Chapter 2 – Hypotheticality and the Source of the Romance Conditional
2.5 Grammaticalisation of the Future and ‘Future in the Past’ Periphrases
The development of the future and ‘future in the past’ periphrases, as described above, is part of the process of grammaticalisation that results in the synthetic, or fused, forms of the future and conditional found in the Romance languages. Grammaticalisation, which may be defined as “that subset of linguistic changes through which a lexical item in certain uses becomes a grammatical item, or through which a grammatical item becomes more grammatical”, is divided into sub-processes or sub-stages, depending on the focus of study.131 Squartini and Klausenburger, for example, follow Heine and refer to the general processes by which grammaticalisation occurs. Desemanticisation v Decategorialisation v Cliticisation v Phonological erosion.132
Ramat, on the other hand, defines the verbal grammaticalisation process in terms of the stages reached in the development of the periphrases, and suggests the stages shown below. Full verb v Predicative construction v Periphrastic forms v Agglutination.133
131 Paul J. Hopper and Elizabeth Closs Traugott, Grammaticalization (Cambridge: Cambridge
University Press, 1993), p. 2.
132 Mario Squartini, Verbal Periphrasis in Romance: Aspect, actionality and grammaticalisation
(Berlin: Mouton de Gruyter, 1998), p. 19.
Jürgen Klausenburger,Grammaticalization: Studies in Latin and Romance morphosyntax(Amsterdam: Benjamins, 2000), p. 77.
Bernard Heine,Auxiliaries: Cognitive forces and grammaticalisation(Oxford: OUP, 1993), pp. 54-8
133
Paolo Ramat, ‘Introductory Paper’, inHistorical Development of Auxiliaries(see Vincent, above), pp. 3-19 (pp. 8-11).
Despite the different focus and terminology of these two examples, they both describe two groups of processes in the overall process of grammaticalisation. The first of these, which can be described as a semantic and morphosyntactic reinterpretation, includes Heine’s processes of desemanticisation and decategorialisation, and Ramat’s stages of full verb, predicative construction and periphrastic forms. The second group of morphonological processes - cliticisation and phonological erosion (Heine), or agglutination (Ramat) - describe the process of fusion through which the Latin analytic periphrasis becomes the synthetic inflected form found in the Romance languages. So far, only the periphrasis-specific reanalysis by which the infinitive +
HABEREperiphrasis came to take on future and ‘future in the past’ meaning has been
described. These developments will be briefly related to the stages of desemanticisation and decategorialisation, and the stages in the second process of fusion will be considered in greater detail.
Reanalysis
Reanalysis is defined by Hopper and Traugott as the process that “modifies underlying representations, whether semantic, syntactic, or morphological, and brings about rule change”.134 As Langacker insists, however, it is a “change in the structure of an expression or class of expressions that does not involve any immediate or intrinsic modification of its surface manifestation”, but is intertwined with, and results in, the processes of desemanticisation and decategorialisation.135 Desemanticisation describes the loss of lexical content of part of the periphrasis, resulting in the allocation of a syntactic role to what was, semantically, a full verb, although Squartini underlines the fact that “loss of lexical content of the auxiliary is always balanced by
134Hopper and Traugott, p. 32.
Hopper and Traugott present a detailed discussion of the mechanisms through which reanalysis takes place, from a broadly generativist point of view. However, their definition of reanalysis is curious, as they suggest that “one of the simplest types of reanalysis, and one very frequently found in grammaticalisation, is fusion: the merger of two or more forms across word or morphological boundaries”, whereas most authors would regard the process of fusion as aresultof reanalysis, not part of theprocess.
Hopper and Traugott, p. 40.
Fleischman,Thought and Language, pp. 58-9.
135
Ronald W. Langacker, ‘Syntactic Reanalysis’, inMechanisms of Syntactic Change, ed. by Charles N. Li (Austin: University of Texas, 1977), pp. 57-139 (p. 58).
an increase of grammatical meaning as a verb morpheme”.136 Decategorialisation occurs when a lexical item loses the ability to function in its full syntactic capacity, and becomes dependent on other, full, lexical items in the sentence. A verb, for example, may be limited to functioning as an auxiliary. Further stages of decategorialisation may occur even within a new, limited category: in the case of auxiliaries, for example, Roberts suggests a subsequent set of shifts from a lexical auxiliary, which still maintains some possessive meaning, to a functional auxiliary, which acts simply as a tense marker.137Once the auxiliary has reached the functional stage, and is dependent on the verb, rather than possessing full freedom to move in the sentence, it is a small shift to cliticisation, at which point the necessary adjacency of the clitic to the verb permits the possibility of fusion.
Parallel to desemanticisation and decategorialisation, lexical split may occur, where a lexical item continues to retain its full meaning in some contexts. The form and meaning of the full verb are retained independently of the grammaticalisation process taking place only in limited, specific, semantic and syntactic contexts. This process took place in the Romance languages, which “separate lexically from one another the two inherited senses of the verb to have”.138 Although habere in the context of an infinitive undergoes semantic bleaching and loss of autonomy, “the reflexes ofhabere
in other contexts have survived as full formal and semantic entities”, and retain their meaning of possession.139
The interaction of reanalysis, decategorialisation and desemanticisation can be used to describe the stages in the development of the infinitive + HABERE periphrasis. The
commonly used example, habebam litteras scribere, ‘I had a letter to write’, where
habebamwas used in its full sense of possession, is re-analysed to include a degree of
136Squartini,Verbal Periphrasis, p. 21. 137Roberts, p. 236.
138 Ernst Pulgram, ‘Latin-Romance habere: Double function and lexical split’, Zeitschrift für Romanische Philologie, 94 (1978), 1-8 (p. 5).
While Pulgram is referring to the past tense auxiliary, the general point he is making may also be applied to the future/conditional auxiliaries which have continued further down the path that he envisages for the past tense auxiliary.
obligative modality, ‘I had a letter that was going to have to be written’. At this point, no desemanticisation or decategorialisation has taken place, as the semantic force of possession is retained, habebamremains a full verb, and the structure of the sentence remains identical. The following stage in the process of grammaticalisation is the complete shift to a modal meaning of ‘obligation’, from ‘I had a letter to be written’ to ‘I had to write a letter’. This is still only an example of semantic reanalysis: desemanticisation has taken place, as the primary meaning ofhabebamas possession has been lost, but the structure remains identical. Decategorialisation has not yet taken place, as habebam continues to function as a full verb, but a shift in syntactic boundaries, referred to by Langacker as reformulation, has occurred, although this is not visible at the surface level.140The main verb habebam is now bracketed with the infinitive rather than the object and the syntactic pattern has become (scribere
habebam) litteras rather than scribere (litteras habebam). It is this syntactic
reanalysis, however, that “permits repositioning of the object complement […] and its eventual deletion” tohabebam scribere, ‘I had to write’ with purely obligative and no possessive meaning.141 This development is no longer an example of reanalysis, as an observable surface change has occurred. Both desemanticisation and decategorialisation have taken place, ashabebam has lost all possessive force. In this context,habebamis no longer acting as a full verb, but as an auxiliary, dependent on the presence of the infinitive. Ultimately, this shift results in the situation where “habere carries the ‘functional’ content while the infinitive carries the lexical, i.e. thematic content”, andhabebamfunctions as a clitic rather than as a full verb.142The semantic shifts that have created syntactic ties between the infinitive and habebam
have paved the way for the subsequent fusional changes, or in Langacker’s terms, resegmentation has provided the conditions in which boundary loss may occur.143
Fusion
Modern work on grammaticalisation and fusion proceeds from the insight of Givón, who coined the frequently cited expression: “today’s morphology is yesterday’s
140Langacker, p. 79. 141
Fleischman,Thought and Language, p. 59.
142
Roberts, p. 237.
syntax”.144 Fusion, or univerbation, is the shift from auxiliary to desinence. It is the second main stage in grammaticalisation, although Schwegler points out that “while semantic reanalysis must have occurred prior to any other type of synthesis, it is clear that diachronically these processes overlapped and interacted.”145 Bybee, Perkins and Pagliuca define two stages in the process, namely “the reduction or loss of phonological bulk and fusion of the grammaticizing material to surrounding material”, as well as distinguishing between dependence (cliticisation) and the actual fusion of the grammatical component with the verb.146
The synthesis of the Romance future and conditional have been summarised by Valesio as a development where “two separate words, members of a verbal phrase,
144 Talmy Givón, ‘Historical syntax and synchronic morphology: An archaeologist's field trip’, in Papers from the 7th Regional Meeting of the Chicago Linguistic Society, (Chicago: University of Chicago, 1971), pp. 393-415 (p. 413).
145Armin Schwegler,Analyticity and Syntheticity: A diachronic perspective with special reference to Romance languages(Berlin: Mouton de Gruyter, 1990), p. 133.
The cyclical introduction of analytic forms, their reduction to synthetic forms, and the subsequent introduction of further analytic forms has been well documented, notably by Schwegler. A welcome addition to the subject has been made by Bichakjian, who emphasises, in contrast to the cyclical approach taken by Pulgram, the linear nature of the individual analytic > synthetic changes that take place in the development of the future and conditional.
Ernst Pulgram, ‘Synthetic and Analytic Morphological Constructs’, in Weltoffene Romanistik: Festschrift für Alwin Kuhn zum 60 Geburtstag, ed. by Guntram Plangg and Eberhard Tiefenthaler (Innsbruck: Leopold-Franzens-Universität, 1963), pp. 35-42.
Bernard H. Bichakjian, ‘Language Change: Cyclical or Linear? The case of the Romance future’, in
Historical Linguistics: Papers from the 8th International Conference on Historical Linguistics (Lille, September 1987), ed. by Henning Anderson and Konrad Koerner (Amsterdam: Benjamins, 1990), pp. 37-49.
146Joan Bybee, Revere Perkins and William Pagliuca,The Evolution of Grammar: Tense, aspect, and modality in the languages of the world(Chicago: University of Chicago, 1994), p. 107-10.
Anderson further delineates the sub-stages of fusion which must be taken into account when describing the process of univerbation, stipulating that an analysis must distinguish between morphological, phonological, segmental and prosodic univerbation.
Henning Anderson, ‘From Auxiliary to Desinence’, in Historical Development of Auxiliaries, (see Vincent, above), pp. 21-51 (p. 31).
combine together as two morphemes, constituting a new single word”.147 While he also notes that the Romance future forms are more deeply grammaticalised than the unfused, analytic, past-tense forms, showing change “both of linguistic category and linguistic level category”, this does not account for the causes of the changes described.148 Nocentini states the difficulties in accounting for the differences between the perfect periphrasis and the future and conditional forms in terms of three problems to be addressed.149