Studies in Language. International Journal sponsored by the Foundation “Foundations of Language” - Volume 37, Issue 3, 2013

Language as a network of dependencies or constructions is a central feature of many cognitive theories of grammar. In this network, inheritance relationships are used to describe synchronic facts about a language whereby members of a less abstract set inherit properties from a more general set, and in the case of multiple inheritance, from more than one general set. This article explores some of the ways in which the language network may change over time, particularly the ways in which more than one constructional type may be considered to be the source of a change in the network.

This paper assumes that in order to explain rather than describe language change, historical linguists should not only consider what happens diachronically at the language output level but also, crucially, what speaker-listeners do at the processing level. The reason for this is that the structure of the language is shaped by the properties of the neurolinguistic mechanism underlying both language use and language learning. It will be argued that analogy as an important principle in grammar formation is the main mechanism in grammaticalization and in change in general when looked at from a processing point of view. The paper discusses the workings of analogy in a number of cases in the history of English which have traditionally been interpreted as unidirectional cases of grammaticalization . It will be shown instead that multiple source constructions were involved, which influenced one another and thus gave direction to the change.

This paper examines two cases of so-called syntactic amalgams. In syntactic amalgams a particular string that is shared by two constructions is exploited to combine them, in such a way that one of the constructions functions as a modifier of the other. Typical examples are after God knows how many years (< after many years + God knows how many years) and a big enough house (< a big house + big enough). In formal theories, these kinds of constructions have been insightfully described as ‘grafts’. However, the exact process through which these amalgams arise remains unexplored. When studied closely, these processes reveal form–function friction not fully accounted for by the graft metaphor.Syntactic amalgams typically serve a subjective function and have been recruited for this purpose. However, because they consist of a syntagm that is still internally parsable, they tend to resist full reanalysis. More precisely, their original syntax continues to constrain their use. As such, amalgams get caught between their original syntax, which remains transparent, and their new function, which suggests a new syntactic status.This appears clearly from contrastive studies of amalgams in Dutch and English that are functionally similar but whose use is constrained in different ways due to structural differences between the two languages. Our first case study deals with the Dutch and English amalgam wie weet / who knows. A contrastive analysis of the development of the respective items shows both the conservative effect of the origin of change and the attraction exerted by the target of change. The second case we discuss in detail involves so-called transparent free relatives. A contrastive analysis shows the role of the overall grammar of a language in licensing change, in this case with Dutch word order posing more difficulties to the new focusing function of transparent free relatives.In general, both case studies show the formation of syntactic amalgams to be sensitive to system pressures both in the course of their development and in the eventual outcome of change.

The interaction of several cases of grammaticalization going back to different source constructions may have dramatic consequences on a linguistic system. This paper discusses the so-called German scandal construction, which consists of a verbal complex displaying a number of synchronic form-meaning mismatches. The scandal is solved if a diachronic point of view is adopted, which shows that the mismatches result from the intraference of two different source constructions expanded in the course of the German linguistic history, namely the Perfekt construction and the zu-infinitive construction.

This paper explores the complex role of language contact in the development of be and have auxiliation in the periphrastic perfects of Europe. Beginning with the influence of Ancient Greek on Latin, it traces the spread of the category across western Europe and identifies the Carolingian scribal tradition as largely responsible for extending the use of the be perfect alongside the have perfect across Charlemagne’s realm. Outside that territory, by contrast, in “peripheral” areas like Iberia, Southern Italy, and England, have came to be used as the only perfect auxiliary. Within the innovating core area, a further innovation began in Paris in the 12th century and spread to contiguous areas in France, Southern Germany, and northern Italy: the semantic shift in the perfects from anterior to preterital meaning. What can be concluded from these three successive instances of diffusion in the history of the perfect is that contact should be regarded as one of the essential “multiple sources” of innovation, and as a fundamental explanatory mechanism for language change.

This article approaches the origin of multilingual innovations in language contact by presenting data from an intensive contact situation between two closely related Finnic languages, Finnish and Estonian. This situation between languages that have complex interconnected morphological systems often leads to the emergence of structures that contain material from both languages. The origin of these structures is discussed in the light of two basic assumptions: that in the mind of the bilingual speaker, the elements of the two languages are connected to each other via phonological and semantic or functional similarities and that both languages of the bilingual speaker always remain active and available during speech processing. The simultaneous activation of synonymous, competing elements or patterns is hypothesised to be the cause for the inevitable presence of both language-internal and cross-linguistic forces at the initial stage of language change.

Multiple sources abound in language, at all levels of linguistic analysis (phonology, syntax, semantics, etc.), and in a range of historical pursuits, including etymology and variationist investigations. From a methodological standpoint, moreover, recognizing multiple sources is often good historical linguistic practice (contrary to inclinations towards neat and elegant solutions that satisfy Occam’s Razor). That is, if we can identify multiple pressures on some part of a language system, it cannot always readily be excluded that some or even all might have played a role in shaping a particular development; if all of the factors represent reasonable pressures that speakers could have been aware of and influenced by, excluding any could simply be arbitrary. In this paper, accordingly, I survey the breadth of multiple sources in a variety of areas of language change, and advance one particular consequence that multiple sources can lead to, namely the hypothesis that recognizing multiple sources can be a basis for positing proto-language variation that is realized in variation within single languages and across related languages.