Book announced at https://linguistlist.org/issues/36-1783

Title: Empirical issues in syntax and semantics
Subtitle: Selected papers from CSSP 2023
Series Title: Empirically Oriented Theoretical Morphology and Syntax
Publication Year: 2025

Publisher: Language Science Press
http://langsci-press.org
Book URL: https://langsci-press.org/catalog/book/519

Editor(s): Gabriela Bîlbîie, Gerhard Schaden

Reviewer: Chloé Debouzie

SUMMARY

The book contains an edited collection of papers from the “Colloque de Syntaxe et Sémantique à Paris” (CSSP) conference, which took place at the École Normale Supérieure in Paris in 2023. After a short preface by the editors, the 231-page volume comprises eight chapters, each presenting a distinct paper. The chapters are preceded by a dozen-line preface by the editors (Gabriela Bîlbîie and Gerhard Schaden) contextualising the volume as a “curated selection of papers” from CSSP 2023. The book’s aim is to present a “snapshot of contemporary linguistic research” in several areas of syntax and semantics, grounded in empirical studies (corpus and experimental) and branching into various theoretical approaches and generalisations.

Broadly speaking, the papers focus on novel methodologies and theoretical approaches by addressing a diverse range of topics regarding syntax (focus structures, rising declaratives, discourse markers), noun phrases (polysemous nouns, nominal complexes, null vs. pronominal subjects), verbs (subjective attitudes verbs, experiencer–object verbs), within a variety of theoretical approaches (Role and Reference Grammar, frame-semantic representations, Head-Functor Approach, Lewisian conversational scoreboards) and languages (mostly English, but also Mandarin Chinese (Chapters 2 and 3), Romanian, German, and French).

This review presents each chapter in turn.

In the opening chapter entitled “A novel representation of focus structure and non-constituent focus”, the three co-authors, Kata Balogh, Laura Kallmeyer, and Rainer Osswald, propose a new formal grammatical model to represent focus structure and specifically analyse “non-constituent focus” domains, such as the examples in (1).

(1) a. What happened to the book? Pete sold the book.

b. What did Pete do with the book? Pete gave the book to Kate.

In (1), the Information Structure (IS) does not correspond to the syntactic structures. In (1a), the focus is on the NP subject “Pete” and the verb “sold”, which do not constitute a single constituent in the sentence. In (1b) the focus is on the main verb “gave” and the PP “to Kate”, which are two non-consecutive constituents. Within the authors’ approach, focus structure is not determined by the nodes of the constituent structure; rather, IS (which contains focus structure) is captured in a distinct module and is linked to syntax and semantics. Information units (IUs) are the core components of focus structure. Their combinations make the content of focus to account for the representation of “non-constituent focus”, whether the sets of IUs correspond to the focus or non-focus of the IS-domains, and whether the main verb (or predicate) constitutes a part of the focus.

The second chapter, “Primary vs. secondary meaning facets of polysemous nouns” by Long Chen, Laura Kallmeyer, and Rainer Osswald, presents a contrastive analysis of English and Mandarin Chinese polysemous nouns, i.e., nouns with several meaning facets related to each other in copredication constructions. The authors’ approach is grounded in semantic frames as a formalisation of their semantic representations. The authors propose two approaches to model restrictions on copredication: one based on top and bottom feature structures on nodes, and a second one based on a default logic allowing for the removal of defeasible facets and which requires only a single feature structure for each syntactic node.

Within the first approach (developed in Section 3.3), building on Kallmeyer & Osswald’s (2013) and Chen et al.’s (2022) syntax–semantic interface framework, the authors distinguish between “facet-picking” predication, where a predicate picks one facet only and prevents other facets from being activated in subsequent predications, and “facet-addressing” predication, where a predicate addresses one specific facet while maintaining other facets available for further predication.

The second approach (Section 3.4) establishes a distinction between “primary facets”, whose meanings are typically more salient and consistent, and which can be targeted in any type of copredication pattern without a predication blocking any of their facets, and “secondary facets”, modelled as a default attribute that can be retracted in case of conflicting frame constraints, thereby becoming unavailable or blocked in some predication patterns.

The two approaches are developed using empirical evidence from the English Web 2021 corpus (enTenTen21) and the corpus of the Center for Chinese Linguistics of PKU. The authors focus on three types of polysemous nouns in the two languages: “book” (which has two primary facets: physical object and information facet), “food” such as lunch or dinner (also with two primary facets: object and event facet), and “speech” or “lecture” (an event and an information facet, both facets being secondary in English, but primary in Chinese).

In the third chapter, “Unifying modifiers, classifiers and demonstratives”, the three co-authors, Chenyuan Deng, Antonio Machicao y Priemer, and Giuseppe Varaschin, introduce the Head-Functor Approach (HFA) to account for the distributional properties of demonstratives (DEMs) and modifiers (MODs) in Mandarin Chinese nominal complexes that pose a problem within Head-driven Phrase Structure Grammar (HPSG) (Pollard & Sag 1994). The authors analyse numeral–classifier–noun sequences as left-branching Noun Phrases (NPs), whereas the numeral–classifier sequence forms a Classifier Phrase acting as a functor over the noun. The notion of functor refers to “non-head daughters that impose selectional requirements on their sisters” (p. 58) and differs from the head-complement DP and the head-specifier NP approaches by rejecting the traditional distinction between lexical and functional categories and by replacing the division between specifiers and modifiers with a unified account of the properties of classifiers, modifiers and demonstratives. The theory is further exploited to offer a two-way typology of languages, distinguishing languages with head–functor phrases (e.g., Mandarin Chinese) and head–specifier languages (e.g., English, German).

The fourth chapter, “Using belief–perception mismatch to assess the meaning of Subjective Attitude Verbs” by Achille Fusco, Cristiano Chesi, and Valentina Bianchi, focuses on two Subjective Attitude Verbs (SAVs), “find” and “consider”, which typically express first-person judgements or opinions that are relevant to the judgements or opinions expressed in their complement clause. However, “find” and “consider” differ in regard to their doxastic status, as shown by empirical observations presented in Section 2 and developed in the literature review in Section 3. To investigate the asymmetry of the doxastic components, the authors use the “Potential Doxastic Conflict” (PDC) to separate belief from first-person perception. PDC refers to contexts in which subjective perception conflicts with a belief over that experience (Section 4). In an experimental study (Section 5), participants are asked to judge the acceptability of sentences containing “find” or “consider” in either PDC or no-PDC contexts. The results show that in no-PDC situations, both SAVs receive similar ratings, while in PDC contexts, which trigger lower acceptability ratings overall, “find” significantly rates as less acceptable than “consider”. These findings seem to confirm previous analyses by Stephenson (2007) and Muñoz (2019) regarding the lack of doxastic component in “find”, which may explain why “find” is specifically used to express a first-person experience, whereas “consider” seems to require less subjective experience.

In the fifth chapter, entitled “Paradigms and discourse effects of English rising declaratives”, Junseon Hong presents a predictable model for rising declaratives (RDs) in English. RDs are non-canonical: they are syntactically declarative sentences, which are prototypically associated with informative speech acts, but their rising intonation is typically associated with interrogative sentences (i.e., questions used to request information). The author distinguishes two types of RDs: (i) assertive rising declaratives (ARDs), which function as an assertion (i.e., fulfilling an informative function), and (ii) inquisitive rising declaratives (IRDs), which are used for questioning speech acts (i.e., fulfilling an inquisitive function). ARDs typically have a weaker rise in intonation, which indicates that the content is more informative than inquisitive, whereas IRDs typically show a steeper rise, indicating that the content is more inquisitive than informative.

The ARDs are further distinguished between Epistemic Uncertainty ARDs, which question the truth value of the proposition, and Metalinguistic Uncertainty ARDs, which question the contextual relevance of the proposition. Regarding IRDs, the author distinguishes two paradigms: Confirmative IRD or Contradictory IRD, with Mirative IRD as a subtype of the latter. In Confirmative IRDs, the speaker expresses a high degree of certainty regarding the proposition, whereas a high degree of suspicion or disbelief is expressed with Contradictory IRDs. Mirative IRDs are used to express the speaker’s surprise.

Finally, the author argues that the rising intonation projects discourse commitment, instead of indicating some lack of commitment, as previously suggested (Rudin 2018, 2022), and affects both the semantic content of the proposition through the level of steepness of the rise and the discourse components, through the interaction with the context.

The sixth chapter, “Next mention biases predict the choice of null and pronominal subjects” by Fabian Istrate, Ruxandra Ionescu, and Barbara Hemforth, examines the choice between null and pronominal subjects in Romanian, a pro-drop language. First, a corpus study (Section 2) of written and spoken Romanian shows that the salience of the referent (such as agentivity, voice, maleness) increases the likelihood of a null subject in subordinate clauses. Moreover, null subjects are preferred in writing (vs. spoken), and in adverbial temporal subordinate clauses. Pronominal subjects are more prevalent in main clauses. The corpus study shows no preference between pronominal and null subjects with the following variables: animacy of the subject, female antecedents, in subordinate clauses with non-subject antecedents, and in causal subordinates. In sum, semantic and pragmatic factors influence the production and interpretation of null or pronominal subjects.

Next, the authors used an experimental study (Section 3) in the form of a completion task to investigate the production of null vs. pronominal subjects after implicit causality verbs (subject–biased vs. object–biased causality verbs). The experimental study confirmed that null subjects were favoured when referring to more salient and explicit antecedents (in subject and non-subject position).

In conclusion (Section 4), both studies concur in showing that a higher referent predictability significantly leads to null subjects. A complex interaction of semantic and pragmatic factors plays a role in the next mention biases that predict the choice between null and pronominal subjects.

In the seventh chapter, “Discourse markers are not special (but they can be complicated)”, Jacques Jayez examines two broad classes of discourse markers (DMs), connective DMs (developed in Section 3) and Hic et Nunc Particles (HNPs) (in Section 4), using examples of French DMs such as “parce que”, “après”, “donc”, “bien que”. The author focuses on three main functions of DMs: (i) to suggest discourse relations (DRs) between semantic objects, (ii) to refer to speakers’ affective or epistemic states, and (iii) to refer to different interactions between conversational agents. The author establishes general, unifying semantic categories of DMs by distinguishing three semantic kinds: (i) modifiers which contribute to the propositional content, (ii) presupposition (PSP) triggers, and (iii) conventional implicature triggers.

Connective DMs express various DRs (e.g., causality, consequence, concession, opposition) between two semantic objects. Despite typically expressing one meaning at a time, connective DMs can be pragmatically ambiguous: for instance, they can refer to states of affairs (SOA) or belief states (BEL). Connective DMs can be PSP triggers, as they express DRs such as relations with an antecedent and speaker-hearer interactions.

HNPs are closer to interjections but are not all categorised as expressives. They are typically outside the propositional content, but they are usually more radically anchored to the situation of utterance than expressives. HNPs can be semantically vague and are context dependent. They are linguistic signs used to express an internal reaction to external events. As such, they are indexical and depend on context and intonation.

The last chapter of the book, “On the (im-)possibility of reflexive binding into the subject of German experiencer-object verbs” by Simon Masloch, Johanna M. Poppek, and Tibor Kiss, focuses on experiencer-object (EO) verbs, that is, when the experiencer of a psych verb is syntactically realised as an object. The authors focus on the acceptability of reflexive binding into the subject of German EO verbs in an experimental study testing acceptability judgement (using a 5-point scale of naturalness). The study assesses two possible word orders (subject–object (SO), where the subject containing the reflexive precedes the object, and object–subject (OS), where the antecedent precedes the reflexive) and two possible cases (accusative–object or dative–object).

The results of the experimental study show that SO sentences received low ratings overall, yet those with accusative-object verbs received slightly better ratings. OS sentences, overall, were rated as more natural, especially those with dative–object verbs; yet a large number of sentences also received low ratings.

In conclusion, these results show that in German reflexive binding into the subject of EO verbs is only licit in the midfield area if the antecedent is before the reflexive “sich” in the surface structure, and thus c-commands it. This is found with accusative–object and with dative–object EO verbs. The authors were surprised to find low levels of acceptability overall and high levels of variation among the participants, requiring further research regarding predicate-based binding theories.

EVALUATION

Throughout the eight chapters of this volume, readers are invited to engage with a variety of methodological and theoretical topics in syntax and semantics.

Each chapter provides a well-researched and well-presented novel approach to the issues addressed. Moreover, all are well structured: each introduction starts from an existing issue raised either from empirical observation (e.g., non-constituent focus in Chapter 1, asymmetry in copredication in Chapter 2) or from ongoing debate (e.g., discourse markers in Chapter 7, backward binding into the subject of an EO verb in Chapter 8). After exposing the issues at stake, the authors clearly state their aims and scope. All introductions (except in Chapter 2) end with an outline of the paper, which usefully guides the reader. Typically, the second section (or third in Chapters 4 and 5) contains the theoretical background or state of the art. The chapters are all grounded in a strong theoretical framework (Role and Reference Grammar (Van Valin & LaPolla 1997, Van Valin 2005), frame–semantic representations, Tree Adjoining Grammar and Tree-Wrapping Grammar (TWG), Head-driven Phrase Structure Grammar, conversational scoreboards) that is well presented and explained, with references to further readings for anyone not well-versed in these approaches. For instance, in Chapter 2, where semantic frames are used, in a footnote (p. 38), the authors refer the reader to Kallmeyer & Osswald (2013) and Chen et al. (2022) in case the reader wants to get more acquainted with frames, logic of attribute-value descriptions and formulas (which I did, and it helped me better understand the rest of the chapter).

The papers based on corpus or experimental studies (i.e., Chapters 2, 4, 6 and 8) provide a clear description of their methodology, followed by the results of the analyses, then a discussion, and a conclusion with the mention of further avenues of research. The data in the studies in Chapters 4, 6 and 8 are analysed using advanced statistical methods (logistic regressions in Chapter 4 and 6, a Bayesian model in Chapter 8). In these chapters, the results are presented in various tables and figures (some of them in colour, as in Chapter 6), which greatly help visualising the findings.

The other chapters rely primarily on the analysis of a few phrases (such as “a blue book” “sold” “in London” in Chapter 1, “She’s home?” as an example of rising declarative in Chapter 5, and a selection of half a dozen French discourse markers analysed in Chapter 7). The figures in Chapter 1 contain syntactic trees used in TWG and RRG. Chapter 3 contains fifteen figures to represent the different analyses (NP/DP/Head-Functor Approach). Chapter 5 uses tables from the conversational scoreboard model to provide formal analyses of the different types of rising declaratives, while Chapter 7 uses Dialogue Game Boards to represent the different discourse markers.

Overall, the book is a valuable resource for researchers in these topics as it presents novel, present-date and state-of-the-art research. The authors provide their observations, analyses and new findings in various areas that are open for further research, which are clearly stated in each chapter. The use of new frameworks and approaches helps better explain some interesting, yet puzzling linguistic phenomena. Some studies confirm existing findings, while others contradict (at least partially) existing theories (e.g., Chapters 4 and 8). Chapter 6, on null vs. pronominal subjects in Romanian, offers valuable insight into cross-linguistic analyses of and comparisons with other pro-drop languages such as Italian and Spanish. Most contributions require a high degree of familiarity with specialised issues and techniques.

The volume is not without limitations, some of which being seemingly unavoidable in a volume which gathers a selection of papers from a conference proceedings covering a wide variety of issues. The book lacks both an introductory and a conclusive chapter, maybe because it does not try to present a unified view of the field. Each chapter deals with a specific issue in a highly specialised manner, which makes it difficult to evaluate the readability of the volume as a whole or to determine precisely whom it is targeted at. Although key concepts are consistently and rigorously defined and explained, it seems that each chapter is addressed to specialists in each theory (e.g., frame-semantic representations, Head-driven Phrase Structure Grammar, conversational scoreboards). As a result, students and early-career scholars may struggle with some chapters if they are not experts in these theoretical frameworks. Moreover, the chapters based on experimental studies use advanced statistical calculations, which may not be easily accessible to everyone (however, I found the figures very useful to make sense of the results when I did not understand the maths, as in Chapter 6). Overall, the content and discourse of the volume is best suited for a specialised audience, as previous knowledge of the terminology and concepts is required (for instance to understand the conversational scoreboards in Chapters 5 and 7).

Finally, the book is well produced on high-quality paper with coloured figures, comprehensive tables and graphs, and a clear layout; yet a few misprints were noted. Most regrettably, the name of the conference itself is misspelt (the -e at the end of “syntaxe” in French is missing in the preface and on the back cover). Another typographic error occurs on p. 58, where the word “modififiers” appears instead of “modifiers”. A couple of careless errors can be found: in the first chapter, p. 9, the text mentions the adjective “red” while the Figures 3 to 6 use the adjective “blue”; on page 185, we read “The fanttype function (for ‘fetch antecedent of type type’) function provides a black box for general antecedent recovery procedures”, with “function” unnecessarily repeated before the brackets. Lastly, in the second chapter, Section 2 lists “three facets” (“event”, “information”, and “object”), but then Section 2.1 uses the label “food” as a facet in the subtype “event food” to describe nouns referring to meals such as “lunch”, “buffet”, and “dinner”. Reading the description in Section 2.1, we understand that meals are types of “event object”, as they have an event and an object facet. The authors do not explain why the label “event food” is used to describe meals in Section 2.1, and not “event object”, which we understand from the explanation on page 31. As the chapter is already quite technical for anyone not familiar with this approach, this confusion between “event food” or “event object” is unfortunate.

REFERENCES

Chen, Long, Laura Kallmeyer & Rainer Osswald. 2022. A frame-based model of inherent polysemy, copredication and argument coercion. In Michael Zock, Emmanuele Chersoni, Yu-Yin Hsu & Enrico Santus (eds.), Proceedings of the Workshop on Cognitive Aspects of the Lexicon, 58–67.

Kallmeyer, Laura & Rainer Osswald. 2013. Syntax-driven semantic frame composition in lexicalized Tree Adjoining Grammars. Journal of Language Modelling 1(2). 267–330. DOI: 10.15398/jlm.v1i2.61.

Muñoz, Patrick Joseph. 2019. On tongues: The grammar of experiential evaluation. The University of Chicago. (Doctoral dissertation).

Pollard, Carl J. & Ivan A. Sag. 1994. Head-Driven Phrase Structure Grammar. Chicago: University of Chicago Press.

Rudin, Deniz. 2018. Rising above commitment. UC Santa Cruz. (Doctoral dissertation).

Rudin, Deniz. 2022. Intonational commitments. Journal of Semantics 39(2). 339–383. DOI: 10.1093/jos/ffac002.

Stephenson, Tamina. 2007. Towards a theory of subjective meaning. MIT. (Doctoral dissertation).

Van Valin, Robert D., Jr. 2005. Exploring the syntax-semantics interface. Cambridge: Cambridge University Press. DOI: 10.1017/cbo9780511610578.

Van Valin, Robert D., Jr. & Randy J. LaPolla. 1997. Syntax: Structure, meaning, and function. Cambridge: Cambridge University Press.

ABOUT THE REVIEWER

Chloé Debouzie holds a PhD in Linguistics from Université Lumière Lyon 2, France. Her doctoral work investigates morphological competition in the formation of new verbs in present-day English. Her research interests include derivational morphology, semantics, onomasiological approaches, morphological competition, contrastive studies (French–English), sociolinguistics, and psycholinguistics. She has taught widely across linguistics, applied linguistics, TESOL, translation, and French as a foreign language.