Grammaticalization and the rise of configurationality in Indo-Aryan by Uta Reinöhl (review)

Although relatively short, this book tackles central issues in historical syntax and grammaticalization theory, with implications far beyond the Indo-Aryan case study that forms the empirical core of the study. It can be read on two levels: on the one hand, as a historical micro-analysis of two postpositions and their source lexemes across three millennia of Indo-Aryan (I-A), and on the other, as a set of theoretical claims regarding the mechanisms by which the claimed ‘strong non-configurationality’ of Vedic ultimately spawned the ‘low-level configurationality’ of the modern I-A languages.

Uta Reinöhl draws on a text corpus spanning several chronological stages of I-A: Vedic Sanskrit, early Middle I-A Old Awadhi, late Middle I-A Apabramsha, early New I-A Awadhi, and contemporary Hindi (9–17). The two postpositions have the forms mẽ ‘in’ and par ‘on’ in Hindi, but R refers to their various I-A cognates as madhye and upari respectively, a convention I also adopt in this review. The corpus includes a total of 1,060 tokens, 844 of madhye and 216 of upari, which are detailed in the Appendix. This review concentrates on the book’s broader implications for diachronic syntax, while an assessment of the philological details is referred to the specialists of historical I-A.

The book’s main hypothesis is that phrasal structure co-evolves with the grammaticalization of lexical elements into purely functional items. The functional items here are the postpositions of contemporary I-A, which constitute a new form class that was entirely absent in ancient I-A. R proposes that with the emergence of postpositions, a previously unattested phrase type (the adpositional phrase) was introduced into I-A syntax, and this in turn contributed to the fixation of phrasal structure in NPs (or DPs). In this sense, then, the development of adpositional phrases spearheaded the development of phrasal structure in I-A. R’s proposals are profoundly influenced by Himmelmann’s (1997) thinking on grammaticalization: grammaticalization is not merely the evolution of individual lexical items down a cline of increasing grammaticalization, but it also creates novel constructional syntax. On this view, grammaticalization works in parallel, in that individual items shift in terms of, for example, obligatoriness or paradigmaticity, and at the same time syntactic structures specifically geared to these functional elements crystallize. In this sense, syntax is ‘emergent’.

With regard to the grammaticalization of adpositions, R rejects the widespread view that the adpositions of I-A languages developed from the Indo-European ‘adverbial particles’. These particles constituted a set of syntactically very heterogeneous items in ancient Indo-European, regularly cited as a main source of adpositions in contemporary Indo-European. According to R, I-A underwent a distinct development, setting it apart from ‘other Indo-European branches’ (87). In Ch. 4, the ‘traditional scenario’ (Section 4.1), widely assumed to hold for Indo-European, is outlined. On this view, the birth of adpositional phrases involved the grammaticalization of what R refers to as ‘symmetrical groups’, groups of nonhierarchically associated syntactic items that included the adverbial particles and case-marked nouns. Through frequent collocation, such groups apparently ultimately crystallized to grammatical phrases (87). In Ch. 4, R assembles three lines of evidence that militate against the ‘traditional scenario’ for I-A. First is the lack of evidence for an etymological connection between the adpositional particles and the later postpositions (the latter being of extremely varied provenience (83)). Second, in a meta-analysis of the corpus studies of Rigvedic syntax undertaken by Hettrich and associates (e.g. Hettrich et al. 2010), she shows that among the attested noun-plus-particle sequences, the ordering particle-noun is actually more frequent than the inverse order, a finding that runs counter to the expectations that the particles became postpositions. Finally, R refers to a distinct pattern of cliticization in I-A, which may have inhibited the reanalysis of the particles as...