*ABA revisited: Evidence from Czech and Latin degree morphology

Glossa a journal of general linguistics De Clercq, Karen and Guido Vanden Wyngaerd. 2017. *ABA revisited: Evidence from Czech and Latin degree morphology. Glossa: a journal of general linguistics 2(1): 69. 1–32, DOI: https://doi.org/10.5334/gjgl.371 RESEARCH *ABA revisited: Evidence from Czech and Latin degree morphology Karen De Clercq1 and Guido Vanden Wyngaerd2 1 FWO/U Gent, Blandijnberg 2, 9000 Gent, BE 2 KU Leuven, Warmoesberg 26, 1000 Brussel, BE Corresponding author: Karen De Clercq () We present a novel account of root suppletion in comparatives and superlatives, and show how it accounts for the presence of ABB and ABC patterns, as well as the absence of ABA patterns. The account assumes that suppletive roots, despite appearances to the contrary, are not contextual allomorphs, but portmanteaus spelling out two distinct features, one belonging to the lexical root, and another one belonging to the comparative. The regular comparative affix then spells out an additional feature relating to the comparative domain. In other words, we show that the comparative (cmpr) head that enters into the morphological makeup of the comparative (Bobaljik 2012) is to be split up into two distinct heads, C1 and C2 (see also Caha 2016). We extend this idea to sprl, which we show is likewise to be split up into S1 and S2, in order to account for suppletive ABC patterns. These four distinct heads receive empirical support from facts of the degree morphology in Czech and Latin. The new account of root suppletion allows a straightforward way of deriving the attested and unattested patterns of (root) suppletion in degree comparison. The analysis developed supports the hypothesis that the absence of AAB patterns in degree comparison is due to a constraint of a different nature altogether. Keywords: *ABA; adjectives; suppletion; comparative; superlative 1 Introduction Bobaljik (2012) argues that it is a universal property of (morphological) comparatives and superlatives that, when comparatives have a suppletive form, the superlative will also be suppletive and vice versa, i.e. there are no ABA-patterns as in (1b), nor AAB-patterns as in (1c). (1) a. good-better-best b.*good-better-goodest c.*good-gooder-best Bobaljik calls this the Comparative-Superlative Generalisation (CSG), and adduces extensive evidence in support of the fact that this generalisation is a language universal.1 Bobaljik’s account for the CSG relies on a number of ingredients. The first ingredient is the Containment Hypothesis, which states that “the representation of the superlative properly contains that of the comparative” (Bobaljik 2012: 4; see Dunbar & Wellwood 2016 for a recent alternative to Bobaljik’s proposal). Concretely, the (derived) structure of comparative adjectives is as in (2), and that of superlatives is as in (3).2 The CSG does not apply to so-called absolute superlatives or elatives, which indicate a very high degree of the adjective, and which lack a comparative meaning component (e.g. Italian buonissimo ‘very good’ vs il migliore ‘the best’). 2 These structures are derived from run-of-the-mill right-branching syntactic representations through an operation M, the exact nature of which Bobaljik is noncommittal about, e.g. Morphological Merger (Marantz 1988) or Local Dislocation (Embick 2010); see Embick & Noyer (2001) for discussion. 1 c. *good-gooder-best De Clercq & Vanden Wyngaerd Bobaljik calls this the Comparative-Superlative Generalisation (CSG), and adduces extensive in support of the factGeneralisation that this generalisation is Bobaljik calls this evidence the Comparative-Superlative (CSG), and b. *good-better-goodest 1 a language universal. Bobaljik’s account forfact thethat CSGthis relies on a number adduces extensive evidence in support of the generalisation is Art.ingredients. 69, page 2 of 32The first De Clercq and Wyngaerd: *ABA revisited c. *good-gooder-best 1 of ingredientaccount is the Containment Hypothesis, which a language universal. Bobaljik’s for the CSG relies on a number states that “the The representation of Generalisation theis superlative properly contains that of of ingredients. first ingredient the Containment Hypothesis, which Bobaljik calls this the Comparative-Superlative (CSG), and the comparative” (Bobaljik 2012: 4; see Dunbar & Wellwood 2016 for a restates that “the representation of the superlative properly contains that of adduces extensive evidence in support of the fact that this generalisation is cent alternative to(Bobaljik Bobaljik’s proposal). the (derived) 1 the comparative” 2012: 4; CSG see Concretely, Dunbar & aWellwood 2016structure for a rea language universal. Bobaljik’s account for the relies on number (2)comparative adjectives is as in (2), and that of superlatives is as in (3).2 cent to Bobaljik’s proposal). Concretely, the (derived) structure of ingredients.of Thealternative first ingredient is the Containment Hypothesis, which 2 comparative adjectives is as in (2), and that of superlatives states that “theofrepresentation of the superlative properly contains that of is as in (3). (2) cmprP (3) sprlP the comparative” (Bobaljik 2012: 4; see Dunbar & Wellwood 2016 for a recmprP (3) (derived) structure sprlP cent alternative(2) to Bobaljik’s proposal). Concretely, the of comparative(3) adjectives in (2), and that of superlatives iscmprP as in (3).2 sprl A is ascmpr 2 De Clercq & Vanden Wyngaerd A cmpr cmprP sprl (2) cmprP (3) sprlP A cmpr b. *good-better-goodest A cmpr c. *good-gooder-best The second ingredient are the rules of exponence A cmpr cmprP sprl which regulate lexical insertion. Thethe relevant rules good-better alternation are given in (4): Bobaljik callssecond this Comparative-Superlative Generalisation and The ingredient arefor thethe rules of exponence which(CSG), regulate lexical adduces extensive evidence in support of the fact that this generalisation is  insertion. The relevant rules for the good-better alternation are given in (4): lexical insertion. The The second ingredient are the of exponence which regulate A rules cmpr (4) a. good → good 1 a language universal. Bobaljik’s account for the CSG relies on a number  relevanta. foringredient the good-better alternation are given in (4): b.rulesgood be(tt)/is the ]Containment cmpr ] (4) → good of ingredients. Hypothesis,  first The second ingredientThe are the rules of exponence which regulate lexical which b. representation good → be(tt)/ superlative ] cmprproperly ] states that “the of the contains that insertion. The These relevant rules for the good-better alternation given (4): rules state good will be inserted are under theinterminal thatofdomi that 2012: the  comparative” (Bobaljik 4; see Dunbar & Wellwood 2016 for a re(4) the a. good nates root good, → except theinserted more specific of cmpr, These rules state willinbe under context the terminal that where domi that good to Bobaljik’s proposal). Concretely, the (derived) structure (4) a.centalternativ (...truncated)