Proceedings of SALT 33 - Two question-embedding strategies and answer-orientedness

Sep 17, 2026

Uploaded by Violet

Abstract Japanese and Turkish attitude predicates combine with two main kinds of embedded clauses: nominalizations and clauses introduced by the morphemes to and diye. The paper analyzes their interrogative variants, arguing that nominalizations yield answer-oriented inferences with responsive predicates, while diye/to interrogatives are question-oriented and entail that the attitude holder linguistically produces the interrogative. A compositional fragment is proposed for how attitude predicates take nominalizations as arguments and how diye/to-clauses enrich meanings via production inferences.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages21

Reading Time53 min

AccessFree to download

1 Introduction

Attitude predicates and clause embedding strategies

Proceedings of SALT 33: 320–340, 2023 Two question-embedding strategies and answer-orientedness* Deniz Özyıldız Universität Konstanz Wataru Uegaki University of Edinburgh Abstract Japanese and Turkish attitude predicates combine with two main kinds of embedded clauses: Nominalizations, and clauses introduced by the morphemes to and diye. We describe their interrogative variants, showing that nominalizations give rise to answer-oriented inferences with responsive predicates (e.g., factivity, belief), but that diye/to interrogatives are question-oriented and entail that the attitude holder linguistically produces the interrogative. We propose a compositional fragment where attitude predicates take nominalizations as arguments, which they may impose semantic restrictions on, and where diye/to-clauses modify and enrich attitude meanings with a linguistic production inference. Keywords: attitude verbs, clause embedding, questions, quotation 1 Introduction Languages have different ways of combining clausal constituents with attitude predicates. In this paper, we focus on combining attitude predicates with nominalized clauses, in (1), and clauses introduced by the morphemes diye and to, in (2), in Turkish and in Japanese (respectively in a. and b.) . (1) a. Ai [ kar ya˘g-dı˘g-ın-a ] ¸sa¸sırdı . Ai snow fall-NMZ-3 S . POSS-DAT was.surprised b. Ai-wa [ yuki-ga hutta-no-ni ] odoroita Ai-TOP snow-NOM fall-NMZ-DAT was.surprised Ai was surprised that it was snowing. ⇒ p ∧Bap (2) a. Ai [ kar ya˘g-ıyor diye ] ¸sa¸sırdı . Ai snow fall-PRES.3 S DIYE was.surprised b. Ai-wa [ yuki-ga hutta-to ] odoroita Ai-TOP snow-NOM fell-TO was.surprised Ai was surprised, thinking “it’s snowing.” ⇏ p * We thank the participants of the MECORE workshop in Edinburgh and at SALT 33, and our reviewers. This work is funded by the project “MECORE: Cross-linguistic investigation of meaning-driven combinatorial restrictions in clausal embedding”(AHRC: AH/V002716; DFG: RO 4247/5-1) . ©2023 Özyıldız, Uegaki Two question-embedding strategies and answer-orientedness As is the case here, the same predicate may often combine with either type of clause, with the choice giving rise to interpretive differences. One difference visible here is that the sentences in (1) are factive, where “be surprised” combines with nominalized declaratives, while the ones in (2) are non-veridical, with D/T-declaratives. This phenomenon, known as a factivity alternation, has been studied in a variety of languages for declarative complements.1 A second difference, which has received less attention in the literature and will play a central role here, is that diye and to give rise to a linguistic production inference. This inference, rendered into English with the adjunct clause thinking “S”, requires that the attitude holder say, mentally entertain, or otherwise produce the clause that diye and to introduce.2 Many predicates that combine with nominalized and D/T-declaratives also combine with nominalized and D/T-interrogatives, in (3) and (4) . (For the nominal status of Japanese embedded questions like in (3b), which don’t feature an overt nominalizer, see Tomioka 2020.) These two interrogative embedding strategies form the novel empirical landscape we explore in this paper. (3) a. Ai [ kar-ın ne zaman ya˘g-dı˘g-ın-a ] ¸sa¸sır-dı . Ai snow-GEN when fall-NMZ-3 S . POSS-DAT was.surprised b. Ai-wa [ yuki-ga itsu hutta-ka-ni ] odoroita. Ai-TOP snow-NOM when fell-Q-DAT was.surprised Ai was surprised by when it had snowed. (4) a. Ai [ kar ne zaman ya˘g-dı diye ] ¸sa¸sır-dı . Ai snow when fall-PST.3 S DIYE was.surprised b. Ai-wa [ yuki-ga itsu huttano-ka-to ] odoroita. Ai-TOP snow-NOM when fell-Q-TO was.surprised Ai was surprised, thinking “when did it snow?” Here too, the choice of a nominalization vs. a D/T-clause makes an interpretive difference. The examples in (3) imply that Ai believes and was surprised by the true answer to the qu

FAQs

What two embedded-clause types does the paper focus on?

It focuses on nominalizations and clauses introduced by the morphemes to and diye, and studies their interrogative variants.

How do nominalizations and diye/to interrogatives differ in orientation?

Nominalizations give rise to answer-oriented inferences with responsive predicates, whereas diye/to interrogatives are question-oriented and require a linguistic production inference by the attitude holder.

What compositional proposal does the paper make?

It proposes a fragment where attitude predicates take nominalizations as arguments and may impose semantic restrictions, while diye/to-clauses modify and enrich attitude meanings through an inference about linguistic production.

Sep 17, 2026

Uploaded by Violet