Lexical Relations Research Articles

Αρκάς: μια πραγματολογική προσέγγιση ενός ελληνικού κόμικ

The paper analyzes certain episodes taken from Arkas comic series: «Με τον κηδεμόνα σου» (Me ton kidemona sou) and «Σκέπτομαι κι ας μην υπάρχω» (Skeptome ki as min iparxo) in terms of Pragmatics. It explores the way language is used to create humor and jokes in one of the most famous Greek comic series. Sense relations, lexical relations and sentence relations are tracked down. In particular, an attempt to test the applicability of Grice’s conversational Maxims, implicatures-explicatures, as well as, the Cooperation Principle is made. In addition, cases of polysemy are discussed. By examining the bubbles and the various symbols, the aim is to identify specific patterns that lead to a humorous outcome. Along with the linguistic elements, reference is made to the aesthetics that surround the sketches. Overall, the paper analyses specific episodes treating language as a means by which we do things.

어휘 관계 및 문맥 정보 기반의 도메인 감성사전 자동 구축 방안 연구

감성사전은 감성 어휘들에 대한 집합으로 각 어휘들에 대한 감성의 극성이 부여되어 있으며, 감성 분석(Sentiment Analysis)을 위한 기초 자료로 활용된다. 하지만 이와 같은 감성 어휘들은 특정 도메인에 따라 극성이 역전되거나 유실될 수도 있으며 의존적인 감성 어휘가 존재할 수 있다. 예를 들면, 일반적으로 ‘잘 잤다’라는 단어는 긍정의 극성을 보이지만, 영화 도메인에서는 그 의미가 부정으로 쓰인다. 그렇기 때문에 감성사전은 분석하고자 하는 도메인의 특징이 반영되어 있어야 하며 도메인에 따라 알맞은 감성사전이 구축되고 활용되어야 한다. 이와 같은 문제를 해결하기 위해 현재 도메인 감성사전을 자동으로 구축하는 다양한 연구들이 나왔지만, 인간의 개입, 문맥적 요소 미고려, 지역적인 정보 반영 등이라는 문제점을 지니고 있다. 본 연구에서는 이와 같은 문제를 해결하기 위해 한국어 범용 감성사전인 ‘KNU 한국어 감성사전’과 글로브 벡터 그리고 접속사 관계를 활용하여, 특정 도메인의 전역적인 감성 정보와 문맥적 특징을 충분히 반영한 도메인 감성사전 구축 방안을 제안한다.

A Study on the Constitution and Structural Properties of Collocation in Russian

Many linguists have proposed various definition and properties of collocations, but there are no unified criteria for the classification of collocations. Sinclair(1996;1998:1999)’s definition of collocation is considered universal, and therefore has applied to not only English but many other languages. Following his theory, many linguists have attempted to treat collocations according to the statistical methods. On the other hand, many linguistics who support machine translation followed to treat collocation by Lexical Function that devised Melcu’k(1980;2007). However, formal and abstract criterions excluding lexical semantics can’t describe systematically all types of collocations. Each word in utterance has a distinctive word that semantically related to extralinguistic situation and reality. E. Coseriu, V. Vinogradov, and N. Arutyunova suggest the relation between word and reality. Through this process, V. Gak claimed the semantic agreement between words in utterance. From this point of view, this article examines some types of collocations and clarify for the constitution and semantic basic properties of collocations. Collocations are formed through word combination rule. In additions, semantic similar collocations make semantic syntagmatic types. Lexical close syntagmatic relations are classified of semantic type. According to this idea, this article shows analysis of collocation of adjective тёмный, using Russian dictionary and corpus.

Multi-SimLex: A Large-Scale Evaluation of Multilingual and Crosslingual Lexical Semantic Similarity

We introduce Multi-SimLex, a large-scale lexical resource and evaluation benchmark covering data sets for 12 typologically diverse languages, including major languages (e.g., Mandarin Chinese, Spanish, Russian) as well as less-resourced ones (e.g., Welsh, Kiswahili). Each language data set is annotated for the lexical relation of semantic similarity and contains 1,888 semantically aligned concept pairs, providing a representative coverage of word classes (nouns, verbs, adjectives, adverbs), frequency ranks, similarity intervals, lexical fields, and concreteness levels. Additionally, owing to the alignment of concepts across languages, we provide a suite of 66 crosslingual semantic similarity data sets. Because of its extensive size and language coverage, Multi-SimLex provides entirely novel opportunities for experimental evaluation and analysis. On its monolingual and crosslingual benchmarks, we evaluate and analyze a wide array of recent state-of-the-art monolingual and crosslingual representation models, including static and contextualized word embeddings (such as fastText, monolingual and multilingual BERT, XLM), externally informed lexical representations, as well as fully unsupervised and (weakly) supervised crosslingual word embeddings. We also present a step-by-step data set creation protocol for creating consistent, Multi-Simlex–style resources for additional languages. We make these contributions—the public release of Multi-SimLex data sets, their creation protocol, strong baseline results, and in-depth analyses which can be helpful in guiding future developments in multilingual lexical semantics and representation learning—available via a Web site that will encourage community effort in further expansion of Multi-Simlex to many more languages. Such a large-scale semantic resource could inspire significant further advances in NLP across languages.

A comprehensive review on feature set used for anaphora resolution

In linguistics, the Anaphora Resolution (AR) is the method of identifying the antecedent for anaphora. In simple terms, this is the problem that helps to solve what the expression referring to a referent refers to. It is considered to be one of the tedious tasks in Natural Language Processing (NLP). AR’s burgeoning popularity among researchers is attributable to its strong relevance to machine translation, text summarization, chatbot, question answering, and many others. This paper presents a review of AR approaches based on significant features utilized to perform this task and presents the evaluation metrics for this field. The feature is a relevant term related to AR that provides vital information regarding anaphor, antecedent, and relation between them. In this context, features represent the lexical, syntactical, semantical, and positional relationship between anaphor and its possible candidate antecedent. The performance of the Anaphora resolution system is profoundly dependent on the features used in the AR system. Hence, the selection of features for the AR system is highly significant. The main emphasis is to provide an overview of the various features needed to extract both the Anaphora and the Antecedent, respectively, used in different AR systems, present in literature. It is observed that syntactical information enhances the correctness of determining the properties for the existence of an anaphor and antecedent identification. Nowadays the trend is changing from hand-crafted feature dependent methods to deep learning approaches which try to learn feature representation. The performance of deep learning is progressing due to the accessibility of additional data and more powerful computing resources. This survey will provide the state-of art for the better understanding of solving AR problem from the feature selection perspective. The findings of this survey are useful to provide valuable insight into present trends and are helpful for researchers who are looking for developing AR system within given constraints.

A Link Prediction Approach for Accurately Mapping a Large-scale Arabic Lexical Resource to English WordNet

Success of Natural Language Processing (NLP) models, just like all advanced machine learning models, rely heavily on large -scale lexical resources. For English, English WordNet (EWN) is a leading example of a large-scale resource that has enabled advances in Natural Language Understanding (NLU) tasks such as word sense disambiguation, question answering, sentiment analysis, and emotion recognition. EWN includes sets of cognitive synonyms called synsets, which are interlinked by means of conceptual-semantic and lexical relations and where each synset expresses a distinct concept. However, other languages are still lagging behind in having large-scale and rich lexical resources similar to EWN. In this article, we focus on enabling the development of such resources for Arabic. While there have been efforts in developing an Arabic WordNet (AWN), the current version of AWN has its limitations in size and in lacking transliteration standards, which are important for compatibility with Arabic NLP tools. Previous efforts for extending AWN resulted in a lexicon, called ArSenL, that overcame the size and the transliteration standard limitation but was limited in accuracy due to the heuristic approach that only considered surface matching between the English definitions from the Standard Arabic Morphological Analyzer (SAMA) and EWN synset terms, and that resulted in inaccurate mapping of Arabic lemmas to EWN’s synsets. Furthermore, there has been limited exploration of other expansion methods due to expensive manual validation needed. To address these limitations of simultaneously having large-scale size with high accuracy and standard representations, the mapping problem is formulated as a link prediction problem between a large-scale Arabic lexicon and EWN, where a word in one lexicon is linked to a word in another lexicon if the two words are semantically related. We use a semi-supervised approach to create a training dataset by finding common terms in the large-scale Arabic resource and AWN. This set of data becomes implicitly linked to EWN and can be used for training and evaluating prediction models. We propose the use of a two-step Boosting method, where the first step aims at linking English translations of SAMA’s terms to EWN’s synsets. The second step uses surface similarity between SAMA’s glosses and EWN’s synsets. The method results in a new large-scale Arabic lexicon that we call ArSenL 2.0 as a sequel to the previously developed sentiment lexicon ArSenL. A comprehensive study covering both intrinsic and extrinsic evaluations shows the superiority of the method compared to several baseline and state-of-the-art link prediction methods. Compared to previously developed ArSenL, ArSenL 2.0 included a larger set of sentimentally charged adjectives and verbs. It also showed higher linking accuracy on the ground truth data compared to previous ArSenL. For extrinsic evaluation, ArSenL 2.0 was used for sentiment analysis and showed, here, too, higher accuracy compared to previous ArSenL.

Automatic Detection of Lexical Functions in Context

In this article, we explore the concept of lexical functions as a formalism to represent recurrent lexical, semantic, and syntactic relations among words. A lexical function takes a word as input and outputs a set of words related to the input in a certain way on the paradigmatic or syntagmatic level. For example, the syntagmatic lexical function termed Oper1 takes the noun decision as input and outputs the verb make with the semantics of ‘Agent realizes the action denoted by the noun’, so Oper1 captures the relation between the noun and the verb in the collocation make a decision. The numeric part of the Oper1 notation reflects two facts: first, the action of make is performed by the agent, which is viewed as the first semantic actant (or argument) in the Tesniere’s model, and second, the syntactic function of the word denoting the agent in utterances with make a decision is subject. In general, lexical functions represent common semantic and syntactic patterns typical for certain word classes and can aid in many tasks of natural language processing, lexical and syntactic disambiguation being the most fundamental one among them. In this article, we review various paradigmatic and syntagmatic lexical functions, their application and identification in natural language processing.

Word embedding in teaching research writing

In research writing as a scientific basis in education, some mistakes related to choosing appropriate term or definition take place. This research discusses the solving such problem by using word embedding in the Research Writing discipline. This is a word representation form, where one word has a vector and its coordinates. The words with close meaning have similar direction, showing lexical compatibility. To calculate lexical relations, the cosine of the angle between two words’ vectors are considered. Value of highly compatible word combinations is equal to 1. On the other hand, lexically incompatible words should approximately have value -1. To test the system the text of the Constitution of the Republic of Kazakhstan was used. Particularly, words which are not related to meaning of article of the Constitution were inserted, and the system had to identify that inserted words. The system for some words showed high accuracy, however some words showed low accuracy. It is suggested that such factor was because even inserted words were not related in meaning, they could be lexically compatible with their neigh bors. This research is carried out within the framework of the Ministry of Education and Science of Republic of Kazakhstan grant project “Developing and implementing the innovative competency-based model of multilingual IT specialist in the course of national education system modernization”.

Evento Coacción de la Construcción ¨Nombre + posposición temporal¨ en chino y estudio comparativo con el español: una perspectiva de la Estructura de Qualia extendida

El presente trabajo, basado en la Teoría del Lexicón generativo (TLG) (Pustejovsky, 1995), procura aportar una interpretación detallada y profunda de los fenómenos del Tipo Discordancia (Type Mismatch) y el Tipo Transición (Type shift) en la construcción ¨Nombre + Posposición temporal¨. Para ello, se realizará un análisis de los casos más representativos, en este caso, las posposiciones temporales后hou/之后zhihou/以后yihou(HOU). Se tomarán a su vez como referencia, sus equivalentes en español, como es el conector adverbial temporal después de (Vanderschueren 2013), con el fin de indagar cómo opera el mecanismo generativo semántico Evento coacción en chino. La construcción canónica (Chomsky 1986) ¨Nombre + posposición temporal¨ tiene la función de localizar un evento dentro de un eje temporal, donde se requiere que el complemento sea un sustantivo del tiempo, como: un punto de tiempo (el momento del evento), un punto de tiempo de referencia (el momento de referencia), o bien, un punto de tiempo de la enunciación (el momento de la enunciación). Cuando en la posición del complemento se presenta un nombre eventivo o no eventivo, la posposición temporal exigirá que dicho nombre sea o se convierta en un delimitado, del cual se pueda derivar un punto en el tiempo. En tal proceso del tipo transición, el nombre eventivo atípico estará coaccionado por su determinante, y pasará de nombre eventivo no delimitado a delimitado. El nombre no eventivo, está directamente coaccionado por la posposición temporal, y pasará de nombre no eventivo, a eventivo delimitado, para satisfacer el requisito de las posposiciones temporales como Hou. Sin embargo, no todos los nombres no eventivos pueden usarse en la construcción coaccionada ¨Nombre + posposición temporal¨. Solamente los nombres secuenciales (Ma 1991) pueden aparecer en la posición del complemento en dicha construcción, gracias a que, el rasgo sub-léxico [Temporalidad contextual] ([TC]) (Liu 2004) les posibilita a obtener la interpretación eventiva. Finalmente, el presente estudio deduce una formal lógica del proceso sintáctico en chino, en el que funciona la coerción de tipos. Dicha lógica forma favorecería una perspectiva dinámica de las relaciones léxicas entre los diferentes tipos de nombres y posposiciones temporales, y a su vez, permite ver claramente las etapas en las cuáles se ha experimentado el Tipo Transición. Los resultados permiten observar, que la propuesta de la TLG puede explicar cómo opera el mecanismo generativo Evento coacción en la construcción ¨Nombre + posposición temporal¨, no obstante, es necesario ampliar la perspectiva del análisis de la Estructura Qualia (EQ) a la Estructura Qualia Extentida (EQE) (Yuan 2014).

Applicability of Wordnet Architecture in Lexical Borrowing Studies

AbstractThe aim of the paper is to present a new, wordnet-based method of studying lexical borrowing. In contrast to the two-dimensional (micro- and macrostructure) organization of traditional dictionaries, describing each unit in isolation, wordnets introduce an additional dimension of inter-word relations in the entire lexicon. The ability to connect wordnets cross-linguistically allows in turn for juxtaposing, comparing, and visualizing a broad spectrum of lexical relations between loanwords within the recipient language and their cognates in the source language for which the term cross-linguistic mapping has been used. The method was applied to a wordnet-based, yet differently structured network of Polish loanwords in Yiddish. This allowed to supplement the traditional typology of transferred words by introducing the notion of ‘rootedness’, that is the degree of their interconnectedness in the source and recipient systems. This in turn has provided a promising means for distinguishing inherited from acquired vocabulary.

Lexico-Semantic Analysis of News Reports in Okexone.com and Tribunnews.com about World War III Issues

Although there have been various semantics studies of news reports, so far the use of lexico-semantics approach in analysing news reports has not been common. The aim of this research is analyze the lexical relations and meaning properties in news reports. It was carried out descriptive-qualitatively. Eight news reports in Okezone.com and eight news reports in Tribunnews.com were studied by using Saeed’s (2009) theory of lexical relations and Siregar (1992) and Leech’s (19981) theory of meaning properties. After completing the analysis, it was found that news news reports in Tribunnews.com are slightly different with the news reports in Okezone.com. There are six types of lexical relations in Tribunnews.com: hyponymy, polysemy, homonymy, synonymy, antonymy, and metonymy, while there are only five types of lexical relations in Okezone.com: hyponymy, polysemy, synonymy, antonymy, and metonymy. In addition, there are five types of meaning properties in Tribunnews.com: meaningfulness, ambiguity, redundancy, anomaly, and contradictory, while there are only four types of meanin properties in Okezone.com: meaningfulness, ambihuity, redundancy, and anomaly.

Indirect associations in learning semantic and syntactic lexical relationships

Computational models of distributional semantics (a.k.a. word embeddings) represent a word’s meaning in terms of its relationships with all other words. We examine what grammatical information is encoded in distributional models and investigate the role of indirect associations. Distributional models are sensitive to associations between words at one degree of separation, such as ‘tiger’ and ‘stripes’, or two degrees of separation, such as ‘soar’ and ‘fly’. By recursively adding higher levels of representations to a computational, holographic model of semantic memory, we construct a distributional model sensitive to associations between words at arbitrary degrees of separation. We find that word associations at four degrees of separation increase the similarity assigned by the model to English words that share part-of-speech or syntactic type. Word associations at four degrees of separation also improve the ability of the model to construct grammatical English sentences. Our model proposes that human memory uses indirect associations to learn part-of-speech and that the basic associative mechanisms of memory and learning support knowledge of both semantics and grammatical structure.

Semantics Analysis of Hyponym in The “Fate and Furious 8”

This study aims to find the hyponym hierarchy and find out the relationship in Fate and Furious 8. The method in this research is descriptive qualitative using observations and non-participatory in collecting data and using informal methods in presenting results. Data were analyzed based on the association with hyponym by referring to semantic analysis. There are found 50 words included in hyponym. There are several categories of hyponym found such as family, transportation, color, appearance, number, object, country, size, animals, and parts of body. Then broken down into (8) hyponym of the family, (7) hyponym of transportation, (3) color of hyponym, (1) hyponym of appearance, (11) hyponym of numbers, (6) hyponym of objects, (4) hyponym of country, (2) hyponym of size, (3) hyponym of animal and (5) hyponym with parts of body. Based on the result, the conclusion can be taken that hyponym of number is the most dominant of hyponym categories.
 Keywords: Semantic, Lexical Relation, Hyponym

Paradigms in construction terminology

The article is devoted to the consideration of the admissibility of lexical and semantic relations in terminology, about the phenomena of synonymy, polysemy and homonymy in the term system of construction, which is one of the controversial in modern terminology and has repeatedly become the subject of research by Russian linguists. The authors focus on the concept of the term as a system. It is noted that the consistency of terminology is an organic synthesis of extralinguistic and actually linguistic factors that the term is systematic, since it is included in the system of concepts of a certain field of knowledge and it is in the system that displays all the properties of the term. The main attention is paid to the consideration of the processes of paradigmatic system taking place in the modern terminological system of the construction industry, and it is proved that the terms (as well as commonly used units of the language) are subject to various lexical and semantic processes. The most active type of paradigmatic relationship is synonymy. In the studied terminology there is also a terminological ambiguity, the basic mechanisms of development of which are isomorphic.

CREATIVITY OF LANGUAGE, LEXICAL RELATIONS, AND CULTURAL VALUES: SUNDANESE HILARIOUS PUNS IN BRILIO.NET

Purpose of the study: The objective is to be studied is the use of Sundanese language with the opposite meaning, cause-effect, comparison, and extension of meaning. The use of the Sundanese language contains cultural values of humility, advice/invitations, and friendliness. Cultural values ​​become the personality of the Sundanese brand adapted to the development of technology to make the speech products of the author also discuss local and global values.
 Methodology: This study seeks to analyze language creativity with local Sundanese nuances that follow global progress. This research used a descriptive qualitative method. The data source from 100 Sundanese hilarious puns found in brilio.net. The descriptive method was used in several stages, namely data collection, data analysis, and data presentation.
 Main Findings: This research shows that Sundanese hilarious puns found in brilio.net use language creativity which contains 35% expansion of meaning (polysemy), opposite meaning (20%), cause-effect (25%), and comparison (20%). The dominant cultural values are friendliness by 52.5% others are humility (20%) and advice (27.5%). They reflect the habits of Sundanese people who like to joke and empower technological advances so that the creativity of the language produced further reflects today's digital development.
 Applications of this study: This research is used as a tool or model in understanding and interpreting language creativity in terms of contrastive meaning, cause-effect, comparison, the extension of meaning and values, i.e.: humility, advice, and friendliness in Sundanese hilarious puns.
 Novelty/Originality of this study: Sundanese hilarious puns show that language creativity not only empowers the play of words, but also the cultural values that become the tradition of Sundanese people, namely the traditions of Sundanese who like to joke (ngabodor). Cultural and lexical relationships are combined and produce funny speech following the development of typical Sundanese technology. Sundanese culture and language will persist if Sundanese people always use and maintain them, one of which is by creating hilarious Sundanese puns that follows technological advances, such as those found in Sundanese hilarious puns in brilio.net.

Lexical Relations in Spanish-Speaking Older Adults.

This study analyzed the types of lexical relations produced by Spanish-speaking older adults with typical aging. A total of 120 older adults completed a word association norms task with 117 stimulus words, which allowed us to explore differences in associations by sex, age, and years of education. We employed two classifications to code the lexical relations: a traditional classification (paradigmatic versus syntagmatic) and a second classification categorizing responses into 17 types of associations (e.g., categorial versus non-categorial). Our results show that participants have a preference for paradigmatic responses (e.g., dog-animal), as well as associations with thematic-contextual co-occurrence plus semantic relations (e.g., cradle-baby). These findings suggest that older adults tend to establish lexical relations based on a combined link, one that is semantic and contextual.

Lexical Cohesion Roles In Speech Rhetorical Strategy of 2016 U.S Presidential Debate, Trump Vs Clinton

This research aims at describing the speech text cohesion and its contribution to the rhetorical strategy used in the 2016 U.S Presidential debate between Trump and Clinton in their first campaign debate on the political topic area on achieving prosperity, America's direction, and securing America. As an effort to reach the purpose of research, the researcher applies qualitative descriptive method on Discourse Analysis (D.A.) as the theory to describe and analyze the lexical cohesion relation and rhetorical devices linguistically used in the discourse of presidential debate. Furthermore, based on the analysis, the data were found around 399 lexical cohesion devices of reiteration and 64 collocations used by Clinton, and as many as 576 lexical cohesion devices for reiteration and 58 markers of collocation relations were applied by Trump in the debate over U.S presidential debates. Repetition is the most dominant type of lexical cohesion at U.S presidential debate, both realized by Clinton and Trump. The repetition achieved by both debate partners was 39.74% and 63.25%. Some of the lexical relation roles used by the two pairs of candidates based on the topics discussed reveal some differences in their political objectives, namely, economic and industrial issues, law, trade, investment, labor, taxes, and terrorism. Meanwhile, the rhetorical tools used by the two candidate pairs include; metaphor, a list of three, parallelism, and contrastive pairs (antithesis). In addition, the debate rhetoric technique used by both candidates includes; anaphora, epiphora, and climax.

A Cognitive Method for Automatically Retrieving Complex Information on a Large Scale

Modern retrieval systems tend to deteriorate because of their large output of useless and even misleading information, especially for complex search requests on a large scale. Complex information retrieval (IR) tasks requiring multi-hop reasoning need to fuse multiple scattered text across two or more documents. However, there are two challenges for multi-hop retrieval. To be specific, the first challenge is that since some important supporting facts have little lexical or semantic relationship with the retrieval query, the retriever often omits them; the second challenge is that once a retriever chooses misinformation related to the query as the entities of cognitive graphs, the retriever will fail. In this study, in order to improve the performance of retrievers in complex tasks, an intelligent sensor technique was proposed based on a sub-scope with cognitive reasoning (2SCR-IR), a novel method of retrieving reasoning paths over the cognitive graph to provide users with verified multi-hop reasoning chains. Inspired by the users’ process of step-by-step searching online, 2SCR-IR includes a dynamic fusion layer that starts from the entities mentioned in the given query, explores the cognitive graph dynamically built from the query and contexts, gradually finds relevant supporting entities mentioned in the given documents, and verifies the rationality of the retrieval facts. Our experimental results show that 2SCR-IR achieves competitive results on the HotpotQA full wiki and distractor settings, and outperforms the previous state-of-the-art methods by a more than two points absolute gain on the full wiki setting.

Analysis of Name Entities in Text Using Robust Disambiguation Method

<em>Named entities are proper nouns or objects contained in a text, such as a person's name, country name, and others. Names of persons in some text are often ambiguous, which makes it difficult for ordinary people to find out these same names are the same person or not. An ambiguity of names also found in hadith, like the name Abdullah in hadith number 86 and 2411, that might be the same person or might be different. Based on this problem, then this study focuses on named entity disambiguation, which considered further semantic and lexical relation between a named entity. Expected in the future, it would help people to understand the ambiguity of the name or distinguish ambiguous names. The method used in this research was Robust Disambiguation because, in this method, the context of the named entity considered. The resulted output obtained was in the form of named entity that grouped based on the same person or different person processed with Density-based Spatial Clustering of Applications with Noise. This research resulted in an accuracy value of 90%, a precision value of 97%, and a recall value of 89% obtained from actual value and predicted value</em>

Zero morphemes in paradigms

Abstract This paper sheds a new light on the notion of zero morphemes in inflectional paradigms: on their formal definition (§ 1), on the way of counting them (§ 2–3) and on the way of conceptualizing them at a deeper, mathematical level (§ 4). We define (zero) morphemes in the language of cartesian set products and propose a method of counting them that applies the lexical relations of homophony, polysemy, allomorphy and synonymy to inflectional paradigms (§ 2). In this line, two homophonic or synonymous morphemes are different morphemes, while two polysemous and allomorphic morphemes count as one morpheme (§ 3). In analogy to the number zero in mathematics, zero morphemes can be thought of either as minimal elements in a totally ordered set or as neutral element in a set of opposites (§ 4). Implications for language acquisition are discussed in the conclusion (§ 5).