Reconstructing grammaticalization and the definition of grammer: an introduction

Gabriele Diewald (University of Hannover, Germany), Ilse Wischer (Potsdam University, Germany)
Recently a number of critical voices have been raised against the concept of grammaticalization as a theoretical framework of language change (cf. especially Campbell ed., 2001). Among other things it has been argued by the critics that

  • the changes described by grammaticalization are reducible mainly to reanalysis;
  • unidirectionality is no essential feature of grammaticalization;
  • grammaticalization being a complex and composite phenomenon has no status of its own, and thus is just epiphenomenal.

A lot of this criticism is factually unjustified as it is based on a superficial treatment of alleged counterexamples or is simply due to misinterpretations. Several of these issues have been convincingly refuted and clarified by Lehmann (in press).

Still, it must be admitted that a certain vagueness concerning the definition of central concepts found in some grammaticalization studies may have provoked some of the criticism. The term "grammaticalization" itself has become so widely overused that it is often applied to any kind of linguistic change involving just semantic generalization or univerbation. In our view, this vagueness goes back to the failure to clarify the defining features of a grammatical item, which is, after all, the target of any grammaticalization process. The aim of the workshop is to solve this problem and thereby to readjust the concept of grammaticalization on the basis of a clear definition of grammar.
Furthermore, we intend to arrive at a more comprehensive understanding of particular peripheral cases of grammaticalization, such as the development of derivational affixes, discourse markers or other linguistic phenomena whose status as elements of grammar have been challenged. In short, the workshop is meant as a contribution to the further development of central theoretical issues of grammaticalization theory.

The papers will focus on the following topics:

  • clarification of the notion of grammar as the target area of grammaticalization
  • description of grammaticalization  as a unified process comprising various mechanisms on several linguistic levels
  • distinguishing grammaticalization from other diachronic processes involved in the rise of grammar
  • discussion of controversial cases of grammaticalization (e.g. evolution of derivational morphemes, minor word classes, pragmatic markers, syntactic word-order).

References:

Campbell, Lyle (ed.) (2001): Grammaticalization: A Critical Assessment. Oxford: Pergamon (Special Issue of Language Sciences 23, 2-3).
Hopper, Paul J. und Elizabeth Closs Traugott ([1993] 2003): Grammaticalization. 2nd ed. Cambridge: CUP.
Lehmann, Christian (in press). "Theory and method in grammaticalization." In: Diewald, Gabriele (ed.). Grammatikalisierung. Themenheft der Zeitschrift für Germanistische Linguistik.
Wischer, Ilse und Gabriele Diewald (eds.) (2002): New Reflections on Grammaticalization. International Symposium, Potsdam, 17-19 June, 1999. Amsterdam: Benjamins (TSL 49).

Invisible Coding and Grammaticalization

Elisabeth Leiss (LMU Munich)
The philosophy behind this paper is that grammatical functions are regulary underspecified as to their overt morphological marking. Morphological underspecification is motivated by reasons of cognitive economy. We assume three strategies of marking grammatical functions: overt morphological marking, non-overt marking, and complex marking.

  1. Non-overt marking uses the "semantics of syntax" to express grammatical functions. The information structure of syntax creates environments which evoke readings of definiteness/indefiniteness. For example, the discourse thema creates a natural environment for definite noun phrases, whereas the rhema evokes indefinite readings. We refer to this marking as "iconic marking". It is based upon natural presuppositions concerning the discourse structure of  human language (invisible grammar)
  2. Overt grammatical marking is achieved by means of bound and non-bound grammatical morphemes (visible grammar).
  3. Complex marking means that grammatical functions are expressed by a combination of two grammatical functions in order to signal an additional grammatical function. It has meanwhile been well documented that the combination of case and aspect creates definiteness effects (complex grammar).

The central aim of this paper is to suggest that the magical formula "one form - one function" leaves considerable room for doubt even when clear cases of overt grammatical marking are concerned. Cases of functional morphemes which are not polysemous are rare. Processes of grammaticalization regularly end up in considerable amounts of polysemy. We will have a look at the grammaticalization of modals.
There is a layering of at least three stages which survived the process of grammaticalization: full verb, deontic modal with narrow scope, deontic modals with wide scope (so called weak epistemics), and epistemic modals. Thus, polysemy seems to be the destiny rather than some sort of by-product of grammaticalization. As polysemy is just another way to underspecify morphology there is reason to assume that "specified morphology" (overt marking) is the worst case, usually avoided for reasons of cognitive economy. As linguistic description tends to be hypnotized by overt morphological marking there are large areas of functional patterning still to be discovered. Just think of the vast continent of prosodic factors in spoken language which are of course audible, not visible. 

"Numeralization": the development of grammar and grammatical markers within numeral systems

Ferdinand von Mengden (FU Berlin)
Cardinal numerals are a linguistic universal. In the vast majority of languages these expressions constitute a numeral system. This numeral system consists of a small  number of simple (mono-morphemic) numerals and a set of morpho-syntactic rules which combine these simple numerals , and which thus generate complex numeral expressions (cf. Greenberg 1978).
The numeral system can be considered as a relatively autonomous subsystem of a language. It is part of the larger linguistic system (lexicon and grammar) in so far as the expressions generated by the numeral system, cardinal numerals, interact with other lexical classes, for instance when quantifying the referent of a noun phrase. Cf. the syntactically parallel structures of numerous books, many books, four books and two hundred and seventy-three books. In these examples, the complex constituent two hundred and seventy-three as a whole occupies the same paradigmatic slot as the simple expression three or as other mono-morphemic quantifiers. For the generation of a complex expression like two hundred and seventy-three completely different rules apply than the grammatical constraints that are valid outside numerals.
The general question I should like to raise in this paper is: in how far does the rise of the internal grammar of a numeral system and its marking strategies follow parallel paths than the emergence of the "outer grammar" of a language and its marking strategies. If grammaticalization is considered as a development from a less grammatical to a more grammatical function of an element or of a structure, then, the development of markers like e.g. the English suffixes -ty �� 10� in seven-ty �7 � 10� and -teen �+ 10� in seven-teen should be considered either as (part of) a process parallel to grammaticalization ("numeralization"?) or as a proper case of grammaticalization. In fact, the data we have from the history of English and the reconstruction of proto-Germanic (and, to a lesser extent, also from that of proto-Indo-European) will show that the features either definitory of or commonly involved in grammaticalization processes, such as phonological attrition, paradigmatic isolation / loss of autonomy, semantic bleaching, etc., also occur in the development from a less complex to a more complex numeral system.
In particular, I should like to discuss the English suffixes -ty �� 10� and -teen �+ 10� in the light of grammaticalization theory and thereby claim that both suffixes are clear cases of grammaticalization. In order to sketch the development of both suffixes it is necessary to go back to proto-Germanic and proto-Indo-European, i.e. into pre-historic stages of English for which we only have indirect evidence through comparative reconstruction. It has recently been claimed (Lehmann forthc.: 5-6) that data from reconstructed stages of a language cannot suffice as evidence for a grammaticalization process. This claim was exemplified by the history of the corresponding German suffix -zig �� 10� (Lehmann forthc.: 17-9). In this context, I should like to take up Lehmann's discussion of the methodological problems with respect to the evidence of reconstructed linguistic data for grammaticalization processes.

References:

  • Greenberg, Joseph H[arold] (1978). "Generalizations About Numeral Systems." in: Universals of Human Language. Volume III: Word Structure. Edited by Joseph H[arold] Greenberg. Stanford, Ca.: Stanford University Press. 249-95.
  • Lehmann, Christian (forthc.). "Theory and method in grammaticalization." in: Gabriele Diewald & Ilse Wischer eds. Grammatikalisierung. Themenheft der Zeitschrift für Germanistische Linguistik.

"Where grammar and lexis meet": Composite predicates in English

Laurel J. Brinton (University of British Columbia)
Composite predicates (CPs) in English of the form "light" verb + (article) + deverbal noun (e.g., give an answer, have a drink, make an attempt, take a walk, do a dive) as well as those of the form verb + (article) + deverbal noun + preposition (e.g., lose sight of, run a test on, put the blame on, pay attention to, offer an apology for) and of the form verb + preposition + (article) + deverbal noun (e.g., take into consideration, stand in need of, come to an end) are typically treated as instances of lexicalization. Wisher (forthcoming) in an article on combinations with have and take, while admitting that "[T]he degree of lexicalisation "is difficult to determine", concludes that they are "semi-lexicalized", that is, syntactically but not always semantically lexicalized. Traugott (1999:259) concludes that historically the development of composite predicates is "best construed as lexicalization", as does Claridge (2000:73). In contrast, Brinton and Akimoto (1999:17) point out that aspects of the development of CPs, such as decategorialization of the noun and its increasing fusion with the verb, resemble changes characteristic of grammaticalization. Others equivocate concerning the status of CPs. Quirk et al. remark that "on the one hand, lose sight of can be viewed as a single lexical item; on the other hand, it comprises three grammatical words, one of them (lose) with some morphological variation" (1985:1530n.). They conclude that such phrasal lexical items illustrate the "gradience between grammar and lexicon, including a gradience in lexicalization" (1530). Likewise, Algeo sees composite predicates as "[S]omewhere near the middle of the magnetic field of language - where grammar and lexis meet" (1995:203).
 This paper will make use of definitions of grammaticalization and lexicalization proposed in Brinton and Traugott (forthcoming), as well as material on the historical development of CPs in English in Brinton and Akimoto, eds. (1999) to argue that there are important distinctions among complex predicate types which would suggest that some CPs are indeed lexicalizations, while others are not. CPs of the second and third types mentioned above (e.g., raise an objection to, pass judgment on, catch hold of, lay claim to or put into practice, come to a stop, call into question, be in doubt) exhibit fossilization, fusion, loss of compositionality, and lack of productivity characteristic of lexicalization. In contrast, CPs consisting of the "light" verbs make, have, take, give, and do - the same set of verbs that have occurred in this construction since late Old English - have acquired the grammatical function of denoting telic aspect in CPs, while the meaning of the combinations remain transparent. CPs with light verbs are highly productive. These may thus be seen as instances of grammaticalization rather than of lexicalization. The papers concludes that CPs do not constitute a unified class, but develop differently along lines of grammaticalization and lexicalization, understood as complementary, not opposing diachronic processes.

References:

  • Algeo, John. 1995. "Having a look at the expanded predicate". The Verb in Contemporary English, ed. by Bas Aarts and Charles F. Meyer, 203-217. Cambridge: Cambridge University Press.
  • Brinton, Laurel J. and Elizabeth Closs Traugott. Forthcoming. Lexicalization and Language Change. Cambridge: Cambridge University Press.
  • Brinton, Laurel J. and Minoji Akimoto, eds. 1999. Collocational and Idiomatic Aspects of Composite Predicates in the History of English. Amsterdam and Philadelphia: John Benjamins.
  • Brinton, Laurel J. and Minoji Akimoto. 1999. "Introduction". In Brinton and Akimoto, eds., 1-20.
  • Claridge, Claudia. 2000. Multi-word Verbs in Early Modern English. Amsterdam and Atlanta, GA:Rodopi.
  • Quirk, Randolph, Greenbaum, Sidney, Geoffrey Leech, Jan Svartvik. 1985. A Comprehensive Grammar of the English Language. London and New York: Longman.
  • Traugott, Elizabeth Closs. 1999. "A historical overview of composite predicate types". In Brinton and Akimoto, eds. 239-260.
  • Wischer, Ilse. Forthcoming. "Lexicalisation of paraphrasal verb constructions with "have" and "take".

Valence promiscuity in Early Romance and its conse-quences for a systematic path in grammaticalization.

Werner Abraham ( University of Vienna, Austria)

The ensuing paper is about the promiscuity of active and passive forms and their functions in Early Romance. Specifically, ever since the synthetic simple passive was gradually abandoned since Late Latin, new composite forms emerged involving several full verbs, among which fakere. The unusual novel use of facere+P(ast) P(articiple) in passive function in Old Sardinian) (Logudorese) first signalled by Meyer-Lübke (1902) was hypothesized to reflect the early equivalence facere/fieri "make, do/be done" in Late Latin, whereby feci "made, did" equalled factus sum "made/done was". It is argued that indeed the grammaticalization of facere as a passive auxiliary might be related to the equivalence between the transitive verb facere and its pro-passive fieri. Other than argued by Cennamo and Meyer-Lübke it is held that this type of promiscuity can only be understood if one places it within the wider phenomenon of promiscuous verbal argument linking (a reflex of which is the equivalence facere/fieri). In order to arrive at the novel (grammar prominent) typological characteristic of S/Top/Th-V-O of Early Romance in succession to discourse prominent Latin Top-Th_V/V_Rh, what had to occur was exactly what happened: The necessary case-inflection supported links between the verbal argument grids and their disambiguating morphological identifications had to be disrupted. In other words, it is argued that this blurred pre-merge lexical syntax is the precedent stage for the rise of SVO and the novel equivalence of S=Top - an emergence against the prior Topic prominent stage of (Late) Latin. The overall claim is that erosion of case inflection alone will not suffice for the fundamental typological shift from Vlast to SVO (as has been claimed for Old English to Middle and Standard Modern English; see Charles Fries). Much rather, the language under pressure to change must undergo a diathetic reorganization as illustrated above.

The role of information structure in grammaticalization

Christian Lehmann (Universität Erfurt)

The contribution investigates how decisions made at the level of information structure predetermine the course of grammaticalization at various syntactic levels between the complex sentence and the clause, and how, on the other hand, grammaticalization levels out contrasts concerning information structure. Empirical domains investigated include

  • from topic to subject,
  • from correlation to relative construction,
  • from emphatic identifier to reflexive pronoun,
  • from purposive cosubordination to instrumental adjunct,
  • from converb construction to either periphrastic verb form or adpositional phrase,
  • from verb-focus / participant-backgrounding construction to verbal agreement.