Schäfer M. The semantic transparency of English compound nouns

Berlin: Language Science Press, 2018. — xiv, 402 p. — (Morphological Investigations 3). — ISBN 978-3-96110-030-9.What is semantic transparency, why is it important, and which factors play a role in its assessment? This work approaches these questions by investigating English compound nouns. The first part of the book gives an overview of semantic transparency in the analysis of compound nouns, discussing its role in models of morphological processing and differentiating it from related notions. After a chapter on the semantic analysis of complex nominals, it closes with a chapter on previous attempts to model semantic transparency. The second part introduces new empirical work on semantic transparency, introducing two different sets of statistical models for compound transparency. In particular, two semantic factors were explored: the semantic relations holding between compound constituents and the role of different readings of the constituents and the whole compound, operationalized in terms of meaning shifts and in terms of the distribution of specifc readings across constituent families.
All semantic annotations used in the book are freely available.Acknowledgments.
Abbreviations.
Introduction.
A first notion of semantic transparency.
Compounds and complex nominals.
Aims and Goals.
Structure.
Semantic transparency in psycholinguistics.
Structure and lexical access.
Morpheme-based models.
Models without morphemes.
Models of conceptual combination.
Conclusion: the different models.
Measuring semantic transparency.
Establishing semantic transparency.
Summary: measuring semantic transparency.
Psycholinguistic studies.
Priming paradigms.
Eye movement studies.
Experiments targeting conceptual combination.
Overview: experimental traces of semantic transparency.
Conclusion: experimental traces of semantic transparency.Related phenomena and notions.
Semantic transparency reflected in other linguistic phenomena.
Semantic transparency and outbound anaphora.
Semantic transparency and compound stress.
Conclusion: semantic transparency and other phenomena.
Other measures and notions.
Quantitative measures.
Semantic overlap.
Compositionality and literality.
Semantic transparency as one dimension of idiomaticity.
Semantic transparency and productivity.
Transparency in other domains.
Phonological transparency.
Orthographic transparency.The semantic analysis of compounds.
Set-theoretic approaches.
Intersective modification.
Subsective modification.
Non-subsective modification.
Problems for a set-theoretic classification of adjectives.
Relation-based approaches: the semantics of compounds.
Levi (1978).
Levi’s complex nominals.
Levi’s recoverably deletable predicates.
Predicate nominalization.
Evaluating Levi’s approach.
Conclusion: the enduring appeal of Levi’s system.
Fanselow (1981).
Compounds involving relational nouns.
Determinative compounds.
Evaluating Fanselow’s approach.
Mixed approaches.
Pustejovsky (1995).
Extending the analysis to compounds 1: Jackendoff (2010).
Extending the analysis to compounds 2: Asher (2011).
Approaches using underspecification.Previous models.
Distributional semantics and word space models.
The basics of distributional semantics: a toy example.
Design decisions.
Two implementations: LSA and HAL.
Conclusion.
Reddy, McCarthy & Manandhar (2011).
Selection procedure.
Reddy et al.’s human judgment data.
Reddy et al.’s distributional semantics models.
Pham and Baayen (2013).
Informativity based measures.
Pham and Baayen: compound selection and variable coding.
Study 3: transparency rating experiment.
Marelli et al. (2015).
Experiment 1: connotations.
Experiment 2: semantic processing.Piloting semantic factors.The Reddy et al. data: a descriptive overview.
Linguistic characterization of the selected compounds.
Descriptive overview of the rating data.
Bell & Schäfer (2013).
Subsetting the Reddy et al. dataset.
Semantic annotation of the compounds.
Annotation results.
Bell and Schäfer (2013): the models.
Bell & Schäfer (2013) revisited.
Classic model criticism.
Linear mixed effects modeling.
The role of the meaning shifts.
Conclusion and consequences.
Compound family based models.
Semantic relations relative to constituent families.
Gagné and Shoben.
Criticism and a corpus-based re-implementation.
Relational distributions in other studies.
Conclusion: relations relative to families.
Assessing the role of constituent meanings.
A database of compound families.
Initial families from the BNC.
Adding items from CELEX.
Usenet frequencies.
Further post-processing.
Semantic coding.
Coding the semantic relations.
Coding the constituent senses.
Variables and predictions.
Variables derived from the semantic coding.
Further explanatory variables.
Tabular overview of the explanatory variables.
Restricting the target dataset.
Predicting semantic transparency.
The models from Bell & Schäfer.
N1 transparency.
N2 transparency.
Whole compound transparency.
The 2016 models: discussion and conclusion.
Re-modeling Bell & Schäfer (2016).
New models for constituent transparency.
New models for compound transparency.
Conclusion: re-modeling Bell & Schäfer (2016).Summary and outlook.Outlook.
Appendix A: Semantic coding for Bell & Schäfer (2013).
Relations.
Shifts.
Appendix B: Semantic coding for Bell & Schäfer (2016).
N1 families.
N2 families.
Appendix C: Multiple readings and the 2016 coding.
Appendix D: Corpus and dictionary sources.
Corpus identifiers.
Online dictionaries.
References.
Index.
Name index.
Language index.
Subject index.