1 The loss of analyticity in the history of Romanian verbal morphology Adina Dragomirescu, Alexandru Nicolae, and Rodica Zafiu Abstract This chapter investigates the relation between syntheticity and analyticity in the history of Romanian. We empirically note that, against the traditional hypothesis that in the passage from Latin to Romance older synthetic forms were extensively replaced by novel analytic formations, there is a set of old Romanian periphrastic constructions which disappeared in the passage to modern Romanian, and we set our goal to provide an explanation for this less frequent linguistic change. By bringing to the fore data from old and modern Romanian, as well as dialectal Daco- Romanian material, we show that the facts are best explained by appealing to structural factors (i.e. the feature matrix of Romanian auxiliaries) rather than functional factors (competition between functionally equivalent forms, learnèd nature, or rare occurrence in usage). Keywords: Romanian, verb morphology, syntheticity, analyticity, periphrasis, inflexional simplicity 8.1 Introduction, outline and minimal theoretical assumptions Addressing the general relation between syntheticity and analyticity in the passage from Latin to Romance, this chapter analyses the diachronic development of verbal periphrases in the passage from old to modern Romanian, including its dialectal varieties; only Daco-Romanian and its regional varieties are considered in this study, while the sub-Danubian historical dialects have not been taken into account. Our analysis starts from the empirical observation that in the passage from old to modern Romanian, despite the loss of some inherited synthetic formations (see Maiden 2018:29-43), there is a subset of novel periphrastic formations which disappeared from the standard language, with just some relics preserved dialectally. While traditional scholarship argued that synthetic forms were replaced tout court by analytic formations in the transition from Latin to Romance (cf. Schlegel 1818 and many references thereafter), more recent work has argued against such a radical view (Schwegler 1990; Ledgeway 2012 and, especially, 2017 1 ), focusing on the cross-Romance variation in this area and highlighting the particular developments of each Romance variety in turn 2 . In this respect, 1 Cf. also Coseriu’s (1987, 1988) distinction between ‘internal’ and ‘external’ structures, onto which he develops the hypothesis that that the Romance languages are distinguished from Latin by an iconic typology whereby relational concepts (external structures) receive relational, ‘syntagmatic’ (i.e. analytic) exponence and non-relational concepts (= internal structures) receive non-relational, ‘paradigmatic’ (i.e. synthetic) exponence. As Coseriu himself admits, this principle is not without exceptions. 2 The disappearance of newly coined periphrases is found in other Romance idioms as well. Consider French, for example. The periphrases made up of the verb estre (être) ‘be’ or aller ‘go’ plus a present participle go extinct (Squartini 1999:27; Buridant 2000:357-8); the loss of periphrases whose lexical verb occurs in the gerund is also attested in Italian, Occitan, Galician, and Portuguese (Squartini 1999:28-9). Of the numerous compound auxiliary forms of French – double auxiliary + past participle: j’ai eu fait (I-have.AUX.1SG have.AUX make.PTPC), j’avais eu fait (I-have.AUX(IMPF).1SG have.AUX make.PTPC), j’aurais eu fait (I-have.AUX(COND).1SG have.AUX make.PTPC), etc., see Ayres Bennett & Carruthers (1992) –, the only one which has been preserved in non-standard and regional French (in the south) is the one employing the auxiliary in the compound past (the “passé surcomposé”) – j'ai eu fait, see Carruthers (1994), Apothéloz (2010), Melchior (2012), etc. Note that there is a formal correspondence between this form and a Romanian compound auxiliary form preserved dialectally – am fost dat (have.AUX.1SG be.AUX give.PTCP) (see example (24) below). In the particular case of French, the demise of many of these periphrases might have also been favoured by the fact that they were largely ignored or condemned by the French grammars (Ayres Bennett & Carruthers 1992).