invited inference will become part of the grammar.

Subjectification and intersubjectification (Traugott, 2010) are based on inferences that take place in interactions between speaker and hearer. Similar to other inferences, they can also become conventionalized into grammatical meaning. In the case of subjectification, they encode attitudes and beliefs of the speaker. Thus, they are associated with meaning change that involves the speaker’s assessment of an event as for instance in the development of concessives out of temporals (English while) or in the use of manner adverbials (Old English anlice ‘simply, especially’) as focus particles in modern English only. At a next stage, meanings that underwent subjectification can be used to express meanings associated with the hearer by intersubjectification. Grammaticalized politeness systems that are based on the speaker’s estimation of the social status of the hearer are often described as typical instances of intersubjectification.

4 Grammaticalization—Problems with Universality and Independence

Newmeyer (1998) started a hot debate on the existence of grammaticalization as a coherent linguistic phenomenon (for a critical assessment of grammaticalization also see Campbell’s [2001] introduction to a special issue in Language Sciences 23 on this topic). In his chapter with the title ‘Deconstructing Grammaticalization’, he argued that ‘there is no such thing as grammaticalization’ (Newmeyer, 1998, p. 226). This is a very strong statement that is based on two main arguments:

  1. (i) There are counterexamples to unidirectionality (see the list in Newmeyer, 1998, pp. 263–278). Therefore, grammaticalization is falsified.

  2. (ii) Grammaticalization is not an independent and coherent phenomenon but rather an epiphenomenon of the three principles of reanalysis, phonetic change, and semantic change. Thus, grammaticalization is not needed to explain the linguistic changes it aims at accounting for.

From the perspective of those who do research on grammaticalization today, these criticisms miss an important part of what they are trying to understand. They are interested in explaining the mechanisms and regularities that recurrently produce phenomena as the ones introduced in Section 1.1 in a multitude of languages. For that reason, exceptions to unidirectionality do not necessarily disprove the whole field. In fact, researchers on grammaticalization are well aware that there are instances of degrammaticalization, i.e., processes of change that go against the direction of grammaticalization clines (e.g., Tabor & Traugott, 1998). More recent research, among it prominently Norde (2009), shows on the basis of more rigid and consistent definitions that the number of examples that revert grammaticalization clines is much smaller than it was claimed by Newmeyer (1998). One of the examples that still counts as an instance of degrammaticalization is the Old Swedish inflectional marker -er ‘nominative singular masculine,’ which is a derivational suffix of nominalization in modern Swedish and thus developed against the cline ‘derivation > inflection’ (Norde, 2009, p. 152). Finally, even Newmeyer (2001, p. 213) had to admit much earlier than the publication of Norde (2009) that unidirectionality ‘is not all that false’, no matter what the universal status of unidirectionality may ultimately be.

If degrammaticalization is limited to a small number of examples, the directionality of grammaticalization needs explanation. Grammaticalization must be more than an epiphenomenon of reanalysis, phonetic change, and semantic change as claimed above in (ii). In this context, the competing motivations of economy vs. iconicity play an important role (Section 1.3). Other important factors are discourse use and pragmatic inference (Section 4). Finally, there is even a very promising generative account based on changes from move to merger (Section 1.3)—the theoretical perspective from which Newmeyer (1998) wrote his deconstruction of grammaticalization.

5 Grammaticalization, Constructions, and Constructionalization

Constructions and the positions within them are very important for grammaticalization. They provide the linguistic environment that sets the course for the grammatical development of a linguistic item in terms of the grammatical domain to which its function will belong. This is illustrated by the Chinese verb zài ‘be at, live at, be alive’ as illustrated in (8a). If zài occurs in the preverbal position of an intransitive construction as in (8b) it will be reanalyzed as a progressive marker. If it gets associated with a prepositional construction it will be interpreted as a preposition with locative meaning (8c):



Constructions are discussed since quite some time in research on grammaticalization (Bybee et al., 1994; Hopper & Traugott, 2003). In the context of grammaticalization clines, it is important to notice that lexical items that have developed in different constructions do not form a cline. Thus, clines of the type [‘be at’ > preposition > progressive marker] or [‘be at’ > progressive marker > preposition] are inadequate because the two grammatical functions developed independently in different constructions. For that reason, a bifurcated path of the following type is much more pertinent:



In this traditional view, constructions are seen as syntactic environments within which processes of grammaticalization (reanalysis, analogy) take place.

In more recent times, research on grammaticalization has taken a new direction by integrating findings from Construction Grammar (Goldberg, 1995, 2006) and its basic assumption that constructions are linguistic patterns whose ‘form or function is not strictly predictable from its component parts or from other constructions recognized to exist’ (Goldberg, 2006, p. 5). This definition applies to all levels of grammatical analysis from morphemes or words to idioms and any type of phrasal structures. The basic notion of constructions as pairings of form and meaning can be illustrated by the use of a verb like laugh in the Caused-Motion Construction (X causes Y to move Z):



In the framework of Construction Grammar, the verb laugh in (10b) is an intransitive verb with its single participant role (cause) in the syntactic function of subject (SUBJ). The other two roles of theme (Y: the poor guy) and goal (Z: out of the room) with their respective syntactic functions of object (OBJ) and oblique (OBL) are contributed by the Caused-Motion-Construction, which also provides the overall meaning of ‘John caused the poor guy to move out of the room’. This model offers a coherent explanation that does not need to take recourse to a more or less ad hoc lexical entry of laugh as a ditransitive verb with an agent, a theme and a goal.

For Construction Grammar, constructions are the fundamental units of linguistic knowledge. If grammar is basically about the inventory of constructions in a given language and the network relations between them, the status and the overall functioning of grammaticalization in this new context need to be reassessed. The following two questions inspired by Gisborne and Patten (2011) are very helpful for that purpose:

  1. (i) New constructions are developed and existing constructions change. What does grammaticalization mean in such a context? Can these changes be called grammaticalization at all?

  2. (ii) What is the relationship between changes affecting the construction as a whole and changes to its components?

On (i): Constructional changes must be described within the existing typology of constructions, which is characterized by a continuum between two prototypical poles, the pole of specificity and the pole of schematicity:

  • Specific/substantive constructions are fully specified phonologically. They prototypically consist of lexical items. Thus, the lexical item tree with its concrete meaning is fully specified phonologically as [tri:]. We have one meaning represented by one formally fully specified sequence of phonemes.

  • Schematic constructions are defined in terms of more abstract categories that function like variables for larger sets of concrete linguistic items. In the case of the Caused-Motion-Construction, the more abstract schematic categories are cause, theme and goal on the semantic side and SUBJ, OBJ, and OBL on the syntactic side. Each of these categories can be productively filled by a large number of lexical items (e.g., John for cause/SUBJ, the poor guy for theme/OBJ, and out of the room for goal/OBL in (10b)).

As pointed out in the above definition of constructions as pairings of form and meaning, Construction Grammar does not make a principled divide between lexical and grammatical expressions. Thus, linguistic change affects both lexical/contentful constructions and grammatical/procedural constructions. Since the former type of construction is associated with lexicalization (see Section 1.4), only the latter type associated with grammar will be discussed in this article. This type develops from more concrete (specific/substantive) to more abstract (schematic) meaning. The properties of this development share many properties with grammaticalization:

  • Directionality (but not unidirectionality as with (iii) in Section 1.2): The development is directional, i.e., constructions move towards increasing abstraction/schematicity.

  • Gradualness (see point (vi) in Section 1.2): The change takes place gradually.

  • Analogy (see Section 4) is a mechanism that increases the productivity of items that can be integrated into a construction (e.g., filling intransitive verbs like laugh into the Caused-Motion Construction in (10b)).

  • Reanalysis (see Section 4) is relevant in the sense that a given surface structure is analysed in terms of a new structure (e.g., a verb with the meaning ‘want’ is reanalyzed as an auxiliary tense marker within an argument structure construction).

On (ii): The process in which new constructions emerge is called constructionalization. Since constructions are defined as pairings of form and meaning, both levels must be equally involved in this process:

Constructionalization is the creation of formnew-meaningnew (combinations of) signs. It forms new type nodes, which have new syntax or morphology and new coded meaning, in the linguistic network of a population of speakers. It is accompanied by changes in degree of schematicity, productivity, and constructionalization.

(Traugott & Trousdale, 2013, p. 22; the original is in italics, W.B.)

Constructionalization takes place gradually, i.e., new constructions are not supposed to arise at one go. Thus, constructionalization is preceded and followed by a succession of constructional changes, which are ‘affecting one internal dimension of a construction’ (Traugott & Trousdale, 2013, p. 26). While constructional changes potentially enable constructionalization, what is crucial for successful constructionalization is that both sides of the construction must be involved, its form side and its meaning side.

To give an example, constructionalization will be briefly illustrated by the case of a lot of in English. Initially, the word lot in its older form of lott had the meaning of ‘part’. With that meaning, it also implied the meaning of ‘a group of’ or ‘a set of’:



At about the same time, the quantification associated with the ‘group/set’-interpretation of lot leads to the pragmatic inference of quantification in general. This use is particularly associated with the plural form for denoting large quantities:



The above changes clearly affect the semantics. The lexical item lot has changed its meaning from ‘part of a whole’ to ‘quantification’. What cannot be determined in (11) and (12) is the syntactic status of lot(s)—is it still the head of a noun phrase or has it changed into a quantificational modifier? In example (13), the change in syntactic function is clearly observable. The noun phrase a lot of goods is taken up by the 3rd person plural pronoun them in the next clause. This indicates that the singular form of lot is no longer the head of the construction—a lot of has become a quantificational modifier of the plural noun goods, which as the new head requires them in the next clause.



The example of a lot of illustrates the lowest level of constructionalization which is concerned with the creation of a new construction based on a single lexical item. But this is not the whole story. As can be observed, instances based on individual lexical items converge into groups of lexical items with similar morphosyntactic properties that express different categories belonging to the same semantic domain. Thus, constructionalization takes place at different levels of abstraction. At the level of individual lexemes, constructionalization produces micro-constructions (e.g., the ‘a lot of’-construction). In principle, one could think of a considerable number of higher level constructions. Traugott and Trousdale (2013) only distinguish a three-level hierarchy for which they use the following terminology:

  • Micro-construction

  • Subschema

  • Schema

The ‘a lot of’-construction is not the only micro-construction expressing large quantity. There is another micro-construction formed by the quantificational modifier many. These two constructions together form a subschema for ‘large quantity’. Another subschema consists of two micro-constructions based on a bit of and a few, which can be subsumed under a subschema for ‘small quantity’. At the highest level, the two subschemata form a schema, the quantifier schema (Figure 1).

GrammaticalizationClick to view larger

Figure 1. The hierarchic organization of the quantifier schema (Traugott & Trousdale, 2013, p. 17).

Constructionalization as it develops from lower to higher degrees of schematization through the three stages of micro-constructions, subschemata, and schemata can account for many phenomena of linguistic change. The question is to what extend models based on constructionalization can account for grammaticalization. Some researchers argue that traditional approaches to grammaticalization with their focus on individual linguistic units (see Sections 1–3) can be subsumed under the constructionalization approach (Trousdale, 2010, 2012). Traditional approaches are criticized because they generally disregard the fact that linguistic units do not grammaticalize in isolation but are always combined with changes that concern the construction as a whole, i.e., with constructionalization. But even if constructionalization can account for many phenomena it is questionable whether it exhaustively covers all processes of grammaticalization. As Noel (2007) argues, there are good reasons to distinguish between grammaticalization and constructionalization even though the two processes often intersect. Such a distinction is straightforward with constructionalization in the domain of lexical/contentful expressions but it is less obvious if constructionalization is involved with grammatical/procedural constructions. Nevertheless, cases like the following two from Chinese (Bisang, 1996, 1998) show that constructionalization alone is not sufficient to account for processes of grammaticalization:

  1. (i) The existence of functional slots within constructions that are filled with new lexical items

  2. (ii) Rigid word order for markers belonging to different grammatical domains

On (i): Trousdale (2010) shows that the development of English modal verbs started out from individual lexical items that produced individual micro-constructions which were integrated into subschemata and schemata at later stages. In the case of Chinese, one can show that this is not the only scenario. If there is a fixed slot within a construction it is possible to fill that slot with new lexical items without concomitant changes in the construction. A good example is the verb ɡěi ‘give’ that developed into a benefactive marker that can occur preverbally (14a) and postverbally (14b):



From Peyraube’s (1988) detailed diachronic description of the benefactive and its expression between the 14th century BC and the 18th century AD, one can see that the option of taking both positions relative to the verb developed between the 6th and the 13th centuries ce. However, this option started out from another verb of giving, i.e., ‘give.’ Between 1250 and 1400, the verb kuì ok ‘give’ took over both functions, those of the full verb and those of the benefactive marker. In earlier periods, kuì ok had the more specific meaning of ‘to make a present of food, to offer in sacrifice’. The verb ɡěi ‘give’ only appeared later and had its breakthrough in novels like The Dream of the Red Chamber (written by Cao Xueqin in the 18th century). In such a scenario, constructionalization is not needed for explaning the use of verbs like kuì ‘give’ and ɡěi ‘give’ as benefactive markers. The verbs simply took over the function of another marker in a given syntactic slot.

On (ii): Chinese has a number of grammatical markers which are clearly derived from verbs. Quite a few of these markers still have maintained at least some of their verbal properties (see (8a) vs. (8b,c)):

  • Tense-aspect-mood (see zài as a progressive marker in (8b))

  • Preposition (see (8c) on zài ‘be at; at, in’)

  • Directionals indicating the direction of the action expressed by the main verb: in (15), the directional verb chū ‘come out’ indicates that the subject moves the object out of something and the verb lái stands for movement towards the speaker/subject/center of interest.



If more than one type of these markers is used in an utterance, their sequence follows strict word-order rules (see Section 2 on rigid word order as one of Lehmann’s [1995] parameters that indicate a high degree of grammaticalization). The following schema shows the possible positions of prepositions (PREP), directionals (DIR), and tense-aspect-mood markers (TAM) relative to the verb (V) in Chinese:



The development of the above three types of markers is clearly due to a considerable degree to constructionalization in the case of individual markers associated with a particular grammatical domain like TAM, but it is not clear how constructionalization can account for the rigid word-order rules in (16) that apply if markers from more than one domain are combined. For such an account, it seems necessary to look beyond constructionalization for other motivations. As is shown in Bisang (1992, 1996), this order has its cognitive-semantic motivation, through which the individual positions in (16) become attractor positions, i.e., positions that attract semantically compatible lexical items into the relevant position associated with the relevant grammatical domain.

6 Areal Differences and Grammaticalization

Processes of grammaticalization are not limited to individual languages. They also operate across languages in situations of contact. This phenomenon is comprehensively discussed by Heine and Kuteva (2005, p. 81), whose model starts out from the observation that speakers of a replica language R become aware that there is a grammatical category X in a model language M with which they are in contact. If for whatever reasons this category turns out to be useful for the speech community of R, they develop an equivalent category X on the basis of patterns and constructions available in their language by adopting universal strategies of grammaticalization. Since the material is taken from R, the newly developed expression format of X in R looks like an element of R even though it was induced by another language M.

Processes of contact-induced grammaticalization are not necessarily limited to two languages. If they take place across a larger number of languages within a coherent geographic area, we get grammaticalization areas like the Balkans, Ethiopia, South Asia, Southeast Asia, and Meso-America (Heine & Kuteva, 2005, pp. 182–217). This type of areal convergence is concerned with the diffusion and the realization of individual grammatical categories across languages. What is not considered in this perspective is the much broader question of the properties of grammaticalization as a whole. Does grammaticalization look the same all over the world or is there at least some area-specific variation? A closer look at the languages of East and mainland Southeast Asia (EMSEA) provides good evidence for the assumption that there is a certain degree of areal variation (Bisang, 2008, 2011, 2015a).

The area of EMSEA languages covers the five families of Chinese or Sinitic (a branch of Sino-Tibetan), Tai (the core group of Tai-Kadai), Mon-Khmer (a branch of Austroasiatic), Hmong-Mien (also called Miao-Yao), and Chamic (a subbranch of Austronesian spoken in Vietnam). In most languages belonging to this area, grammaticalization is characterized by the following two properties:

  1. (i) High relevance of pragmatic inference

  2. (ii) Limited co-evolution of meaning and form (see Section 1.2, point (vii))

On (i): Markers that express grammatical categories which are generally associated with high degrees of grammaticalization cross-linguistically are still subject to a large extent to pragmatic inference. Thus, even if an EMSEA language has markers for categories like tense-aspect-mood or (in)definiteness they are not obligatory (see Lehmann’s [1995] parameter of paradigmatic variability/obligatoriness; Section 2), i.e., they are omitted if they are inferable from context. In the following short text passage from Khmer, the events are explicitly situated in the past (‘once upon a time’) but there is no TAM-marking at all, even if the language has a rich inventory of such markers. The noun kɔma·rỳj ‘daughter’ is explicitly marked as indefinite by m‐nὲʎk [one-CL]. The other nouns have no (in)definiteness marking. Thus, the definiteness of mda·j ‘mother’ and the indefiniteness of dɔmlo·η‎ ‘sweet potatoes’ must be inferred from context (definiteness can be expressed by the distal demonstrative nùh ‘that’).



Another factor that keeps pragmatic inference high is the multifunctionality of many of its markers. Thus, numeral classifiers are not only used for individuating nouns in the context of counting as in Mandarin sān liàng chē [three-CL car] ‘three cars’, they can also mark definiteness as well as indefiniteness in a number of Sinitic languages like Wu Chinese and Cantonese (Mandarin classifiers only express indefiniteness postverbally; see Li & Bisang, 2012). In the following example from Cantonese, the classifier in the [CL-N]-construction is definite in the preverbal position (18a) and definite or indefinite in the postverbal position (18b). The definite or indefinite interpretation in the postverbal position depends on context.



On (ii): The clearest indicator of a high degree of grammaticalization in EMSEA languages is rigid word order (see the parameter of syntagmatic variability in Lehmann, 1995; Section 2). Other parameters associated with the co-evolution of form and meaning harmonize only to a limited extent with ongoing grammaticalization. Thus, the integrity of the linguistic sign only shows some effects of morphophonological reduction because there is a strong tendency to avoid morphemes which are smaller than a syllable. Moreover, morphological paradigms of the Latin type am-o [love-1SG:PRS] ‘I love’, am-as [love-2SG:PRS] ‘you love’, am-at [love-3SG:PRS] ‘s/he loves’ rarely develop in EMSEA languages (for exceptions, see Arcodia, 2013; Gerner & Bisang, 2010).

The data from EMSEA languages provide good evidence that there is cross-linguistic variation with regard to the extent to which the parameters developed by Lehmann (1995) matter. Finding out more about the limits of variation in these parameters beyond EMSEA languages looks like a promising research project.

7 A Final Remark from an Evolutional Perspective

In an evolutionary framework, linguistic structures as we find them in the population of the world’s languages are the result of the following factors that determine their successful diffusion within a speech community (Bisang, 2004; Croft, 2000):

  • Cognitive factors (e.g., economy vs. iconicity)

  • Communicative factors (e.g., pragmatic inference, discourse structure)

  • Physiological factors (production and perception of sounds)

  • Social and cultural factors

  • Universal Grammar (if assumptions on innateness are right)

In such an evolutionary scenario, one would expect a certain degree of cross-linguistic homogeneity in processes of linguistic change like grammaticalization, but there is no necessity for rigid uniformity. After all, the input situation differs within each language or variety, and the individual factors are also characterized by a certain degree of leeway. For reasons like these, unidirectionality with no exceptions (Newmeyer, 1998) is not to be expected. Similarly, there is no a priori reason to assume that the results of grammaticalization share exactly the same properties across the globe. This shows that there is a lot of space for future research.

Further Reading

Bisang, W. (2015). Problems with primary vs. secondary grammaticalization: The case of East and mainland Southeast Asian languages. Language Sciences, 47, 132–147.Find this resource:

Bisang, W., Himmelmann, N. P., & Wiemer, B. (Eds.). (2004). What makes grammaticalization?—A look from its fringes and its components. Berlin: Mouton de Gruyter.Find this resource:

Boye, K., & Harder, P. (2012). A usage-based theory of grammatical status and grammaticalization. Language, 88, 1–44.Find this resource:

Heine, B., & Kuteva, T. (2002). World lexicon of grammaticalization. Cambridge, U.K.: Cambridge University Press.Find this resource:

Heine, B., & Kuteva, T. (2005). Language contact and grammatical change. Cambridge, U.K.: Cambridge University Press.Find this resource:

Himmelmann, N. P. (2004). Lexicalization and grammaticalization. In W. Bisang, N. P. Himmelmann, & B. Wiemer (Eds.), What makes grammaticalization? A look from its fringes and its components (pp. 21–42). Berlin: Mouton de Gruyter.Find this resource:

Hopper, P. J., & Traugott, E. C. (2003). Grammaticalization (2d ed.). Cambridge, U.K.: Cambridge University Press.Find this resource:

Lehmann, C. (1995). Thoughts on grammaticalization. Munich: LINCOM.Find this resource:

Narrog, H., & Heine, B. (Eds.). (2011). The Oxford handbook of grammaticalization. Oxford: Oxford University Press.Find this resource:

Newmeyer, F. J. (1998). Language form and language function. Cambridge, MA: MIT Press.Find this resource:

Norde, M. (2009). Degrammaticalization. Oxford: Oxford University Press.Find this resource:

Roberts, I., & Roussou, A. (2003). Syntactic change: A minimalist approach to grammaticalization. Cambridge, U.K.: Cambridge University Press.Find this resource:

Traugott, E. C. (2010). (Inter)subjectivity and (inter)subjectification: A reassessment. In H. Cuyckens, K. Davidse, & L. Vandelanotte (Eds.), Subjectification, intersubjectification and grammaticalization (pp. 29–69). Berlin: Mouton de Gruyter.Find this resource:

Traugott, E. C., & Trousdale, G. (2013). Constructionalization and constructional changes. Oxford: Oxford University Press.Find this resource:

Van Gelderen, E. (2004). Grammaticalization as economy. Amsterdam: John Benjamin.Find this resource:


Arcodia, G. F. (2013). Grammaticalization with coevolution of form and meaning in East Asia? Evidence from Sinitic. Language Sciences, 40, 148–167.Find this resource:

Bisang, W. (1992). Das Verb im Chinesischen, Hmong, Vietnamesischen, Thai und Khmer. Vergleichende Grammatik im Rahmen der Verbserialisierung, der Grammatikalisierung und der Attraktorpositionen. Tübingen: Gunter Narr.Find this resource:

Bisang, W. (1996). Areal typology and grammaticalization: Processes of grammaticalization based on nouns and verbs in East and mainland South East Asian languages. Studies in Language, 20(3), 519–597.Find this resource:

Bisang, W. (1998). Grammaticalization and language contact, constructions and positions. In A. Giacalone Ramat & P. Hopper (Eds.), The limits of grammaticalization (pp. 3–58). Amsterdam: John Benjamins.Find this resource:

Bisang, W. (2004). Grammaticalization without coevolution of form and meaning: The case of tense-aspect-modality in East and mainland Southeast Asia. In W. Bisang, N. P. Himmelmann, & B. Wiemer (Eds.), What makes grammaticalization?—A look from its fringes and its components (pp. 109–138). Berlin: Mouton de Gruyter.Find this resource:

Bisang, W. (2008). Grammaticalization and the areal factor: The perspective of East and mainland Southeast Asian languages. In M. J. López-Couso & E. Seoane (Eds.), Rethinking grammaticalization. New perspectives (pp. 15–35). Amsterdam: John Benjamins.Find this resource:

Bisang, W. (2011). Grammaticalization and typology. In H. Narrog & B. Heine (Eds.), The handbook of grammaticalization (pp. 105–117). Oxford: Oxford University Press.Find this resource:

Bisang, W. (2015a). Problems with primary vs. secondary grammaticalization: The case of East and mainland Southeast Asian languages. Language Sciences, 47, 132–147.Find this resource:

Bisang, W. (2015b). Modern Khmer. In M. Jenny & P. Sidwell (Eds.), The handbook of Austroasiatic languages (pp. 677–716). Leiden, The Netherlands: Brill.Find this resource:

Bisang, W. (2016). Linguistic change in grammar. In A. Keith (Ed.), The Routledge handbook of linguistics (pp. 366–384). Oxford: Routledge.Find this resource:

Boye, K., & Harder, P. (2012). A usage-based theory of grammatical status and grammaticalization. Language, 88, 1–44.Find this resource:

Brinton, L. J., & Traugott, E. C. (2005). Lexicalization and language change. Cambridge, U.K.: Cambridge University Press.Find this resource:

Bybee, J., Perkins, R., & Pagliuca, W. (1994). The evolution of grammar. Tense, aspect and modality in the languages of the world. Chicago: University of Chicago Press.Find this resource:

Campbell, L. (2001). What’s wrong with grammaticalization? Language Sciences, 23, 113–161.Find this resource:

Chomsky, N. (1995). The minimalist program. Cambridge, MA: MIT Press.Find this resource:

Croft, W. A. (2000). Explaining language change. An evolutionary approach. Essex, U.K.: Pearson Education.Find this resource:

Diewald, G., & Smirnova, E. (2012). Paradigmatic integration: The fourth stage in an expanded grammaticalization scenario. In K. Davidse, T. Breban, L. Brems, & T. Mortelmans (Eds.), Grammaticalization and language change. New reflections (pp. 111–133). Amsterdam: John Benjamins.Find this resource:

Fischer, O. (2007). Morphosyntactic change: Functional and formal perspectives. Oxford: Oxford University Press.Find this resource:

Gerner, M., & Bisang, W. (2010). Classifier declinations in an isolating language: On a rarity in Weining Ahmao. Language and Linguistics, 11, 579–623.Find this resource:

Gisborne, N., & Patten, A. (2011). Construction Grammar and grammaticalization. In H. Narrog & B. Heine (Eds.), The Oxford handbook of grammaticalization (pp. 92–104). Oxford: Oxford University Press.Find this resource:

Givón, T. (1971). Historical syntax and synchronic morphology. Chicago Linguistics Society Proceedings, 7, 394–415.Find this resource:

Givón, T. (1979). On understanding grammar. New York: Academic Press.Find this resource:

Goldberg, A. (1995). Constructions: A construction grammar approach to argument structure. Chicago: University of Chicago Press.Find this resource:

Goldberg, A. (2006). Constructions at work. The nature of generalization in language. Oxford: Oxford University Press.Find this resource:

Gries, S., & Stefanowitsch, A. (2004). Extending collostructional analysis: A corpus-based perspective on “alternations.” Journal of Corpus Linguistics, 9, 97–129.Find this resource:

Haiman, J. (1983). Iconic and economic motivation. Language, 59, 781–819.Find this resource:

Haspelmath, M. (1998). Does grammaticalization need reanalysis? Studies in Language, 22, 315–351.Find this resource:

Heine, B. (2002). On the role of context in grammaticalization. In I. Wischer & G. Diewald. (Eds.), New reflections on grammaticalization (pp. 83–101). Amsterdam: John Benjamins.Find this resource:

Heine, B., Claudi, U., & Hünnemeyer, F. (1991). Grammaticalization. A conceptual framework. Chicago: University of Chicago Press.Find this resource:

Heine, B., & Kuteva, T. (2002). World lexicon of grammaticalization. Cambridge, U.K.: Cambridge University Press.Find this resource:

Heine, B., & Kuteva, T. (2005). Language contact and grammatical change. Cambridge, U.K.: Cambridge University Press.Find this resource:

Hilpert, M. (2008). Germanic future constructions: A usage-based approach to language change. Amsterdam: John Benjamins.Find this resource:

Himmelmann, N. P. (2004). Lexicalization and grammaticalization. In W. Bisang, N. P. Himmelmann, & B. Wiemer (Eds.), What makes grammaticalization?—A look from its fringes and its components (pp. 21–42). Berlin: Mouton de Gruyter.Find this resource:

Hoffmann, S. (2005). Grammaticalization and English complex prepositions: A corpus-based study. London: Routledge.Find this resource:

Hopper, P. J., & Traugott, E. C. (2003 [1993]). Grammaticalization (2d ed.). Cambridge, U.K.: Cambridge University Press.Find this resource:

Kastenholz, R. (1986). Grundkurs Bambara (Manding) mit Texten. Cologne: Rüdiger Köppe.Find this resource:

Krug, M. (2000). Emerging English modals: A corpus-based study of grammaticalization. Berlin: Mouton de Gruyter.Find this resource:

Kuryłowicz, J. (1965). The evolution of grammatical categories. Diogenes, 13, 55–71.Find this resource:

Lehmann, C. (1995). Thoughts on grammaticalization. Munich: LINCOM.Find this resource:

Li, X., & Bisang, W. (2012). Classifiers in Sinitic languages: From individuation to definiteness-marking. Lingua, 122, 335–355.Find this resource:

Lindquist, H., & Mair, C. (Eds.). (2004). Corpus approaches to grammaticalization in English. Amsterdam: John Benjamins.Find this resource:

Marchese, L. (1986). Tense/Aspect and the development of auxiliaries in Kru languages (Summer Institute of Linguistics, Publications in Linguistics, 78). Arlington: Summer Institute of Linguistics and University of Texas.Find this resource:

Matthews, S., & Yip, V. (1994). Cantonese. A comprehensive grammar. London: Routledge.Find this resource:

Meillet, A. (1912). L’évolution des formes grammaticales. Scientia (Rivista di Scienza), 12(26), 384–400.Find this resource:

Narrog, H., & Heine, B. (Eds.). (2011). The Oxford handbook of grammaticalization. Oxford: Oxford University Press.Find this resource:

Newmeyer, F. J. (1998). Language form and language function. Cambridge, MA: MIT Press.Find this resource:

Newmeyer, F. J. (2001). Deconstructing grammaticalization. Language Sciences, 23, 187–230.Find this resource:

Noel, D. (2007). Diachronic construction grammar vs. grammaticalization theory. Functions of Language, 14, 177–202.Find this resource:

Norde, M. (2009). Degrammaticalization. Oxford: Oxford University Press.Find this resource:

Peyraube, A. (1988). Syntaxe diachronique du chinois. Évolution des constructions datives du XIVe siècle av. J.-C. au XVIIIe siècle. Paris: Colège de France, Institut des hautes études chinoises.Find this resource:

Post, M. W. (2007). Grammaticalization and compounding in Thai and Chinese: A text frequency approach. Studies in Language, 31, 117–175.Find this resource:

Roberts, I., & Roussou, A. (2003). Syntactic change: A minimalist approach to grammaticalization. Cambridge, U.K.: Cambridge University Press.Find this resource:

Schiering, R. (2005). Cliticization and the evolution of morphology: A cross-linguistic study on phonology in grammaticalization (unpublished PhD diss.). Konstanz, Germany, University of Konstanz.Find this resource:

Tabor, W., & Traugott, E. C. (1998). Structural scope expansion and grammaticalization. In A. Giacalone Ramat & P. J. Hopper (Eds.), The limits of grammaticalization (pp. 229–272). Amsterdam: John Benjamins.Find this resource:

Traugott, E. C. (2002). From etymology to historical pragmatics. In D. Minkova & R. Stockwell (Eds.), Studying the history of the English language: Millennial perspectives (pp. 19–49). Berlin: Mouton de Gruyter.Find this resource:

Traugott, E. C. (2010). (Inter)subjectivity and (inter)subjectification: A reassessment. In H. Cuyckens, K. Davidse, & L. Vandelanotte (Eds.), Subjectification, intersubjectification and grammaticalization (pp. 29–69). Berlin: Mouton de Gruyter.Find this resource:

Traugott, E. C. (2011). Grammaticalization and mechanisms of change. In H. Narrog & B. Heine (Eds.), The Oxford handbook of grammaticalization (pp. 19–30). Oxford: Oxford University Press.Find this resource:

Traugott, E. C., & Trousdale, G. (Eds.). (2010). Gradience, gradualness and grammaticalization. Amsterdam: John Benjamins.Find this resource:

Traugott, E. C., & Trousdale, G. (2013). Constructionalization and constructional changes. Oxford: Oxford University Press.Find this resource:

Trousdale, G. (2010). Issues in constructional approaches to grammaticalization. In E. Gehweiler, E. König, & K. Stathi (Eds.), Grammaticalization: Current views and issues (pp. 51–72). Amsterdam: John Benjamins.Find this resource:

Trousdale, G. (2012). Grammaticalization, constructions and the grammaticalization of constructions. In K. Davidse, T. Breban, L. Brems, & T. Mortelmans (Eds.), Grammaticalization and language change. New reflections (pp. 167–198). Amsterdam: John Benjamins.Find this resource:

Van Gelderen, E. (2004). Grammaticalization as economy. Amsterdam: John Benjamins.Find this resource:

Wichmann, A. (2008). Speech corpora. In A. Lüdeling, M. Kytö, & T. McEnery (Eds.), Corpus linguistics: An international handbook (pp. 187–207). Berlin: Mouton de Gruyter.Find this resource:

Wichmann, A. (2011). Grammaticalization and prosody. In H. Narrog & B. Heine (Eds.), The Oxford handbook of grammaticalization (pp. 331–341). Oxford: Oxford University Press.Find this resource: