Predicting Declension Class from Form and Meaning

The noun lexica of many natural languages are divided into several declension classes with characteristic morphological properties. Class membership is far from deterministic, but the phonological form of a noun and/or its meaning can often provide imperfect clues. Here, we investigate the strength of those clues. More specifically, we operationalize this by measuring how much information, in bits, we can glean about declension class from knowing the form and/or meaning of nouns. We know that form and meaning are often also indicative of grammatical gender---which, as we quantitatively verify, can itself share information with declension class---so we also control for gender. We find for two Indo-European languages (Czech and German) that form and meaning respectively share significant amounts of information with class (and contribute additional information above and beyond gender). The three-way interaction between class, form, and meaning (given gender) is also significant. Our study is important for two reasons: First, we introduce a new method that provides additional quantitative support for a classic linguistic finding that form and meaning are relevant for the classification of nouns into declensions. Secondly, we show not only that individual declensions classes vary in the strength of their clues within a language, but also that these variations themselves vary across languages.

[1]  Mark Aronoff,et al.  Noun classes in Arapesh , 1992 .

[2]  M. Wertheimer,et al.  The relation between the sound of a word and its meaning. , 1958, The American journal of psychology.

[3]  Thomas M. Cover,et al.  Elements of Information Theory , 2005 .

[4]  Welch Bl THE GENERALIZATION OF ‘STUDENT'S’ PROBLEM WHEN SEVERAL DIFFERENT POPULATION VARLANCES ARE INVOLVED , 1947 .

[5]  Josef Vachek Some Remarks on Writing and Phonetic Transcription , 1945 .

[6]  Richard Sproat,et al.  Book Reviews: A Computational Theory of Writing Systems , 2006, CL.

[7]  M. Wertheimer,et al.  Some Physiognomic Aspects of Naming, or, Maluma and Takete Revisited , 1964, Perceptual and motor skills.

[8]  Greville G. Corbett,et al.  Gender assignment: a typology and a model , 2000 .

[9]  Eva Maria Vecchi,et al.  (Linear) Maps of the Impossible: Capturing Semantic Anomalies in Distributional Space , 2011 .

[10]  Ryan Cotterell,et al.  Meaning to Form: Measuring Systematicity as Information , 2019, ACL.

[11]  H. Theil On the Estimation of Relationships Involving Qualitative Variables , 1970, American Journal of Sociology.

[12]  J. Volín,et al.  Phonological spelling errors among dyslexic children learning a transparent orthography: the case of Czech. , 2001, Dyslexia.

[13]  Zellig S. Harris,et al.  Distributional Structure , 1954 .

[14]  Sebastian Kürschner,et al.  The interaction of gender and declension in Germanic languages , 2011 .

[15]  Gereon Müller,et al.  Class features as probes , 2005 .

[16]  Olivier Bonami,et al.  Joint predictiveness in inflectional paradigms , 2016 .

[17]  Zdeněk Matějček Reading in Czech. Part I: Tests of reading in a phonetically highly consistent spelling system , 1998 .

[18]  Y. Benjamini,et al.  Controlling the false discovery rate: a practical and powerful approach to multiple testing , 1995 .

[19]  E. Sapir A study in phonetic symbolism. , 1929 .

[20]  Marcus Hutter,et al.  Distribution of Mutual Information , 2001, NIPS.

[21]  Richard Sproat,et al.  The Consistency of the Orthographically Relevant Level in Dutch , 2002 .

[22]  James W. Harris,et al.  The form classes of Spanish substantives , 1992 .

[23]  Petr Sojka,et al.  Software Framework for Topic Modelling with Large Corpora , 2010 .

[24]  N. Schiller,et al.  Semantic gender assignment regularities in German , 2004, Brain and Language.

[25]  Jonathan W. Pillow,et al.  Bayesian and Quasi-Bayesian Estimators for Mutual Information from Discrete Data , 2013, Entropy.

[26]  Damaris Nübling WAS TUN MIT FLEXIONSKLASSEN? DEKLINATIONSKLASSEN UND IHR WANDEL IM DEUTSCHEN UND SEINEN DIALEKTEN , 2008 .

[27]  Ga Miller,et al.  Note on the bias of information estimates , 1955 .

[28]  Ryan Cotterell,et al.  On the Relationships Between the Grammatical Genders of Inanimate Nouns and Their Co-Occurring Adjectives and Verbs , 2020, Transactions of the Association for Computational Linguistics.

[29]  Olivier Bonami,et al.  A comprehensive view on inflectional classification , 2016 .

[30]  E. Newport,et al.  Learning at a distance I. Statistical learning of non-adjacent dependencies , 2004, Cognitive Psychology.

[31]  Roy Schwartz,et al.  How Well Do Distributional Models Capture Different Types of Semantic Knowledge? , 2015, ACL.

[32]  Liam Paninski,et al.  Estimation of Entropy and Mutual Information , 2003, Neural Computation.

[33]  Morten H. Christiansen,et al.  Phonological typicality influences on-line sentence comprehension , 2006, Proceedings of the National Academy of Sciences.

[34]  Richard Sproat,et al.  The Relation of Writing to Spoken Language , 2002 .

[35]  Wolfgang Ullrich Wurzel,et al.  Inflectional Morphology and Naturalness , 1989 .

[36]  Robert L. Mercer,et al.  An Estimate of an Upper Bound for the Entropy of English , 1992, CL.

[37]  Morten H. Christiansen,et al.  Arbitrariness, Iconicity, and Systematicity in Language , 2015, Trends in Cognitive Sciences.

[38]  É. Benveniste,et al.  Origines de la formation des noms en indo-européen , 1938 .

[39]  Jeffrey Dean,et al.  Efficient Estimation of Word Representations in Vector Space , 2013, ICLR.

[40]  Wido La Heij,et al.  Phonological facilitation of grammatical gender retrieval , 2004 .

[41]  Richard M. Hogg Orthography and Phonology , 2011 .

[42]  D. Lakens,et al.  Why Psychologists Should by Default Use Welch's t-test Instead of Student's t-test with Unequal Group Sizes , 2017 .

[43]  I. Watson,et al.  In the Beginning Was the Word , 2009 .

[44]  Ryan Cotterell,et al.  UniMorph 3.0: Universal Morphology , 2018, LREC.

[45]  James W. Harris,et al.  The exponence of gender in Spanish , 1991 .

[46]  Student,et al.  THE PROBABLE ERROR OF A MEAN , 1908 .

[47]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[48]  Patrick Pantel,et al.  From Frequency to Meaning: Vector Space Models of Semantics , 2010, J. Artif. Intell. Res..

[49]  Yong-Yeol Ahn,et al.  Element-centric clustering comparison unifies overlaps and hierarchy , 2017, Scientific Reports.

[50]  Alec Marantz,et al.  Revisiting form typicality of nouns and verbs A usage-based approach , 2017 .

[51]  Jasper Snoek,et al.  Practical Bayesian Optimization of Machine Learning Algorithms , 2012, NIPS.

[52]  Morten H. Christiansen,et al.  The phonological-distributional coherence hypothesis: Cross-linguistic evidence in language acquisition , 2007, Cognitive Psychology.

[53]  Mirella Lapata,et al.  Composition in Distributional Models of Semantics , 2010, Cogn. Sci..

[54]  Robert M. Gonyea,et al.  Learning at a Distance : , 2009 .

[55]  Patricia J. Brooks,et al.  Acquisition of Gender-like Noun Subclasses in an Artificial Language: The Contribution of Phonological Markers to Learning , 1993 .

[56]  Janet L. McDonald,et al.  Properties of Phonological Markers That Affect the Acquisition of Gender-Like Subclasses☆☆☆★ , 1998 .

[57]  Jürgen Schmidhuber,et al.  LSTM can Solve Hard Long Time Lag Problems , 1996, NIPS.

[58]  Stephen Clark,et al.  Vector Space Models of Lexical Meaning , 2015 .

[59]  A. Carstairs-McCarthy Inflection classes, gender, and the principle of contrast , 1994 .

[60]  Annette D'Onofrio,et al.  Phonetic Detail and Dimensionality in Sound-shape Correspondences: Refining the Bouba-Kiki Paradigm , 2014 .

[61]  Bernd Wiese,et al.  Warum Flexionsklassen? Über die deutsche Substantivdeklination , 2000 .

[62]  James P. Blevins,et al.  Parts and wholes: Implicative patterns in inflectional paradigms , 2009 .

[63]  Lise M. Dobrin,et al.  The morphosyntactic reality of phonological form , 1998 .

[64]  Jonathan W. Pillow,et al.  Bayesian entropy estimation for countable discrete distributions , 2013, J. Mach. Learn. Res..

[65]  Xuanjing Huang,et al.  Investigating Language Universal and Specific Properties in Word Embeddings , 2016, ACL.

[66]  Ryan Cotterell,et al.  Quantifying the Semantic Core of Gender Systems , 2019, EMNLP.

[67]  E Miles,et al.  Dyslexia may show a different face in different languages. , 2000, Dyslexia.

[68]  Simon Kirby,et al.  How arbitrary is language? , 2014, Philosophical Transactions of the Royal Society B: Biological Sciences.

[69]  Jessica Maye,et al.  Infant sensitivity to distributional information can affect phonetic discrimination , 2002, Cognition.

[70]  Alec Marantz,et al.  Some key features of distributed morphology , 1994 .

[71]  Gemma Boleda,et al.  Distributional Semantics in Use , 2015, LSDSem@EMNLP.

[72]  Robert Malouf,et al.  Morphological Organization: The Low Conditional Entropy Conjecture , 2013 .

[73]  Mark Dingemanse,et al.  Redrawing the margins of language: Lessons from research on ideophones , 2018 .

[74]  D. Maurer,et al.  The shape of boubas: sound-shape correspondences in toddlers and adults. , 2006, Developmental science.