This book presents a new theory of grammatical categories - the Universal Spine Hypothesis - and reinforces generative notions of Universal Grammar while accommodating insights from linguistic typology.
AUTHOR: Krämer, Sabine TITLE: Synchrone Analyse als Fenster zur Diachronie SUBTITLE: Die Grammatikalisierung von werden + Infinitiv SERIES: LINCOM Studien zu Germanistik 23 PUBLISHER: Lincom GmbH YEAR: 2005
REVIEWER: Martin Hilpert, Department of Linguistics, Rice University, USA
This book combines synchronic and diachronic perspectives in a study of the German construction 'werden' plus infinitive. In present-day German, this construction conveys the meanings of future time reference and epistemic modality. This ambiguity has been the subject of several synchronic and diachronic studies, but the author points out that any historical investigation needs to be informed by a satisfactory account of the present-day facts. Finding such an account to be absent, the author aims to take a fresh look at the syntax and semantics of 'werden' in present-day German, and to use synchronic analysis as a window on diachrony.
The book is divided into five chapters. Chapter one is a short introduction. Chapter two presents a synchronic account of 'werden' plus infinitive. Chapter three surveys previous approaches to its diachronic development. Chapter four lays out an alternative scenario based on the synchronic findings. Chapter five gives a summary.
Chapter one presents the subject matter of the analysis, which is the German auxiliary verb 'werden' with an infinitive complement. Sentences such as 'Peter wird singen' are ambiguous between the temporal interpretation 'Peter will sing' and the modal interpretation 'Peter probably sings (right now)'. The aim of the book is to give a synchronic account of this ambiguity as a basis for the reconstruction of the historical development that the construction has undergone.
The synchronic analysis in chapter two is preceded by an overview of previous work on 'werden' plus infinitive. The ambiguity of the construction has given rise to different and quite strongly opposed accounts. The recurring theme in these accounts is the question whether the construction instantiates the grammatical domain of future tense, or whether it is primarily an epistemic modal verb. Krämer argues that both of these positions are problematic: if 'werden' is to be classified as a tense marker, the definition of tense has to be extended considerably in order to accommodate modal meaning. Conversely, an account viewing 'werden' as a modal verb is at a loss to explain why it in some cases does not convey modal meaning, and why its syntactic behavior differs in several respects from the German epistemic modals 'müssen' and 'können'. These problems cast doubt on any unified account of 'werden' as either temporal or modal, so that Krämer advances a model that distinguishes between two separate senses. Krämer goes on to support her model with several pieces of evidence, showing that the temporal and modal meanings of 'werden' exhibit different characteristics.
First, assertive speech acts that commit the speaker to the truth of the expressed proposition occur only with temporal meaning. Examples such as 'Der Verlag wird Ihnen 200 Euro zukommen lassen' can only be interpreted as 'The publisher will pay you 200 Euros', if they are uttered as a felicitous assertive speech act, that is, by someone who has the authority to make this proposition, and who honestly intends to make it true at a later point in time.
Similarly, in complement clauses that are headed by the verb 'wissen', meaning 'know', 'werden' can only receive a temporal interpretation. Sentences such as 'Maria weiß, dass Peter gehen wird' can only be understood as 'Mary knows that Peter will leave'.
A third difference can be observed in relative clauses, where the distinction between restrictive and non-restrictive relative clauses differentiates the modal and temporal meanings of 'werden'. The example 'Die Kinder, die Mittagsschlaf machen werden, dürfen mit in den Zoo' translates temporally as 'The children who will take a nap may come along to the zoo.', if the relative clause is restrictive. By contrast, the non-restrictive reading 'The children, who will take a nap, may come along to the zoo.' induces a modal interpretation. As the glosses show, English 'will' behaves in exactly the same way.
Lastly, emphasis as conveyed through verum focus enforces a temporal interpretation of 'werden'. A sentence such as 'Hans WIRD seine Prüfung wiederholen' must be interpreted temporally as 'Hans is indeed going to repeat his exam'. This not only suggests that temporal and modal 'werden' are distinct, it also sets 'werden' apart from the German epistemic modals 'müssen' and 'können', which do tolerate verum focus. The sentence 'Hans KANN seine Prüfung wiederholen' can be interpreted epistemically as 'Hans could indeed repeat his exam'.
From this evidence, Krämer conludes that there are two separate senses of 'werden', and argues that these senses correspond to different underlying syntactic structures. She therefore posits two different lexical entries with different semantic and syntactic characteristics. Using the minimalist framework, she proposes that temporal 'werden' is generated in the functional projection T0 (Klein 1994). In this projection, 'werden' exclusively conveys temporal meaning. This assumption accounts for those examples in which modal meaning of 'werden' is not recognizable, or where a modal interpretation would even render the sentence ungrammatical. By contrast, modal 'werden' is generated in the relatively higher functional projection MoodP (Cinque 1999). This projection also hosts evidential adverbs that, like modal 'werden', serve to qualify the truth value of a given proposition. The assumption that modal 'werden' occurs in this syntactic position also explains that it does not tolerate verum focus, which evidential adverbs do not either.
Chapter three critiques previous suggestions regarding the diachronic development of 'werden'. There is a general consensus that the construction emerged in the late 13th century, and that its future interpretation grew out of the ingressive semantics of 'werden'. Beyond that, the proposed scenarios differ considerably.
The so-called 'erosion theory', first proposed in Weinhold (1883) derives the modern construction from uses of 'werden' with a participial complement. Middle High German sentences such as 'er wirt mich gerne sehende' - 'he will like seeing me' are viewed as the source construction. The present participle ending in '-ende' is assumed to reduce over time to the ending '- en', making it indistinguishable from a regular infinitive. The erosion theory of 'werden' suffers from the fact that attested infinitive complements predate the proposed erosion process, and that the process itself is poorly supported by the diachronic evidence. Hence, not only Krämer but in fact most modern accounts discard this theory.
Leiss (1985) proposes that the German construction came about as a borrowing from Czech. Krämer doubts that the contact situation between German and Czech was suitable for the proposed grammatical borrowing. Also, quantitative data suggests that the construction did not propagate westward from eastern Germany.
Schmidt (2000) argues that 'werden' came to be used with infinitive instead of participial complements by way of analogy to the modal verb 'sollen'. Functional overlap between the two verbs motivates the assimilation of complementation patterns. Krämer criticizes that the functional overlap of 'werden' and 'sollen' is merely a stipulation that does not fall out of the meaning of the two verbs. Also, Krämer doubts that the proposed model could result in the high type frequency of infinitive complements that is attested for early 'werden' plus infinitive. Krämer concludes that a satisfactory theory of the emergence of 'werden' that converges with findings about its synchrony is still missing. The fourth chapter aims to fill this gap and presents an alternative account of the history of 'werden'.
Starting from the synchronic account in chapter two, Krämer aims to develop a diachronic account that accommodates the modern status of 'werden' as a polysemous item with a temporal and an epistemic function. With most other accounts, she assumes that the source of the modern construction was 'werden' with a present participle complement. The source construction conveyed ingressive meaning, which is still present in modern sentences such as 'Peter wird wütend' - 'Peter gets angry'. Like Schmidt (2000), Krämer views the shift to infinitive complements as a process of analogy. However, whereas Schmidt takes the modal auxiliary 'sollen' to be the model for the analogy, Krämer suggests the verb 'beginnen' - 'begin'. This verb shares the ingressive semantics of 'werden', and is therefore a more convincing candidate.
The second step in the development of 'werden' is its reanalysis from an ingressive lexical verb to a temporal future auxiliary. Krämer adopts a view of reanalysis that includes changes in the type of syntactic head that is instantiated by a given element. To illustrate, the sentence 'Er wird trinken' - 'He will drink' is assumed to contain a number of empty categories such as TP, AgrP, and CP. The reanalysis from the ingressive interpretation 'Er [[ [wird trinken]VP ]]' - 'He starts to drink' thus consists of 'werden' being reanalyzed as one of the available empty categories. In this case, this is TP, as in 'Er [[ wird [trinken]VP ]TP ]. The temporal auxiliary 'werden' is therefore represented on a higher constituent of the syntactic tree than its lexical ingressive counterpart, such that the proposed development instantiates 'upward reanalysis' (Roberts and Rousseau 2003). The development of future 'werden' into an epistemic modal involves another upward reanalysis in which 'werden' comes to be located in the relatively higher constituent MoodP.
This book is highly instructive, as the author manages to make a clear case of a complex subject matter in less than 150 pages. The criticisms of previous accounts are presented clearly, and the general characterization of 'werden' as a polysemous item that has undergone reanalysis from a temporal auxiliary to an epistemic auxiliary is well-argued and supported by relevant evidence. The following criticisms therefore apply at the level of the theoretical assumptions that the author makes.
Regarding the synchronic account of 'werden' as an element with two lexical entries, it can be disputed whether a complementary distribution of temporal and epistemic meanings across different constructions really warrants the postulation of two separate senses. The fact that 'werden' has a temporal interpretation in restrictive relative clauses and an epistemic interpretation in non-restrictive relative clauses is quite striking, but if it is the syntax that disambiguates 'werden', why is it then necessary to posit lexical entries to do the same job?
Likewise, the postulation of two different syntactic structures for temporal and epistemic 'werden' is open to debate. Krämer motivates her claim that epistemic 'werden' is represented at a higher syntactic constituent with the fact that it encodes non-propositional information that can not be negated or questioned. For example, in the sentence 'Hans wird sich das nicht gefallen lassen' - 'Hans will not submit to that', epistemic 'werden' is outside the scope of the negator. Questions such as 'Wird Peter schlafen?' - 'Will Peter sleep?' can only be interpreted temporally. This evidence is accurate, but it only captures the well-known semantic fact that increasingly grammaticalized forms take on non-propositional meanings (Traugott 1989). It is quite possible, and indeed likely, that the semantic change of 'werden' was accompanied by a syntactic change, but such a change has to be documented through independent formal evidence.
Since the postulation of two syntactic structures in synchrony is the basis of the diachronic analysis, the latter stands or falls with the former. The 'upward reanalysis' of 'werden' in the syntactic tree offers an elegant formalization of the development of 'werden', but it requires a number of theory-internal assumptions of a minimalist approach to syntax (Roberts and Rousseau 2003), and it is not very explicit with respect to the question how the semantic change came about.
These criticisms notwithstanding, Krämer's account is of great interest not only to scholars of German, but to students of tense, modality, and grammaticalization in general.
Cinque, Guglielmo. (1999) Adverbs and Functional Heads. A Cross-Linguistic Perspective. Oxford: OUP.
Klein, Wolfgang. (1994) Time in Language. London: Routledge.
Leiss, Elisabeth. (1985) Zur Entstehung des neuhochdeutschen analytischen Futurs. Sprachwissenschaft 10, 250-73.
Roberts, Ian and Anna Rousseau. (2003) Syntactic Change. A Minimalist Approach to Grammaticalization. Cambridge: CUP.
Schmidt, Hans U. (2000) Die Ausbildung des werden-Futurs. Zeitschrift für Dialektologie und Linguistik 67/1, 6-27.
Traugott, Elizabeth C. (1989) On the rise of epistemic meanings in English: An example of subjectification in semantic change. Language 57/1, 33-65.
Martin Hilpert is a graduate student at Rice University. He is interested in grammaticalization, construction grammar, cognitive linguistics, and corpus linguistics. He is currently writing his dissertation, in which he compares the synchronic use and diachronic development of future constructions in the Germanic Languages from a usage-based perspective.